From: pwoody@nyc.pipeline.com (Paul Woodworth) Newsgroups: comp.text.sgml Subject: US, SGML, Webmaster Duties, Oversite of .com Server Date: 31 May 1995 23:20:56 -0400 Organization: The Pipeline Message-ID: <3qjbmo$pas@pipe3.nyc.pipeline.com> Please forgive the Interruption, I have been asked to find some experts, with skills like yours. It occurred to me that I may ask the type person I was looking for to point me in the right direction. I would appreciate your letting me know where to post a position available ad for your profession. Thanks in advance for your assistance. -- Paul L. Woodworth pwoody@pipeline.com From: ldefurio@sgml.com Newsgroups: comp.text.sgml Subject: ANNOUNCE:Rochester NY SGML User's Group Date: 1 Jun 1995 06:03:29 GMT Organization: ServiceTech, Inc. Message-ID: <3qjl7h$4ep@murphy.servtech.com> X-Newsreader: AIR News 3.X (SPRY, Inc.) ANNOUNCEMENT Kick-off Meeting of the Rochester, NY SGML User's Group Where: The Depot Restaurant, 441 N. Main St., Pittsford When: Wednesday, June 21, at 6pm What: Appetizers and Cash Bar This meeting, sponsored by Eastman Kodak Company, is a chance for SGMLers in the Rochester, NY area to meet each other and discuss future goals for the group. Any questions can be directed to: Terry Badger, Eastman Kodak Company, lockovm2.le2j33@kodako.kodak.com, 716-477-6869 or Cindy Mazza, Information Architects, Inc., cmazza@sgml.com, 315-568-8718 ldefurio@sgml.com. ----- Lori A. DeFurio Information Architects, Inc. ldefurio@sgml.com tel: 315-568-8718 fax: 315-568-0157 Newsgroups: comp.text.sgml Subject: Re: [2Q's] image maps? (1st post short-sheeted?) Message-ID: <1995Jun1.122328.411@gw2.admin.ch> From: michael.straessle@gw2.admin.ch (Michael Straessle) Date: 1 Jun 95 12:23:28 +0100 References: <3q5blh$f7v@uuneo.neosoft.com> Organization: Verwaltungskontrolle des Bundesrates X-Newsreader: WinVN 0.99.2 MIME-Version: 1.0 In article <3q5blh$f7v@uuneo.neosoft.com>, hav@neosoft.com says... > > >I have 2 questions on image-maps. > .... > >My second question is more general. I tried to put an image-map >in a page but when I click on the image, IT SIMPLY LOADS THE >MAP-FILE - not the intended target page. ??? >My map file (test.map) looks something like... > rect trees.htm 0,0 200,75 > rect bonsai.html 0,76 200,145 >(note: these files are all in the same directory) >My html looks something like... > \ > \\\ Michael From: pdavis@copernicus.bbn.com (Peter Davis) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,comp.infosystems.www.misc Subject: Re: Summary, was Crux, was PDF vs HTML Followup-To: comp.text.pdf,alt.hypertext,comp.text.sgml,comp.infosystems.www.misc Date: 1 Jun 1995 15:01:43 GMT Organization: Bolt, Beranek and Newman Inc. Message-ID: <3qkkon$mee@info-server.bbn.com> References: <3q2aql$cgc@icarus.convex.com> \ <3q50kr$3tv@icarus.convex.com> <3q5717$d6r@mars.mcs.com> <3q6crk$ghs@ecl.wustl.edu> Reply-To: pdavis@bbn.com X-Newsreader: TIN [version 1.2 PL2] Fritz Lehmann (fritz@rodin.wustl.edu) wrote: : Jorn Barger asked: : ---begin quote--- : Is anybody willing to summarize this thread? : ---end quote--- : My attempt to summarize it is: : 7. If your chief relationship to the user is helping a willing user : by providing _information_ that he or she wants, then WWW/HTML is good. : If your chief relationship is influencing a user and getting him or : her to want something or to feel a certain way, through your presentation : skills, then WWW/HTML may not be what you want. This is a bit of an oversimplification, and implies that people who use PDF are sinister, manipulative mercenaries. I think the distinction could be better expressed as follows: If the content of your message will withstand reformatting according to reader preferences, then use HTML. If, however, the design and layout of your material are crucial to its impact, then use PDF. -- Peter Davis "Education is not the 617/873-4145 BBN Educational Technologies filling of a pail, but FAX: 617/873-2455 150 Cambridge Park Drive the lighting of a fire." pdavis@bbn.com Cambridge, MA 02140 -- W. B. Yeats URL: http://copernicus.bbn.com/0/people/PDavis/homepage.html From: swong@pobox.com (Sylvia Wong) Newsgroups: comp.text.sgml Subject: Lonely Lily Date: 1 Jun 95 15:15:22 HKT Organization: Asia On-Line Limited, Wanchai, Hong Kong. Message-ID: <3qk499$c91@news.asiaonline.net> Reply-To: swong@pobox.com (Sylvia Wong) My friend Lily lives in Hong Kong, like me. She loves to receive phone calls from foreign men. She does not have computer, so I am sending this message for her. If you want to call her and you are in the United States the number is 011 852 1747 3094. Callers from other countries need to put the international code then 852 1747 3094. No e-mail please. Sylvia Wong Newsgroups: comp.text.sgml From: john@nmt.edu (John Shipman) Subject: Re: Application of the DSSSL STTP Message-ID: <1995Jun1.160702.18805@nmt.edu> Organization: Zoological Data Processing References: \ Date: Thu, 1 Jun 1995 16:07:02 GMT Bjorn Hell Larsen \ wrote: +-- | ...So, I wonder, would it be possible to use the STTP to | down-translate a document from a given DTD to a simpler one, | such as HTML? Would it be wise to utilize DSSSL/STTP-conforming | tools in this manner, or would other tools (OmniMark, SGML | Hammer, whatever) be better? +-- I too wondered if I could use STTP to transmogrify some other SGML document type into HTML. Here's an example. I'm writing a help system in HTML. One of our standards is that the text of a link to another help page should be the same as the title of that page. That way, when a user chases a link and the new page appears, its title gives positive feedback that they went to the right place. However, in practice this means that if the title of a page is changed, we have to go and find all the references to it and change the link text to match the new title. I'd much rather design a DTD that allows pages to refer to each other by ID strings (which are not visible to the user, and don't ever have to change). Then presumably I could use the STTP or a similar tool to create the final HTML form by inserting the \...\ text of the target page as the link text at each reference. This is just one example of how the enforcement of design rules could be simplified by designing a DTD for the essentials of the document, and then using STTP to create a specific HTML presentation. Another example: we have specific standards for navigational links at the bottom of each page ("Next", "Previous", "See also", etc.). The way these links connect is actually determined by the overall outline of the help system. I'd rather maintain the outline, and have an automated process build these links, than do it manually---by looking at the outline in one window while reworking links in another. So, assuming that some robust, affordable (if not free) STTP implementations are available soon, is that the way to go? -- John Shipman (js@cs.nmt.edu) ``Let's go outside and commiserate with nature.'' --Dave Farber From: jeffrey@hks.com (Glenda Jeffrey) Newsgroups: comp.text.sgml Subject: Re: sqml2asc? Is there such a biest? Date: 1 Jun 1995 17:46:43 GMT Organization: Hibbitt, Karlsson & Sorensen, Inc. Message-ID: <3qkue3$7ks@argo.hks.com> References: <3q1p0o$su5@news.cs.tu-berlin.de> \ X-Newsreader: TIN [version 1.2 PL2] Someone wrote: : >Is there a tool to convert a sgml document to plain ascii? Yes. It comes with Earl Hood's perlSGML package. The following comes from the README file: stripsgml* Simple program to remove SGML markup. See comments at begining of source file for usage information. And the comments at the top of the source file: ## Description: ## Remove SGML markup. Reads from STDIN, writes to STDOUT. ## Character entities are converted to ASCII equivalents. However, ## due to the nature of some special characters, some special ## characters may get lost in the output. ## ## Usage: ## % stripsgml [-html] < file.sgml > file.txt ## ## The -html option cause URLs in anchor elements in an HTML document ## be preserved in the text output. To get a copy of this software, check out http://www.oac.uci.edu/indiv/ehood/dtd2html.doc.html Happy stripping ;-) -- Glenda Jeffrey Email: jeffrey@hks.com Hibbitt, Karlsson & Sorensen, Inc Phone: 401-727-4200 1080 Main St. Fax: 401-727-4208 Pawtucket, RI 02860 From: jeffrey@hks.com (Glenda Jeffrey) Newsgroups: comp.text.sgml Subject: US-RI/near MA-Need SGML tech writer Date: 1 Jun 1995 17:54:26 GMT Organization: Hibbitt, Karlsson & Sorensen, Inc. Message-ID: <3qkusi$7ks@argo.hks.com> X-Newsreader: TIN [version 1.2 PL2] HKS seeks an experienced technical writer to produce online and printed user documentation for ABAQUS, an engineering analysis software package. In order to produce the documentation, the writer is expected to use ABAQUS, interview developers, and read product specifications written by developers. The documentation will be written in SGML, on a Unix workstation. The successful applicant will be intellectually curious, detail-oriented, and able to work independently. Significant mathematical and/or engineering background is _required_; a bachelor of science degree in either field is highly desirable. Experience using engineering analysis software, though not necessarily in a documentation capacity, is also desirable. HKS develops finite element software for engineering calculations. We are located in Pawtucket, Rhode Island, just north of Providence and convenient to I-95. Reply with a resume, salary requirements, and list of references to: Documentation Department Hibbitt, Karlsson & Sorensen, Inc. 1080 Main St. Pawtucket, RI 02860. Email resumes are also acceptable; send them to jeffrey@hks.com. Please do not call. -- Glenda Jeffrey Email: jeffrey@hks.com Hibbitt, Karlsson & Sorensen, Inc 1080 Main St. Pawtucket, RI 02860 From: reh@wam.umd.edu (Richard Huddleston) Newsgroups: comp.text.sgml Subject: SGML in, PS out? Date: 1 Jun 1995 20:14:32 GMT Organization: Empire of the Senses Message-ID: <3ql738$5al@cville-srv.wam.umd.edu> Given an SGML instance, and related DTD, FOSI and ATD, I'm looking for a PD batch-mode utility which can convert this mess to PS output. I've looked around on the WWW.SIL.ORG site, and scanned the FAQs, but haven't found anything. Source files may be generated by either ArborText or SoftQuad editors, although they share a common DTD, etc. Any pointers, even to PD utilities which will get me part of the way there (with hints as to what has to be done to finish the job ;) will be greatly appreciated. I'll happily post a summary. Richard -- Richard Huddleston | "Any use of a human being, in which less is Anthropology/Computer Science | demanded of him and less is attributed to University of Maryland | him than his full status, is a degradation ...the usual disclaimers... | and a waste." --Norbert Wiener From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 1 Jun 1995 23:30:31 GMT Organization: \ Message-ID: <3qliin$38i@news1.delphi.com> References: <19950530T051645Z@naggum.no> X-Newsreader: Galahad 1.1b Erik Naggum wrote: >the purpose of the backward compatibility is to protect user investment in >SGML documents. investing in RANK (and a few other minimization >"features") >has never been smart, but from there to chopping it is not necessarily a >straight route. Ok, I will rephrase the question: I thought there was some talk about developing an intelligent standard with its focus on the future needs and not tied to backward compatibility. Your message seams to imply that backward compatibility is of \major\ importance. Is it? My personal peeves are with the interactions between ps, re, and rs in the ISO standard. I was hoping a revision of the SGML standard would replace the "record" based processing as defined in the current standard with a more "stream" based model. The current model is reminiscent of IBM mainframe style processing. ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: S177946@KUB.NL (U627 177946 M. TIMMERMANS) Newsgroups: comp.text.sgml Subject: levels of information Date: 2 Jun 1995 08:24:06 GMT Organization: Tilburg UniversityTilburg University, The Netherlands Message-ID: <3qmhr6$juq@mailnews.kub.nl> X-News-Reader: VMS NEWS v1.25 Marnix> Hello, I'm currently working on my thesis and I am trying to Marnix> rank a number of ways to store information. I've come up Marnix> with the following: Marnix> 1. Hardcopy Marnix> 2. Bitmap-image (e.g. pagescan) Marnix> 3. PDL (page description language, e.g. pcl/postscript/pdf) Marnix> 4. Page in DTP-program (e.g. Quark XPress, PageMaker) Marnix> 5. Text with style sheets / markup with style tags / HTML Marnix> 6. Structured: SGML, database or spreadsheet; Marnix> This list is more or less ordered from information in a low Marnix> level of abstraction to a higher level. My questions are Marnix> these: Marnix> - do you think this makes any sense? Marnix> - do you know of any other person or web-site/book/magazine Marnix> that addresses this subject? > Hi, > > I think that this is a `correct' ranking according to level of > abstraction (though one could begin a discussion on whether 3 and 4 > should be exchanged). If that is what you're after, you've already > got it. > > It would be interesting to see what you want to do with this. You > said you would like to store `information'. What should be done > with that? For example, hardcopy might be necessary for legal > purposes (if the document is signed). A bitmap image is obviously > easier to manage (can be stored on digital media, like CD-ROM) and > conveys information about the looks of the document to the most > people. In all other formats, you need some more software to > exactly reproduce the look of the pages stored. We could presume a > bitmap viewer on every system, though. I was asked by a printing company and an educational publisher to research the possibilities for "printing on demand". They would have liked me to set up a database (presumably in SGML). I've made some business proposals for other publishers and they all seem more interested in page-oriented database publishing. (e.g. storing pages in Quark XPress and use these to put together books in a flexible manner. It was my idea for the thesis to make a chapter with some sort of technical explenation of different ways to store information. It seemed like a nice idea to make up a list with different levels of information. I could then make a matrix: 1 2 3 4 5 1 2 3 4 5 to describe the processes of going from level to level. (e.g. 3 to 4 would be 'ripping', 1 to 2 would be 'scanning' etc.) Although this aproach feels nice to me, it causes quite a lot of headaches. E.G. Where to fit in plain ascii files? > However, for information retrieval purposes, you need a format > where the structure of the document is explained in as much detail > as possible. You would also need to create a kind of `database > schema' for that structure information to be useful. > > An intermediate level of storing things would be in a format that > lets you see just the text (possibly in addition to a rendering of > the looks on paper, eg via scanned image). Having the text would > give you the possibility of doing full-text searches which often > suffice to retrieve the kinds of documents one is looking for. Obviously each method of storing has got it's own pros and cons. I would like to analyse some publishers and their audience, their current way of doing things, future possibilities they might want etc. This could be related to the technical possibilities to come up with some recommendations and conclusions. > > Also, one needs to consider whether the document base is > homogeneous or heterogeneous (which can happen at different Unfortunately the product of the educational publisher I mentioned is making textbooks for primary schools (e.g. how to learn to read). These books are not very well structured; creative pagedesign is important in these kind of books. > levels). For instance, suppose you're storing documents in SGML > format. If there is only one DTD involved, you can just process > these documents fairly easily, but if several DTDs are involved, > you need to store the DTDs, too. Further, one DTD might give > information about author and editor and reviewer separately, > whereas another DTD might give information about `persons' only. > Thus, you should create some sort of mapping between the entities > on a semantic level. > How often do you need conversion from SGML to SGML? For the reason you mentioned it is better to keep the number of DTD's low though. My concern is more that I don't know how to get from SGML to some backend. Most publishers seem to feel restricted by SGML. Quark XPress is more aimed at designing page layout and last minute editting can be done in Quark. When using SGML however, any changes other than in the SGML-source file should be avoided. > Have I addressed your concerns at all? > > Looking forward to hear (read) your comments, > \\kai{} Thanks a lot for your reply. I hope there are more people interested in this subject, which is why I put the reply in the newsgroup. Marnix Timmermans From: mvulpe@i4i.org (Michel Vulpe) Newsgroups: comp.text.sgml Subject: SGML Application Server (Product Announcement) Date: 2 Jun 1995 13:09:34 GMT Organization: Infrastructures for Information Inc. Message-ID: \ SGML Application Server©: Product Announcement Infrastructures for Information Inc. is proud to announce a fundamental SGML tool, the SGML Application Server or "SAS"©. SAS© is a set of DLL's "Dynamic Link Libraries" for MS-Windows/DOS or C objects for Unix or Macintosh that brings the full power of SGML to the desktop. SAS is available now. Desktop tools such as MS-Word©, Oracle©, FrameMaker©, etc. can now be configured to support SGML. SAS is not a postprocess or a style translator, it is the toolkit that allows you to embed SGML into your applications. SAS has been used to provide SGML functionality to the word-processors, databases and composition systems used in Infrastructures complete SGML systems. It has been tested and proven in the C, C++, PowerBuilder©, Delphi©, and Visual Basic© environments. Technical Specifications: Approximately one megabyte of disk space inclusive of SAS, documentation and sample code on the use of SAS. The sample code includes a C program for DTD navigation, and a structure viewer in Delphi, PowerBuilder and Word Basic. SAS SGML functions are divided into service classes as follows: ° Validation ° Parsing ° DTD Navigation ° I/O ° Entity ° Session ° Error ° Misc. SAS-HTML© limited to HTML 2 and limited to non-commercial applications is available as shareware. For a copy of SAS-HTML mail us your request with your return e-mail and postal address to sashtml@i4i.org. Commercial versions of SAS with support for arbitrary DTD's can be licensed from Infrastructures for Information for $99.00 US. For information on the commercial implementation please mail us your request with your return e-mail and postal address to SAS@i4i.org or write us at Infrastructures for Information Inc. 330 Dupont, Suite 302 Toronto, Ont., M5R 1V9 416.920.6489 (voice) 416.920.6493 (fax) From: emv@mail.coast.net (Edward Vielmetti) Newsgroups: comp.text.sgml,bit.listserv.vpiej-l Subject: ISO 10283 "Serials, Electronic Manuscript" DTD Date: 2 Jun 1995 15:10:55 GMT Organization: Coast to Coast Telecommunications, Inc. Message-ID: <3qn9lv$n06@simtel.coast.net> X-Newsreader: TIN [version 1.2 PL2] It looks like I am going to be using this DTD pretty heavily over the course of my next project, and I was hoping to find some kindred souls who were in the same situation. The eventual delivery platform is some combination of specialized SGML viewers and general purpose HTML viewers, and there is a body of page images that goes along with the marked up text just to make things special. From reading the comp.text.sgml archives it appears clear that the maths part is a challenge; I'm inclined to believe that the entire citation / bibliographic apparatus is going to be a good piece of effort as well esp. if the logical progression toward Hytime linking is used. I have an undated, unmarked draft of the standard, and have managed to find a pointer to EPSIG ("Electronic Publishing Special Interest Group") at http://www.sil.org/sgml/epsig.html (Robin Cover's archive). I'm sure there's other projects underway that have as their goal some substantial body of work completed within this DTD, or perhaps using DTDs based substantially on this one - I'd like to see a big more of that compiled in one place to help the process along. thanks Ed Edward Vielmetti U of Michigan Humanities Text Initiative emv@coast.net emv@umich.edu (not working just yet) From: twayne@julian.uwo.ca (Tim Wayne) Newsgroups: comp.text.sgml Subject: Microstar location needed. Help Date: Fri, 2 Jun 1995 15:19:25 GMT Organization: SLIS Message-ID: \ X-Newsreader: Trumpet for Windows [Version 1.0 Rev A] I would like to know how to contact Microstar, a company involved with various SGML utilities located in Ottawa Ontario. An e-mail address or URL would be great. Thanks in advance. Tim Wayne twayne@julian.uwo.ca Tim Wayne Graduate School of Library and Information Science From: amiano@marburg.roc.servtech.com Newsgroups: comp.text.sgml Subject: Re: Application of the DSSSL STTP Date: 2 Jun 1995 16:54:04 GMT Organization: ServiceTech, Inc. Message-ID: <3qnfnc$cer@murphy.servtech.com> References: <3qf9a0$4en@Mercury.mcs.com> X-Newsreader: AIR News 3.X (SPRY, Inc.) > svb@MCS.COM (Stephan Bechtolsheim) writes: ..[trimmed for brevity]... > Let's say you have a DTD for the papers of a conference, where the DTD > accomodates one paper at a time. Now you would like to publish a book > containing all the papers of a conference. This book has its own DTD, > the abstracts of all the papers need to be moved into the first part > of the book, all paper titles become titles of chapters of the book - > and after all this is done you need to format the document. > > You use the STTP to make one book out of the individual papers, subsequently > you use the STFP to format the book. ..[trimmed for brevity]... Now I am curious. A superficial look at DSSSL makes it look interesting from the standpoint of automated composition systems, but the closer looks I have taken indicate that it may not be appropriate for certain uses. For instance, I noticed a restriction on rotations to 90 degree increments. DSSSL really looks interesting, and I'd like to know the opinion of those who are actively engaged with it, if something like a database publishing system, or a document & image composition system are inappropriate applications, or just what was intended? Is DSSSL general enough to allow a standardized way of doing as-yet unimagined ways of doing style, semantics, and formatting? (Are there any descriptions of working DSSSL pilots or systems out there?) From: nettagger@aol.com (NETTAGGER) Newsgroups: comp.text.sgml Subject: Re: sgml for educational publishers Date: 2 Jun 1995 16:12:48 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3qnrc0$sue@newsbf02.news.aol.com> References: <3qinas$8ss@mailnews.kub.nl> Reply-To: nettagger@aol.com (NETTAGGER) The issues in terms of costs are the issues of your legacy data conversion. Our organization has the technology and low cost labor force to allow you the financial lattitude to exploit the benefits brought by SGML. As well, my background is in educational publishing. My suspicions are that you could be looking at internal costs reductions at between 50 to 70% on new product development for text, ancillaries and pedagogical tools. Flash your needs, I'll be glad to help Nettagger From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: DTD for resume Date: 3 Jun 1995 03:43:24 GMT Organization: \ Message-ID: <3qolos$j1t@news1.delphi.com> References: \ X-Newsreader: Galahad 1.1b Marcy Thompson had some interresting comments about resumes. >Anyway, if I try to write a resume DTD that doesn't itself contain >a lot of information about how to lay stuff out on the page, then >I would anticipate that every single person would need to design >a customized post-process in order to print it. One of the things I do besides SGML is develop macros for TeX. For example, I wrote a set of macros to typeset a resume based on the DTD I wrote. But I actually did a bit more than that. The TeX macros were designed in a modular way to allow three separate documents to be produced from the same source file. The first document would be a full resume in the order the data falls. A second format is the same information formatted in a monospaced font (so it can be converted back into ASCII after using TeX as the formatter). The third format was a 1 page chronological summary. This goes to the issue of separating form from content. It should be possible to develop a DTD for a resume that contained all the possible information about a person. But the DTD should \not\ contain the formatting specifications of how to present the information. One of my co-workers has been looking for a better paying job. He ran across an authoring tool that lets you put text, graphics, and sound into a stand alone program. He took some of his ray-traced artwork, some digitized sounds, and the text of his resume to create a program that shows off his artistic skills and technical expertise. Now all he needs to do is find a program to help him create an easy installation procedure since the resume is around 3 meg compressed. It also seems a bit of overkill to require 10 meg of disk space to install someone's resume. Using a HTML style DTD it should be possible to develop a set of program that could automatically do that. The basic idea is to have "late binding" of the presentation of the information. Isn't that was DSSSL is all about? (I am only starting to work my way through DSSSL.) One of the concepts in the DSSSL documents I have been reading is that you should be able to have multiple output specifications for one DTD. Now the last time I did some serious job hunting, I was re-arranging my resume for each job. I also suppressed most of the information about skills that were not mentioned in the job description. I would read the job description and re-order my resume to match the order in the job description. That was a lot of work. But it got me out of job I hated and into one that I love (most of the time). ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: reh@wam.umd.edu (Richard Huddleston) Newsgroups: comp.text.sgml Subject: Rhetorical question Date: 3 Jun 1995 04:17:52 GMT Organization: Empire of the Senses Message-ID: <3qonpg$5er@cville-srv.wam.umd.edu> Forgive my ignorance, but--as my question will indicate--I am quite new to SGML. What would be involved if one were to wish to translate an SGML application to its equivalent PostScript form? Using a commerical publisher, and its command-line switches, this is easy enough to accomplish; what I'm trying to understand is just what the hell is going on inside of the black box. Simple pointers to references appreciated, in lieu of a written-out explanation. Richard -- Richard Huddleston Angel wannabe From: bois@irisa.fr (Philippe Bois) Newsgroups: comp.text,comp.text.sgml,comp.text.tex Subject: The publisher Date: 3 Jun 1995 08:29:56 GMT Organization: IRISA, Campus de Beaulieu, 35042 Rennes Cedex, FRANCE Message-ID: <3qp6i4$qap@news.irisa.fr> Reply-To: bois@irisa.fr I'am searching for a script with 'The publisher' editor commands to sort elements with the same tags. Before tag1 tioga tag1 tag1 grif tag1 After tag1 grif tag1 tag1 tioga tag1 Any help is appreciated. Thanks. From: bois@irisa.fr (Philippe Bois) Newsgroups: comp.text.sgml Subject: The Publisher Date: 3 Jun 1995 08:34:12 GMT Organization: IRISA, Campus de Beaulieu, 35042 Rennes Cedex, FRANCE Message-ID: <3qp6q4$qap@news.irisa.fr> Reply-To: bois@irisa.fr I'am searching for a script with 'The publisher' editor commands to sort elements with the same tags. Before tag1 tioga tag1 tag1 grif tag1 After tag1 grif tag1 tag1 tioga tag1 Any help is appreciated. Thanks. From: photopres@aol.com (PHOTOPRES) Newsgroups: comp.text.sgml Subject: SGML acceptance Date: 3 Jun 1995 08:25:36 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3qpkc0$bas@newsbf02.news.aol.com> Reply-To: photopres@aol.com (PHOTOPRES) I've been following this list for a while and I've read 'Practical SGML' by Eric van Herwijnen. I find SGML interesting, but I'm wondering what the scope of its acceptance in the marketplace is. My perspective is from running a small digital composition service for small to medium sized book publishers. In our discussions with probably 20 to 25 publishers in the past 3-4 months, only one has mentioned a requirement for SGML. I would like to move in this direction, but I'm uncertain if the market is (or will be) there. Any comments would be appreciated. Jeff Schmidt Appalachian Prepress From: pwoody@nyc.pipeline.com (Paul Woodworth) Newsgroups: comp.text.sgml Subject: DTD for resume Date: 3 Jun 1995 12:16:20 -0400 Organization: The Pipeline Message-ID: <3qq1sk$sgq@pipe1.nyc.pipeline.com> References: <3q261a$6v6@yama.mcc.ac.uk> <3q85vs$4el@news1.delphi.com>\ In comp.text.sgml mlt@netcom.com (Marcy Thompson) said: >And frankly, I think that's the right advice for most people who are not >compiling huge databases of resumes. (It has crossed my mind that it might be >useful to make a sort of generic resume-printing tool based on SGML, for use by >college employment services or something. But as pepole get more experienced, >resumes get more individualized, and I don't really see the benefit from using >SGML to create them.) > >Marcy Marcy, I currently access several points on the net that are publishing using the Adobe .pdf files. Even though this requires a reader, the reader is available for free, this format allows formatted documents to be delivered, particularly important for resumes. Y'all may want to try pointing a browser at nytimesfax.com and see how the technology works. Scanning and Mac type setting are a couple of the advantages that I see. Anyway, what more an you ask, the times puzzle delivered to your desktop daily!!! -- Paul L. Woodworth pwoody@pipeline.com From: clive@d-supp.demon.co.uk (Clive Bruton/Typonaut) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 3 Jun 1995 20:08:51 +0100 Organization: Digital Support Sender: news@news.demon.co.uk Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> <3ph57d$175@nntp.interaccess.com> <3pi6dp$8g6@news.iag.net> <3q0673$qnc@frame.frame.com> <3q7muf$d0e@martha.utk.edu> X-Nntp-Posting-Host: d-supp.demon.co.uk In article <3q7muf$d0e@martha.utk.edu>, syost@hephp01.phys.utk.edu (Scott Yost) wrote: > It is not less important, just different. That is why it would be unfortunate > to see paper publishing formats mindlessly duplicated in the computer > medium. The rapid increase in the information-handling capabilities implies > that it is just a matter of time before the readers of most web documents are > no longer human. They will be programs which intelligently extract information > from the Web and present it in a way most useful to the reader. This is > already happening: most people find subjects of interest in my own WWW pages > through robotic indexes. Such indexes are now clearly primitive compared > to the robots which will evolve as the web's complexity grows. Only document > formats flexible enough to survive the rapid transitions ahead will continue > to have relevance. Are you saying that cannot be done with pdf? You may like scanning screens of information, fortunately most people still prefer books. Clive Bruton (AKA The Typonaut) clive@d-supp.demon.co.uk From: clive@d-supp.demon.co.uk (Clive Bruton/Typonaut) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 3 Jun 1995 20:12:40 +0100 Organization: Digital Support Sender: news@news.demon.co.uk Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> <3mkm90$1q0@everest.pinn.net> \ \ <3q0673$qnc@frame.frame.com> <3q9i94$eeo_002@huji.ac.il> X-Nntp-Posting-Host: d-supp.demon.co.uk In article <3q9i94$eeo_002@huji.ac.il>, MARDER@agri.huji.ac.il (Jonathan B. Marder) wrote: > The question really is how much *content* in conveyed in the > presentation. Sure a picture says a thousand words, but it can also > cripple the utility for modern computer usage. I could put up > beautifully crafted pages of poetry all in GIF - you'd see exactly what > I wanted you to see, but could you search for a particular expression? > Newpapers and books etc. are/were designed for human reading. Electronic > information also serves as databases for various type of searching. But that is the point of pdf, you can have great typography and illustrations, and they can be searchable an editable. And what's more they can be a hell of a lot smaller than gif files. Clive Bruton (AKA The Typonaut) clive@d-supp.demon.co.uk From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 03 Jun 1995 20:48:36 GMT Organization: Naggum Software; +47 2295 0313 Expires: 17 Jun 1995 20:48:35 GMT Message-ID: <19950603T204836Z@naggum.no> References: <19950530T051645Z@naggum.no> <3qliin$38i@news1.delphi.com> [Jeffrey McArthur] | I thought there was some talk about developing an intelligent standard | with its focus on the future needs and not tied to backward | compatibility. these are not mutually exclusive options. a "new" SGML that "fixed" all the "broken" aspects of the "old" SGML would have to jump through the same political and practical hoops to become accepted as has taken more than ten years for the "broken" SGML. not building upon the foundation of general acceptance for the "broken" SGML would be madness. that said, there are many things that can be fixed without breaking well-formed and benignly conforming documents. there are also a number of stupid things people do that were never intended to be allowed, but which was also very hard to outlaw before the depth of the bottomless pit of human ignorance could be gauged through practical experience. e.g., using non-punctuation characters in short reference delimiters. | Your message seams to imply that backward compatibility is of | \major\ importance. Is it? yes, of course. but there's backward _compatibility_ and _backward_ compatibility. the former addresses the legitimate investment concerns that will make or break SGML:TNG and possibly take SGML with it. the latter addresses the "requirement" by ignorant users who have made a whole slew of design mistakes and don't want to fix them in SGML, but rather want to prove that their perl wizardry is a necessary precondition to success. ("perl" is just the most likely tool, but _any_ tool that is touted as "essential" is a likely candidate for this kind of "backward" thinking.) one of the fundamental assumptions in the SGML world is that in terms of value and costs data > software < specification if you don't think individual pieces of software are basically worthless, consider your willingnes to "upgrade" to new versions of said software, and the fact that most of the highest quality software is free, anyway. SGML's strength lies in protecting the value and investment in the data and in the specifications. a conscientious software engineer has already planned the obsolesence of the _next_ generation software to implement his requirements and constantly evaluates the benefits of dumping particularly megalomaniac vendors against the costs of adopting tools with slightly higher training needs. so we're not talking bug compatibility the way most vendors never get out of their caves, but about the freedom to make long-term decisions in a world of moving targets. SGML should spell "stability", but mainaa | My personal peeves are with the interactions between ps, re, and rs in | the ISO standard. I was hoping a revision of the SGML standard would | replace the "record" based processing as defined in the current | standard with a more "stream" based model. The current model is | reminiscent of IBM mainframe style processing. the difference between "record" and "line" is that "line" could be interpreted as something that occurred on the output side of the system, at least at the time it was defined, while the "record" was unambiguously an "input concept". further, the record concept in SGML is abstract and general enough that very few actual systems operate this way, least of all IBM mainframes (they don't have codes to start or end records). note that those systems that use line terminators don't obey the literal meaning of the codes as "start" and "end", either. how would you like to see PS, RE, and RS interact? I would not rule out that we can use a "mode switch" (see Tracy Kidder: The Soul of a New Machine for a discussion of the evils of mode switches to handle incompatible advances in technology), but it should be well argued and no options to maintain backward compatibility should be left unexamined. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: amiano@marburg.servtech.com (Mitch C. Amiano) Newsgroups: comp.text.sgml Subject: Re: levels of information Date: 3 Jun 1995 21:13:23 GMT Organization: ServiceTech, Inc. Message-ID: <3qqj9j$o4k@murphy.servtech.com> References: <3qmhr6$juq@mailnews.kub.nl> Reply-To: amiano@marburg.roc.servtech.com X-Newsreader: WinVN 0.92.6+ In article <3qmhr6$juq@mailnews.kub.nl>, S177946@KUB.NL (U627 177946 M. TIMMERMANS) says: > > Marnix> Hello, I'm currently working on my thesis and I am trying to > Marnix> rank a number of ways to store information. I've come up > Marnix> with the following: > > Marnix> 1. Hardcopy > Marnix> 2. Bitmap-image (e.g. pagescan) > Marnix> 3. PDL (page description language, e.g. pcl/postscript/pdf) > Marnix> 4. Page in DTP-program (e.g. Quark XPress, PageMaker) > Marnix> 5. Text with style sheets / markup with style tags / HTML > Marnix> 6. Structured: SGML, database or spreadsheet; > I think the idea is interesting. What you are trying to do appears to be to produce a taxonomy of information representation methods. Might I suggest that you may be muddling the concepts of how something is percieved by an information consumer versus how infomation is encoded for transmission in a message? For instance, we percieve text as a series of symbolic words which are themselves a series of symbolic characters. To a mechanized system on the other hand, text is a series of (usually) fixed size bit strings which serve as indexes into a table of vector-art or bitmaps, as well as a (usually implied) specification of a table for validating the bit strings (eg ASCII, EBCDIC). Perhaps "how abstract is the information" is not the right question to be asking. After all, most humans generally consider writen and verbal language to be first class representations... they are intuitively able to derive meaning from them. Mechanized representations (like ASCII) and not "more abstract"... they are totally unreadable to a human without some intermedate translation step into some sensory form amenable to human interpretation. I think a salient question is, 'How many steps must be taken to make data in a given form amenable to interpretation as information, either by a computer or a human?' After all, data is just data, but information has meaning. This seems more directed toward your goal, too. By the way, some may argue that PDF is a fair bit'well structured'; at least, it beats the pants off of PostScript in this respect, and is as structured as one of those DTP formats you mention above. We seem to be doing very similar things. Drop me an email line if you would like to discuss the publishing topics further offline. From: sthompso@ix.netcom.com (Scott Thompson) Newsgroups: comp.text.sgml Subject: Job Posting Date: 4 Jun 1995 04:51:33 GMT Organization: Netcom Message-ID: <3qre4l$9ti@ixnews4.ix.netcom.com> Metatec Corporation is a leading information technology company serving two key markets; organizations seeking to publish information on CD-ROM, and personal computer users who own CD-ROM equipped, multimedia equipment. Metatec offers state-of-the-art manufacturing operations and distribution facilities, complemented with comprehensive software development services, for organizations creating custom CD-ROM products. To serve the consumer market Metatec publishes the NautilusCD, a monthly multimedia CD-ROMS delivering the latest software and multimedia content to the significantly expanding base of multimedia computer enthusiasts worldwide. Metatec's common stock is traded on the NASDAQ National Market System under the symbol "META" Currently there is a position open for a Data Developer in our Saratoga, CA office and in out Columbus, OH office. Primary responsibilities include converting SGML, or other data tagging formats, to our data format (FulText) using OmniMark or other similar tools. The job also include data analysis; DTD development; and working with customers to resolve data issues. Some travel is required Skills Required * Programming Experience * Data Analysis * Conversion of data * Good communication and writing skills. Additional Skills Preffered * Visual Basic * Visual C++ * OmniMark * SGML * FulText CONTACT: Metatec Corporation Attn: Jerry Wein 1821 Saratoga Avenue Suite 204 Saratiga, CA 95070 Fax: (408) 777-0717 E-MAIL: jwein@metatec.com Metatec Corporation Attn: Scott Thompson 7001 Metatec Blvd. Dublin, OH 43017 Fax: (614) 766-3150 E-Mail sthompson@metatec.com From: Peter Murray-Rust \ Newsgroups: comp.text.sgml Subject: Conversion of ASN.1 to SGML? Date: 4 Jun 1995 09:23:01 +0100 Organization: Myorganisation Sender: news@news.demon.co.uk Message-ID: <802224869snz@ursus.demon.co.uk> Reply-To: Peter@ursus.demon.co.uk X-Newsreader: Demon Internet Simple News v1.29 X-Posting-Host: ursus.demon.co.uk I would be grateful to know if there is a way of converting ASN.1 to SGML. (Chemical Abstracts, and the National For Biotechnology Information both have dictionaries based on this standard.) Having looked at these it seems that it might be possible to convert fairly automatically (although I am sure the reverse is not true as ASN.1 does not provide for references in its syntax.) TIA Peter -- Peter Murray-Rust, Glaxo Wellcome, domestic net connexion From: nik@dreamland.tcp.co.uk (Nick Barron) Newsgroups: comp.text.sgml Subject: Re: Getting started in SGML Date: Sun, 4 Jun 95 14:01:41 GMT Organization: None Message-ID: <19950604.312578.CD4A@dreamland.tcp.co.uk> References: <3q8f6c$sl4$1@mhafc.production.compuserve.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-NewsSoftware: GRn 2.1 Feb 19, 1994 In article <3q8f6c$sl4$1@mhafc.production.compuserve.com> Aaron Minnick <73404.1504@CompuServe.COM> writes: > I am a part of a startup enterprise looking to focus on online > documentation of various types (SGML, HTML, WinHelp, PDF). I am I've a couple of bits I can snail-mail you; a good general introductory article and a few extracts from reports I've produced. Let me know your address if this is of interest, Cheers, Nik From: dmatthewson@cix.compulink.co.uk (David K Matthewson) Newsgroups: comp.text.sgml Subject: Browser for stand alonbe non net use Date: 4 Jun 1995 17:52:54 +0100 Organization: Demon Internet News Service Sender: news@news.demon.co.uk Message-ID: <3qsod6$pe6@imp.demon.co.uk> X-Nntp-Posting-Host: spol-lon.demon.co.uk X-Newsreader: Forte Free Agent v0.55 I'm looking for an SGML browser for stan alone use to run under Windows. This would be used as a document viewer for an in house manual. Whilst Netscape etc works fine, it has lots of controls on it which an in house user doesn't need and which confude them. All suggestions gratefully recived. TiA DMATTHEWSON@CIX.COMPULINK.CO.UK FAX: +44-(0)-1580-200493 ------------------------------- From: "Andre J. Emmell" \ Newsgroups: comp.text.sgml Subject: Purpose of SGML in 3 paragraph Date: 4 Jun 1995 21:04:58 GMT Organization: Global-X-Change Message-ID: <3qt75q$po8@news.globalx.net> This may sound obvious and trivial a question but I wonder if there is an elegant, clear and simple answer to the following question that keeps coming up during my workshops. In what way is an SGMl environment better than a well organized Word Perfect (MsWord, or mainframe) environment? I can think of a million reason for believing an insisting on SGML but I have yet to "convince" a "reluctant" manager that SGML is a better way to go in organizing his or her library of documents. I often find myself coming up with loads of technical reasons and success stories but I want a VERY simple "here this is the simple reason why you would invest all this time, effort and money in SGML". Can some help me in formulating such a simple and convincing answer? From: kmc@sfs1.specialform.com (Keith M. Corbett) Newsgroups: comp.text.sgml Subject: Re: sgml2asc? Is there such a beast? Date: 05 Jun 1995 01:27:58 GMT Organization: Special Form Software Message-ID: \ References: <3q1p0o$su5@news.cs.tu-berlin.de> <19950525T142043Z@naggum.no> Reply-To: kmc@specialform.com In-reply-to: Erik Naggum's message of 25 May 1995 14:20:43 UT Erik, I don't doubt that your reply to Mr. Adam is entirely correct: there is no general solution for translating SGML into "plain ASCII text". As you say, we don't know what kind of translation is required, nor do we know what tools are available to him. However, it's worth mentioning some specific tools that can be used for such purposes, if one makes some very basic (conventional) assumptions regarding input and output. Since I don't have a copy of the original posting, I'll pass along my suggestion here: consider using the "sgmlsasp" program with a null replacement file. Someone posted this trick to the www-html list a while back, and I have found it useful several times since then. Here's a self-contained example for Unix: #!/bin/sh sgmls - <\ \ ]> \ \

This is a simple text sample. \

It contains hardly any SGML markup at all, and does not stress this idea of "stripping" away SGML markup. !EOF Running this script produces the following: sfs1:~/sgml$ plain-example This is a simple text sample.It contains hardly any SGML markup at all, and does not stress this idea of "stripping" away SGML markup.sfs1:~/sgml$ This technique strips the markup right down to the bare bones, running the text together without any formatting, e.g. line breaks. Presumably, if one knows the ASP replacement file format, one could perform more sophisticated conversions. The sgmlsasp program comes with the sgmls parser. It is not included with "sp", which leads me to suspect it is deprecated. -- Keith M. Corbett kmc@specialform.com Special Form Software +1 617 596 7021 Newsgroups: comp.text.sgml From: papresco@calum.csclub.uwaterloo.ca (Paul Prescod) Subject: Re: DTD for resume Message-ID: <3qtmkg$j1d@calum.csclub.uwaterloo.ca> Sender: news@undergrad.math.uwaterloo.ca (news spool owner) Organization: University of Waterloo Computer Science Club References: <3q261a$6v6@yama.mcc.ac.uk> <3q85vs$4el@news1.delphi.com> \ Date: Mon, 5 Jun 1995 01:28:48 GMT In article \, Marcy Thompson \ wrote: >However, I remain convinced that individuals have very little >need for a resume DTD. To me, making a resume has two purposes: >attracting someone's attention and then selling myself to them >once they decide to read it. The second obviously has to do >somewhat with content, but not entirely. And the first is almost >entirely format. [...] >Anyway, if I try to write a resume DTD that doesn't itself contain >a lot of information about how to lay stuff out on the page, then >I would anticipate that every single person would need to design >a customized post-process in order to print it. The term "customized post-process" makes it sound harder than it needs to be. It could be as simple as modifying a template in your favourite word processor. Then you can lend the customizations to your friends who could "pour" their resume into it, or you could sell it to those with lesser design skills. I think it would be useful to be able to buy a hundred resume templates and easily pour my resume data into my favourite. >But as pepole get more experienced, resumes get more individualized, >and I don't really see the benefit from using SGML to create them.) You are right, SGML is usually not currently the best way. An unfortunate property of the paper resume is that there is little content and a great deal of formatting. Things change, however, if you are looking at online resumes. Online resumes are very interesting because they combine the brevity of traditional resumes with the ability to click a link to get more information. Breaking up the document and creating the links between sections is a reasonable SGML application. If you get a customizable printed version out of the deal, that is icing on the cake. And if there were some form of standard resume DTD, then someone could set up a "for-hire" database. A question: As an employer, would you rather see resumes stripped of the formatting "on equal footing" so that you could judge the content without prejudice? Or do you like the formatting so that you can weed out some based on their poor layout skills (or poor effort)? Is the fancy layout something that we should work to conserve in online resumes, or to suppress? Paul Prescod Newsgroups: comp.text.sgml From: hyama@kansai.oki.co.jp (Hideki Yamamoto) Subject: Machine translation system for SGML text? Message-ID: \ Sender: news@kansai.oki.co.jp Reply-To: hyama@kansai.oki.co.jp Organization: Oki Electric Industry Co.,Ltd. Kansai Lab. Osaka, Japan. Date: Mon, 5 Jun 1995 02:48:04 GMT Does anyone have a pointer to a machine translation for SGML text(or other formated text)? I am searching for a paper on the machine translation system that saves document format after translating. Thanks in advance. Hideki ---------- * ---------- * ---------- * ---------- * --------- Hideki Yamamoto Kansai Laboratory Oki Electric Ind. Co., Ltd. Crystal Tower, phone: +81-6-949-5101 2-27 Shiromi 1-Chome Chuo-ku, fax: +81-6-949-5108 Osaka 540,Japan email: hyama@kansai.oki.co.jp ---------- * ---------- * ---------- * ---------- * --------- From: mackie@krsun.nslab.ntt.jp (Naoto Makinae) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 05 Jun 1995 04:11:46 GMT Organization: NTT Network Service Systems Laboratories, Tokyo, Japan Message-ID: \ References: <3qt75q$po8@news.globalx.net> In-reply-to: "Andre J. Emmell"'s message of 4 Jun 1995 21:04:58 GMT In article <3qt75q$po8@news.globalx.net> "Andre J. Emmell" \ writes: | This may sound obvious and trivial a question but I wonder | if there is an elegant, clear and simple answer to the | following question that keeps coming up during my workshops. | | In what way is an SGMl environment better than a well organized | Word Perfect (MsWord, or mainframe) environment? | | I can think of a million reason for believing an insisting on SGML | but I have yet to "convince" a "reluctant" manager that SGML | is a better way to go in organizing his or her library of | documents. I often find myself coming up with loads of | technical reasons and success stories but I want a VERY | simple "here this is the simple reason why you would | invest all this time, effort and money in SGML". | | Can some help me in formulating such a simple and convincing | answer? | A word processor marks up and stores a document in its proprietary way. In order to read such a document file, you must use the word processor which has stored the document, or use a processor which supports the file format in which the file has been stored. Suppose there are one hundred word processors and each has its own mark up method. A processor has to support ninety nine methods as well as its own method to read any document which has been written by other processors. If a new processor emerges, the existing processor has to implement a new markup method. This sounds costly and troublesome. Hence, a standard markup method is desirable. Any document stored in this format will be read by any word processor (or SGML viewer) no matter which processor (or SGML editor) has witten the document. SGML (Standard Generalized Markup Language) is such a standard markup language and is defined in the ISO standards. -- Naoto Makinae Network Service Systems Laboratories Nippon Telegraph and Telephone Corporation E-Mail : mackie@krsun.nslab.ntt.jp From: donday@bga.com (Don Day) Newsgroups: comp.text.sgml Subject: Re: Machine translation system for SGML text? Date: 5 Jun 1995 07:24:40 GMT Organization: Real/Time Communications - Bob Gustwick and Associates Message-ID: <3qubfo$59p@giga.bga.com> References: \ Reply-To: donday@bga.com (Don Day) X-Newsreader: IBM NewsReader/2 v1.03 In \, hyama@kansai.oki.co.jp (Hideki Yamamoto) writes: > >Does anyone have a pointer to a machine translation >for SGML text(or other formated text)? >I am searching for a paper on the machine translation system >that saves document format after translating. You might find some articles that relate to your question among these references: - http://www.oclc.org/oclc/research/publications/publications.html (the OCLC Fred project) - ftp.cis.ohio-state.edu:pub/chameleon (the Integrated Chameleon Architecture project) - http://www.ebt.com/ebtbooks/dynatag/ (the Rainbow and DynaTag translation technologies) -- Don R. Day donday@bga.com (hobby mail and WWW access) donday@vnet.ibm.com (document systems analyst) From: dgdillaman@aol.com (DGDillaman) Newsgroups: comp.text.sgml Subject: Re: Microstar location needed. Help Date: 5 Jun 1995 03:25:58 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3qubi6$eom@newsbf02.news.aol.com> References: \ Reply-To: dgdillaman@aol.com (DGDillaman) The following is from their manual: Phone: 1-800-267-9975 or (613)727-5696 Fax: (613) 727-9491 Email: cade@microstar.com From: donday@bga.com (Don Day) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 5 Jun 1995 07:50:54 GMT Organization: Real/Time Communications - Bob Gustwick and Associates Message-ID: <3qud0u$59p@giga.bga.com> References: <3qt75q$po8@news.globalx.net> \ Reply-To: donday@bga.com (Don Day) X-Newsreader: IBM NewsReader/2 v1.03 In \, mackie@krsun.nslab.ntt.jp (Naoto Makinae) writes: > >In article <3qt75q$po8@news.globalx.net> "Andre J. Emmell" \ writes: > | This may sound obvious and trivial a question but I wonder > | if there is an elegant, clear and simple answer to the > | following question that keeps coming up during my workshops. > | > | In what way is an SGMl environment better than a well organized > | Word Perfect (MsWord, or mainframe) environment? (clip!) >A word processor marks up and stores a document in its proprietary >way. In order to read such a document file, you must use the word >processor which has stored the document, or use a processor which >supports the file format in which the file has been stored. > (clip!) More than being just printed pages, your documents are databases of information, and SGML helps ensure access to that information. See http://www.ucc.ie/info/net/markup.html for one clear example of the value of SGML over proprietary word processor formats. -- Don R. Day donday@bga.com (hobby mail and WWW access) donday@vnet.ibm.com (document systems analyst) From: S177946@KUB.NL (U627 177946 M. TIMMERMANS) Newsgroups: comp.text.sgml Subject: levels of information Date: 5 Jun 1995 12:51:34 GMT Organization: Tilburg UniversityTilburg University, The Netherlands Message-ID: <3quukm$j2j@mailnews.kub.nl> X-News-Reader: VMS NEWS v1.25 Hello Marnix> It was my idea for the thesis to make a chapter with some Marnix> sort of technical explenation of different ways to store Marnix> information. It seemed like a nice idea to make up a list Marnix> with different levels of information. I could then make a Marnix> matrix: Marnix> 1 2 3 4 5 Marnix> 1 Marnix> 2 Marnix> 3 Marnix> 4 Marnix> 5 Marnix> to describe the processes of going from level to Marnix> level. (e.g. 3 to 4 would be 'ripping', 1 to 2 would be Marnix> 'scanning' etc.) Although this aproach feels nice to me, it Marnix> causes quite a lot of headaches. E.G. Where to fit in plain Marnix> ascii files? > I also think that this is a good idea. You might want to leave > ascii files by the wayside and explain about them in a footnote. > OTOH, though, it might be possible to fit them in: In a plain > ascii file, all formatting is already specified except for the > choice of the font used to render it. Does that fit in any of your > categories? What formatting is in an ascii file? There's no underlining, bold/italic, font etc.? Marnix> How often do you need conversion from SGML to SGML? For the Marnix> reason you mentioned it is better to keep the number of Marnix> DTD's low though. > I wasn't thinking of a conversion between two SGML formats. I was > solely thinking of the retrieval aspect where you need to map the > semantics of the query onto the different DTDs. > > As I understand your application, you won't need to do any > retrieval, a document id (eg the ISBN) would be sufficient. To > print another copy of the book with id FOO, you would retrieve it > from the database and print it. That was what I think you are > trying to do. The only reason for SGML would be to make different kind of output methods (backends: e.g. offset-print, toner-print, internet, cd-rom etc. etc.) Marnix> My concern is more that I don't know how to get from SGML to Marnix> some backend. > What'a a backend? There must be a way of converting SGML to a > bitmap (otherwise printing of SGML documents would be impossible). > Why couldn't there be a way of converting SGML to postscript, too, > or to DVI, or whatever. Yes, but the work that is normally done in Quark XPress (specifying fonts, typographical issues, placement of graphics) has to be done somewhere. Information about the representation of the information has to be added. The ISO-committee is (as far as I understand) trying to solve this with the DSSSL-standard. I suppose that the document-management systems in use today have built in possibilities for this. It is also possible to use TEX and macro's and so on, but this doesn't seem very user-friendly and flexible. Marnix> Most publishers seem to feel restricted by SGML. Quark Marnix> XPress is more aimed at designing page layout and last Marnix> minute editting can be done in Quark. > I think there is this other standard, ODA (I think the format is > called ODIF or something), where there are two views onto each > document, a logical view and a layout view. Thus, storing > documents in this format would enable you to specify the layout as > well. ODA is not very popular at the moment with publishers (and in the rest of the world). The amount of software for SGML far exceeds that of ODA. IMHO ODA is on a dead track. Marnix Timmermans From: suthers+@pitt.edu (Daniel D Suthers) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 5 Jun 1995 13:53:31 GMT Organization: LRDC, University of Pittsburgh Message-ID: <3qv28r$1tg@usenet.srv.cis.pitt.edu> References: <3qt75q$po8@news.globalx.net> \ <3qud0u$59p@giga.bga.com> Reply-To: suthers+@pitt.edu X-Newsreader: TIN [version 1.2 PL2] I want to emphasize the importance of Don's response to this thread (i.e., that SGML treats a document as a structured database of information). Word processors are increasingly able to read each other's formats, and I suspect that whoever Andre needs to sell SGML to will know this. So the argument won't fly on that alone. Your argument will be stronger if you can identify a need to treat your documents as a database -- e.g., to select portions in response to a query, automatically generate specialized documents, etc. For example, suppose your company suddenly needed to make portions of the documents available on the internet for use by customers or service personnel at other sites. With SGML based documents it would be much easier to have a server automatically find relevant portions and convert them to HTML on the fly. [Don Day (donday@bga.com) 5 Jun 1995 07:50:54 GMT]: | In \, mackie@krsun.nslab.ntt.jp (Naoto Makinae) writes: | > | >In article <3qt75q$po8@news.globalx.net> "Andre J. Emmell" \ writes: | > | In what way is an SGMl environment better than a well organized | > | Word Perfect (MsWord, or mainframe) environment? | (clip!) | >A word processor marks up and stores a document in its proprietary | >way. In order to read such a document file, you must use the word | >processor which has stored the document, or use a processor which | >supports the file format in which the file has been stored. | > | (clip!) | More than being just printed pages, your documents are databases of | information, and SGML helps ensure access to that information. See | http://www.ucc.ie/info/net/markup.html for one clear example of the | value of SGML over proprietary word processor formats. -- ================================================================ Dan Suthers | Learning Research & Development Center suthers+@pitt.edu | University of Pittsburgh (412) 624-7036 voice | 3939 O'Hara Street (412) 624-9149 fax | Pittsburgh, PA 15260 (412) 363-3992 home | http://www.pitt.edu/~suthers/ ================================================================ From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 05 Jun 1995 14:01:42 GMT Organization: Naggum Software; +47 2295 0313 Expires: 19 Jun 1995 14:01:42 GMT Message-ID: <19950605T140143Z@naggum.no> References: <3qt75q$po8@news.globalx.net> \ [Naoto Makinae] | A word processor marks up and stores a document in its proprietary way. | In order to read such a document file, you must use the word processor | which has stored the document, or use a processor which supports the | file format in which the file has been stored. | | Suppose there are one hundred word processors and each has its own mark | up method. A processor has to support ninety nine methods as well as | its own method to read any document which has been written by other | processors. If a new processor emerges, the existing processor has to | implement a new markup method. This sounds costly and troublesome. the problem is that if you wish to sell a quadrillion copies of a product (or any other unimaginably large number), you can afford to do anything, including making reader modules for literally every other format in wide use. the only party to lose in this game is the user, despite his (and the trade press') myopic perception that this kind of "interchangeability" is a plus. one _less_ format would have been a real plus, not one more. since these formats are conceptually _very_ close (as can be seen by the ability to convert between any two of them without much loss of either features or information), it's pathetic that vendors continue to invent incompatible changes to them. | Hence, a standard markup method is desirable. Any document stored in | this format will be read by any word processor (or SGML viewer) no | matter which processor (or SGML editor) has witten the document. this applies to SGML seen as regular text files, but SGML embeds a whole new world of information in a textual representation, instead of a binary form. you don't want users to work directly on the representation, except in some decreasingly common circumstances, but on what is represented, whic is what those other word processors do -- in fact, you don't get access to their internal representation at all, and at best carefully controlled subsets of it. there is no difference between SGML and other environments in the essential aspect that the meaning of the representation must be supplied in order for a document to have any use. it's just that those other word processors have the meaning built in and the details of the representation is largely unknown because it is decided by the word processor programmers instead of by the users who know more about their data and their representatino needs. | SGML (Standard Generalized Markup Language) is such a standard markup | language and is defined in the ISO standards. no, SGML is not such a standard markup language. it is an ufortunate and common misconception that SGML is a "markup language". it isn't. it's a "generalized markup language". in particular, this means that it has no innate application-level semantics, it provides only the syntax that that application can use to represent its data. this is immensely valuable, although it may take some getting used to that one may choose a high-level textual representation that is able to accomodate data of any kind, but this is conceptually no different from the way we choose low-level bits and bytes to represent data of any kind. the bits don't mean anything per se, either, but that hasn't stopped us from assigning any useful meaning to them. ditto with SGML. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: thiel@acsu.buffalo.edu (Kathleen M Thiel) Newsgroups: comp.text.sgml Subject: dtd and sgml Date: 5 Jun 1995 17:06:48 GMT Organization: UB Message-ID: <3qvdj8$4dt@azure.acsu.buffalo.edu> I need to write a parser (I need to parse a dtd looking for specific things) for a dtd. But, I have a problem with the fact that the dtd is not normalized or not always going to be in the form I need. So I thought of sending it through sgml to get a normalized document. But of course it wont let me. I need the dtd to be in the same form as the output from a regular document being sent through sgml. Does anybody know how I can send a dtd through sgml and get this output? Or have a better suggestion? thanks kathy From: ht@chestnut.ling.upenn.edu (Henry S. Thompson) Newsgroups: comp.text.sgml Subject: Re: dtd and sgml Date: 05 Jun 1995 18:26:38 GMT Organization: University of Pennsylvania Message-ID: \ References: <3qvdj8$4dt@azure.acsu.buffalo.edu> In-reply-to: thiel@acsu.buffalo.edu's message of 5 Jun 1995 17:06:48 GMT I have a perl-based dtd-normaliser, released as ftp://ftp.ifi.uio.no/pub/SGML/DEMO/true-dtd.pl It's far from perfect, but it may help you get started. Alternatively, I have a not-in-releasable-state dtd normaliser based on James Clark's SP, which is available for any C++ hackers out there who want to clean up the rough edges and fill in the gaps. ht -- Henry S. Thompson, Human Communication Research Centre, University of Edinburgh Visiting Fellow at Institute for Research in Cognitive Science University of Pennsylvania, 3401 Walnut Street, Philadelphia, PA 19104-6228 Phone: (+1) 215 898-0366 Fax: (+1) 215 573-9247 Email: ht@unagi.cis.upenn.edu Newsgroups: comp.text.sgml From: mlt@netcom.com (Marcy Thompson) Subject: Re: DTD for resume Message-ID: \ Organization: Netcom Online Communications Services (408-241-9760 login: guest) References: \ <3qolos$j1t@news1.delphi.com> Date: Tue, 6 Jun 1995 01:24:04 GMT Sender: mlt@netcom8.netcom.com Jeffrey McArthur writes: >Marcy Thompson had some interresting comments about resumes. > >>Anyway, if I try to write a resume DTD that doesn't itself contain >>a lot of information about how to lay stuff out on the page, then >>I would anticipate that every single person would need to design >>a customized post-process in order to print it. > >One of the things I do besides SGML is develop macros for TeX. For >example, I wrote a set of macros to typeset a resume based on the DTD I >wrote. But I actually did a bit more than that. The TeX macros were >designed in a modular way to allow three separate documents to be produced >from the same source file. The first document would be a full resume in >the order the data falls. A second format is the same information >formatted in a monospaced font (so it can be converted back into ASCII >after using TeX as the formatter). The third format was a 1 page >chronological summary. Right. I would expect that any *reasonable* post process would be able to do something like this. The point I was trying to make is that even if you and I had entirely similar educational and employment histories, along with similar skills, chances are we would not want our resumes to look the same. So your great TeX macros won't do for me. Also, I'm afraid that I consider that your resume DTD *does* have basic what-it-should-look-like information in it (see all that roman | italic stuff). However, It doesn't really have enough that I could specify my personal preference for presenting myself in a resume, and it surely won't work for my friend Tamsen-the-graphic-designer whose resume involves very complex make-it-look-easy layout processing. So if I pushed your DTD to allow you and Tamsen and me all to be able to do what we really want with our resumes, then a single post-process will be super-hard to write properly. Whereas with the other approach (that is, keep the formatting informaiton out altogether), then we really *do* have to write our own post-processes. Now if you work in a place like Passage (me) or Atlis (like Jeffrey), you can probably reasonably expect that a large number of your coworkers are capable of writing such a post-process. But why should they? Maybe there's a reason, but I haven't thought of one yet. Marcy -- Marcy Thompson work: marcy@passage.com play: marcy@world.std.com mlt@netcom.com From: erabena@eraj.ericsson.se (Naoufel Ben Cheikh) Newsgroups: comp.text.sgml Subject: Japanese SGML editor Date: 6 Jun 1995 11:00:33 GMT Organization: Ericsson Radio Systems - Mobile Telephone Systems PDC Message-ID: <3r1cgi$dns@erinews.ericsson.se> Reply-To: erabena@eraj.ericsson.se Hi, Anyone who knows something about the market situation for SGML editors in Japanese: we are interested in translating our existing documents from english SGML to the japanese equivalent. Any information is welcome. Regards/ Naoufel Ben Cheikh From: Steven Haber <72027.263@compuserve.com> Newsgroups: comp.text.sgml Subject: Re: Word Tables to ASCII text Date: 6 Jun 1995 11:58:31 GMT Organization: CompuServe Incorporated Message-ID: <3r1ft7$195@dub-news-svc-4.compuserve.com> References: \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) Have you considered using the generic terminal type print driver in Windows and printing to a file? Not elegant, but workable.tactics@atlantis.actrix.gen.nz (Tactics Wellington) wrote: >Does anybody know if there are any products out there that will convert >tables in Word for Windows documents and convert them to be readable and >understandable in ASCII format. > >I need to do this for a proprietory online system that does not handle >formatting at all. Any product that can recognise the tables and mark >them or whatever so that we can write some filters to take that >information and format the tables would also be useful. Newsgroups: comp.text.sgml Approved: erik@naggum.no Message-ID: <9506061732.AA6148@notes.microstar.com> From: Matt Timmermans \ Date: 06 Jun 1995 10:27:50 -0400 Subject: Re: Purpose of SGML in 3 paragraph [Andre J. Emmell] | In what way is an SGML environment better than a well organized Word | Perfect (MsWord, or mainframe) environment? : | Can some help me in formulating such a simple and convincing answer? Assuming you can do everything you want in a WP environment, then the answer is that it's not -- but it will be. In such situations, SGML is for the future. If your information is an investment of time or money, then SGML is a way to protect that investment by ensuring you'll still be able to read it when all your tools are obsolete, or when you find a _new_ use for it. \ -- Matt Timmermans | Phone: +1 613 727-5696 Microstar Software Ltd. | Fax: +1 613 727-9491 34 Colonnade Rd. North | BBS: +1 613 727-5272 Nepean Ontario CANADA K2E-7J6 | E-mail: mtimmerm@microstar.com From: steve@poota.dungeon.com (Steven Burrows) Newsgroups: comp.text.sgml Subject: Re: Word Tables to ASCII text Date: Tue, 06 Jun 1995 16:16:31 GMT Organization: Bainton Computing Ltd Message-ID: <3r2358$lod@news.dungeon.com> References: \ Reply-To: steve@poota.dungeon.com X-Newsreader: Forte Free Agent v0.55 tactics@atlantis.actrix.gen.nz (Tactics Wellington) wrote: >Does anybody know if there are any products out there that will convert >tables in Word for Windows documents and convert them to be readable and >understandable in ASCII format. You could try the "Internet Assistant" for word, which converts WFW docs to html, including outputting table data as fixed-pitch ascii. It's crude, but it does work. Available from the Microsoft FTP site ftp.microsoft.com /Steve ======================================================================= OFSOD - The Office For Sending Offensive Despatches - +44 (0)850 919269 Have You Got A Complaint?? Call Us - We Guarantee To Offend Your Target Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: Tue, 6 Jun 1995 16:31:07 -0400 From: bullardc@source.asset.com (Claude L. Bullard) Message-ID: <9506062031.AA24795@source.asset.com> References: <3r1cgi$dns@erinews.ericsson.se> Subject: Re: Japanese SGML editor [Naoufel Ben Cheikh] | Anyone who knows something about the market situation for SGML editors | in Japanese Hi! Call the Intergraph Corporation (US: +1 205 730 2000) and ask for Deane Dayton. Put the same question to him. They have translation software on the market and they handle SGML. Can't say if they can do exactly what you need, but Deane is quite knowledgeable about SGML and these problems. He may have a good answer. Cheers, Len Bullard From: kimber@passage.com (W. Eliot Kimber) Newsgroups: comp.text.sgml Subject: A New Era of Afforable Tools Date: 6 Jun 1995 21:27:57 GMT Organization: Passage Systems, Inc. Message-ID: <3r2h8t$9gn@pubxfer4.news.psi.net> Mime-Version: 1.0 X-Newsreader: WinVN 0.93.11 I attended the recent SGML Europe '95 conference. On display were a number of soon-to-be-released SGML tools, primarily editors, that will, I think, dramatically change the economics (and thus acceptance) of SGML. Both Novell/Wordperfect and Microstar Software, Ltd. showed true SGML editors that will be priced in the 300 dollar US range and that provide functions comparable to or better than the current slate of editors (Adept Edit, Author/Editor, Incontext, Grif) available on Windows platforms. The product that excited me the most was Wordperfect's SGML Edition. While I haven't had time to work with it myself, I spent a long time watching the demo and talking to the product manager. He was able to answer my hard questions to my satisfaction, sufficient to convince me it is very probably really an SGML editor and not crude hack. What excites me most about Wordperfect SGML Edition is not really that particular product but the fact of its existence and what it represents. It will probably be priced as existing Wordperfect products are, which is to say available for upgrade if you already have WP or for competitive upgrade. I just bought WP 6.1 for Windows as a competive upgrade for about 150 dollars US, so using that as a ball-park, this means you can buy a reasonably complete SGML editor *and* print composition system for the price of nice dinner for four. Couple this with a product like SoftQuad Panorama PRO, which at low volumes will be about 150 dollars US. With Panorama, you have a pretty complete online publishing system. And considering that EBT, the makers of DynaText, must certainly be crafting their answer to Panorama, there will be several free or low-priced SGML browser/retrieval systems to choose from. This means that for an investment of 300 dollars US you can put together a complete personal SGML print and online publishing system. Couple this with existing free tools like Rexx (if you are Warped[tm] or using PC-DOS 7), Perl, SP and NSGMLS, the various Rainbow converters, and whatever else you might find, and you can build a lot of stuff for very low cost. This means that useful SGML tools are well within the reach of anyone who cares to buy them (and isn't so militantly committed to free software that they're not even willing to pony up a few scores of dollars--there's always emacs and PSGMLS, which is a darn good system by itself, if not necessarily appealing to your average Windows user). The Microstar product, which is a Word for Windows[tm] add-on has similar qualities, and is, for the same reason, very exciting. As a consultant, this is particularly exciting because now I have two viable options for customers who want to use SGML but already have lots of Wordperfect or Word around and are not keen to toss it out for Adept Edit or Author/Editor. This makes SGML, and particularly, native SGML editing, much more attractive, which is always good for SGML consultants. Frame also showed an upgrade to their FrameBuilder product that by all accounts fixes a number of problems that made it not quite a true SGML editor (I didn't have time to look at it, so I can't comment on how true this is, but I trust the people who told they thought so). However, unless they change their pricing policy, it doesn't count as an Everyperson SGML editor. However, it will be attractive to those customers who are already using a Frame-class DTP system and want to move to SGML. Let's hope they at least follow the Novell trend and price it similarly to their base FrameMaker product. This development should also cause the existing SGML editor vendors to take a hard think about what distinguishes their products from these challengers. Given what I've seen so far, I would be hard pressed to find a way that any of the existing Windows editors offer any significant competitive advantage over the Wordperfect product, with the exceptions of Incontext, which is unique in being able to edit documents with element declarations in their internal DTD subsets, and Grif, which has uniquely powerful composition and editing features (Grif is hampered, however, by high up-front training and customization costs and weak North American distribution). Certainly the historical mainstays, Author/Editor and Adept Editor, have a tough row to how at this point. Synex AG (the developers of SoftQuad Panorama and SoftQuad Explorer) also announced an SGML toolkit (of which Panorama is an application). This tool, along with Michel Vulpe's recent SGML Application Server announcement, means that you have access the same sort of generalized integration technology that database folks have taken for granted for some time now. The price is still a bit high (I'm holding out for a 300 dollar Visual Basic VBX package), but certainly within reason. I think all of these developments point to a time in the not at all distant future when we can expect SGML tools to be much more complete, competetive, and affordable. I'm looking forward to it. Speaking personally, the Wordperfect SGML Edition should finally provide me with my Holy Grail, a sufficiently complete way to print from SGML from my Windows laptop without having to write a lot of code (e.g., writing TeX macros was out, even though I new it could be done in theory). I can't wait. -- \

W. Eliot Kimber (kimber@passage.com) Systems Analyst and HyTime Consultant Passage Systems, Inc., 2608 Pinewood Terr., Austin TX 78757 (512)339-1400 465 Fairchild Dr., Suite 201, Mountain View, CA 94043, (415) 390-0911 "SGML or die" \
 From: Martin Bryan \ Newsgroups: comp.text.sgml Subject: Re: SGML Application Server (Product Announcement) Date: 6 Jun 1995 22:46:03 +0100 Organization: The SGML Centre Sender: news@news.demon.co.uk Message-ID: <802460124snz@sgml-cen.demon.co.uk> References: \ Reply-To: mtbryan@sgml-cen.demon.co.uk X-Newsreader: Demon Internet Simple News v1.29 X-Posting-Host: sgml-cen.demon.co.uk What is the SGML System Declaration for SAS? What character sets does it support? -- Martin Bryan @ The SGML Centre, Churchdown, Glos. GL3 2PU, UK (+44 1452 714029) From: Martin Bryan \ Newsgroups: comp.text.sgml Subject: Re: Application of the DSSSL STTP Date: 6 Jun 1995 22:46:08 +0100 Organization: The SGML Centre Sender: news@news.demon.co.uk Message-ID: <802459800snz@sgml-cen.demon.co.uk> References: <3qf9a0$4en@Mercury.mcs.com> <3qnfnc$cer@murphy.servtech.com> Reply-To: mtbryan@sgml-cen.demon.co.uk X-Newsreader: Demon Internet Simple News v1.29 X-Posting-Host: sgml-cen.demon.co.uk In article <3qnfnc$cer@murphy.servtech.com> amiano@marburg.roc.servtech.com writes: > ..[trimmed for brevity]... > > Now I am curious. A superficial look at DSSSL makes it look interesting from > the standpoint of automated > composition systems, but the closer looks I have taken indicate that it may not > be appropriate for certain uses. For > instance, I noticed a restriction on rotations to 90 degree increments. > The initial release of DSSSL does have restrictions on things like 90 degree rotations of text blocks and limited drawing facilities. This is deliberate. Originally we were trying to do everything but this just got too big to bite off in one chunk. So it was decided to go for an intermediate stage which would provide the functionality of most existing DTP type systems with in-built programmability to allow users to add on those features they cannot live without in their environments (knowing that enhanced documents won't be transportable). Once the first stage has been fully defined, tested and accepted I have been tasked with producing DSSSL-plus sometime in the next century! If you have things that you feel that DSSSL should provide before you would find it useful let me know and I will add it onto our "Wanted Features" list. -- Martin Bryan @ The SGML Centre, Churchdown, Glos. GL3 2PU, UK (+44 1452 714029) British Standards Institute representative on ISO panels for SGML, DSSSL, etc. Newsgroups: comp.text.sgml From: kaikow@standards.com Subject: Re: Word Tables to ASCII text Originator: kaikow@mv.mv.com Message-ID: \ Sender: kaikow@standards.com Reply-To: kaikow@standards.com Organization: MV Communications, Inc. Date: Wed, 7 Jun 1995 00:47:55 GMT References: \ <3r2358$lod@news.dungeon.com> I'm walking in on the middle of this thread, but why not use the options in Word to convert a table to text. It gives you a choice of delimiters. PS: I know that this in Word 6. I forget whether Word 2 has this. From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 7 Jun 1995 05:51:51 GMT Organization: \ Message-ID: <3r3epn$c7i@news1.delphi.com> References: <19950603T204836Z@naggum.no> X-Newsreader: Galahad 1.1b Erik Naggum wrote: >how would you like to see PS, RE, and RS interact? First I would like to forbid the use of RE and RS as short reference tags. I dislike short refs in general. But the examples on page 76-78 of The SGML Handbook give me heartburn. My simple solution would be to bannish RS and RE. Remove then entire concept from SGML. I also dispise shorttag. Rank is worthless. The ISO standards process is slow. Any changes to SGML could follow the same course that HTML is trying to follow: define some practices as depricated. The idea is that some currently used methods will continue to work, but there is an easy way to test your documents to see if they will conform to the new, stricter standard. Take the approach the N. Writh has taken to programming languages: cut out all the needless crap. If in doubt take it out. Then if a case can be made that the construct is really useful, you can always put it back in. ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. Newsgroups: comp.text.sgml From: beweiche@si.bosch.de (Bernhard Weichel) Subject: TeX representation for ISO* entities Message-ID: <1995Jun7.062715.135@fr.bosch.de> Sender: news@fr.bosch.de Organization: Robert Bosch GmbH, Schwieberdingen Date: Wed, 7 Jun 1995 06:27:15 GMT Hi, is there any TeX notation available for the ISO* entities. What I am looking for is a set of files, with statements like this: \\entity{auml}{\\"a} \\entity{lt}{{\\tt <}} or how LaTeX would do it... \\newcommand{ISOauml}{\\"a} \\newcommand{ISOlt}{{\\tt <}} and so on. Pointers are welcome. ========================================================================== Bernhard Weichel Office: [49] (711)811-8322 Robert Bosch GmbH FAX: [49] (711)811-8262 Dept. K3/EES4 P.O. Box 30 02 40 Internet: k3ees4_wl@siiks_a1.bosch.de D-70442 Stuttgart x400: C=DE;ADMD=DBP;PRMD=BOSCH-01; Germany O=BOSCH;G=BERNHARD;S=WEICHEL Newsgroups: comp.text.sgml From: robh@cwi.nl (Rob Hoksbergen) Subject: public domain sgml parser (Mac/Windows/Dos)? Message-ID: \ Keywords: parser, public domain Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam Date: Wed, 7 Jun 1995 07:22:30 GMT Hi, sorry for this question, which must have been asked before, but are there any public domain sgml parsers running on either Macintosh or Windows/Dos? I have been unable to find them. If they exist, do they read all public domain dtd's? In general: where can i find information on this? Thanks in advance. Rob Newsgroups: comp.text.sgml From: robh@cwi.nl (Rob Hoksbergen) Subject: public domain dtd's? Message-ID: \ Keywords: dtd, public domain Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam Date: Wed, 7 Jun 1995 07:26:20 GMT Hi, I am looking for the ISO-standard DTD's. I do have ISO12083:1993 for article, but it calls for several entity sets that i cannot trace. Also, where can i find other public domain dtd's, like the AAP ones (article & maths). Thanks in advance for answering. Rob From: Neil Hurn \ Newsgroups: comp.text.sgml Subject: Re : Omnimark From Exoterica Date: 7 Jun 1995 10:19:28 GMT Organization: UnipalmPIPEX server (post doesn't reflect views of UnipalmPIPEX) Message-ID: <3r3ufg$7kr@soap.pipex.net> I am looking for UK training in this product. Can anyone help. Thanks in advance for answering. From: censign@interserv.com Newsgroups: comp.text.sgml Subject: Re: SGML acceptance Date: 7 Jun 1995 11:29:41 GMT Organization: Logical Design Solutions, Inc. Message-ID: <3r42j5$otq@data.interserv.net> References: <3qpkc0$bas@newsbf02.news.aol.com> X-Newsreader: AIR News 3.X (SPRY, Inc.) Jeff Schmidt (photopres@aol.com) writes: > I'm wondering what the > scope of its acceptance in the marketplace is. My perspective is from > running a small digital composition service for small to medium sized book > publishers. In our discussions with probably 20 to 25 publishers in the > past 3-4 months, only one has mentioned a requirement for SGML. Couple of anecdotes that may help. The SGML Forum of New York started out with a core group of maybe 20 people four years ago. It's mailing list is now over 500, paid membership is over 100 with some dozen or so corporate sponsors, and several meetings this year had over 100 people in attendance. Membership crosses all industry boundaries -- publishing, insurance, financial services, defense, pharmaceutical, telecommunications -- and most of the members are actively involved in some sort of project, although most are probably R\&D. Several presentations to the Forum over the past year have been from publishers using SGML as their data content language. I had the good fortune to interview (and the bad luck to offer a job too late to) someone from another medium-sized book publisher located in the mid-west. She was extremely well versed in SGML because it was what they were doing for several of their clients. I believe that those clients were scientific journals. The University of Chicago Press is also using SGML for many of their journals. Those publishers who are starting to focus on the value of their content instead of the value of their pages are likely to be looking into SGML. It is going to take awhile, but if HTML doesn't eventually drive everyone away in disgust, then we'll start to see some broad-based things happen. /chet -- Chet Ensign censign@interserv.com Logical Design Solutions, Inc. +1 908 771 9221 From: boo@primenet.com (Walter Ian Kaye) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: Wed, 07 Jun 1995 05:29:15 -0700 Organization: Getting organized... a little more... a little more... Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> <3mkm90$1q0@everest.pinn.net> \ \ <3q0673$qnc@frame.frame.com> <3q9i94$eeo_002@huji.ac.il> \ X-Web-Site: http://www.primenet.com/~boo/ X-FTP-Site: ftp://ftp.primenet.com/users/b/boo/ X-Quote: More insomniacs get their news from World News Now than from any other source! X-Newsreader: Value-Added NewsWatcher 2.0b24.3+ In article \, clive@d-supp.demon.co.uk (Clive Bruton/Typonaut) wrote: > But that is the point of pdf, you can have great typography and > illustrations, and they can be searchable an editable. > > And what's more they can be a hell of a lot smaller than gif files. For line art, sure. But what about pixel/halftone/etc graphics -- how does pdf compare with gif in that area? -Walter # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # Walter Ian Kaye: (602) 942-6390 FoxPro/Excel Programmer; Guitarist # # Correspond to: boo@primenet.com, boodlums@genie.com # # BinHex files: boo@primenet.com WWW: http://www.primenet.com/~boo/ # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # From: Mark@scotts.demon.co.uk (Mark Scott) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 7 Jun 1995 14:43:05 +0100 Organization: (none) Sender: news@news.demon.co.uk Message-ID: <3r4ad9$t4t@imp.demon.co.uk> References: <3mgo02$mtp@doc.jmu.edu> <3mkm90$1q0@everest.pinn.net> \ \ <3q0673$qnc@frame.frame.com> <3q9i94$eeo_002@huji.ac.il> \ Reply-To: Mark@scotts.demon.co.uk (Mark Scott) X-Nntp-Posting-Host: scotts.demon.co.uk X-Newsreader: IBM NewsReader/2 v1.01 In \, boo@primenet.com (Walter Ian Kaye) writes: >For line art, sure. But what about pixel/halftone/etc graphics -- how does >pdf compare with gif in that area? pdf uses a variety of compression schemes, the default for pictures is a form of jpeg compression. So it should be as good if not better than GIF. Mark Newsgroups: comp.text.sgml Subject: Re: Japanese SGML editor Message-ID: <1995Jun7.144206.1255@ittpub> From: miked@ittpub.nl (Mike Davis) Date: 7 Jun 95 14:41:58 WET References: <3r1cgi$dns@erinews.ericsson.se> erabena@eraj.ericsson.se (Naoufel Ben Cheikh) wrote: >Hi, >Anyone who knows something about the market situation for SGML editors in >Japanese: we are interested in translating our existing documents from english >SGML to the japanese equivalent. >Any information is welcome. >Regards/ Naoufel Ben Cheikh Perhaps there's a Japanese version of SoftQuad's Author/Editor which'll run under the Japanese version of Windows. Have a look around at http://www.sq.com/ Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: Wed, 7 Jun 1995 11:02:13 -0400 From: "Claude L. Bullard" \ Message-ID: <9506071502.AA31001@source.asset.com> References: <3r2h8t$9gn@pubxfer4.news.psi.net> Subject: Re: A New Era of Afforable Tools Thanks for the review, Eliot. It seems that the early pieces for a complete SGML-centered architecture are starting to appear. Tools that use the SGML to drive the Word or WordPerfect are, IMO, the better of the breed as that satisfies the requirement to use SGML to design the catalog or schema types independent of the styling systems, then enable the local site to tailor the "look". If DSSSL proves adequate, then some reuse of the style descriptions among the media should be possible. Exciting stuff but still not quite there yet. You are quite right that soon, the vendors will have to strive to distinguish themselves which is not so easy in standards-based systems. Expect a fierce competition to create more intuitive user interfaces. This points even more clearly to the integration of SGML tools into complete architectures in which neither the standards or the technology, but the applications themselves are sold. The SGML will be hidden from the user who doesn't need to know more, but available to those who do. Whereas, I am one who always "hacks the SGML", some years of CALS experience makes me sympathize with the writers whose plates are already too full and regard these systems as tools, not vocations. The MID research in its second phase bears out that architectures will separate look, feel, content, and push and pull interactivity. One type will support widget behavior, another will do the stylesheets, content types become schemata and catalogs, and another will define the communications layer. The stylesheets will take on prominence as the content types become more abstract. The fixed mosaic look will become a historical artifact. We still need the communications layers for document databases that provide interoperability among different intelligent object notations (hmm.. CGM, SGML, VRML, etc). The SGML communications layer might be based on notation registration and HyTime types and property definitions. Similar to the ODBC for relational systems, this layer would provide standard function types (i.e, message passing) that all multimedia systems (SGML-based, PDF, VRML, etc) would be expected to handle in a distributed client-server environment with appropriate levels of conformance (all do level 1, most do level 2, some do level 3, and so on). Hopefully, the work on SMSL which appears to favor a CORBA/DSOM like solution will move swiftly to provide this for SGML. All of this points to the hard work ahead for WG8 to ensure that all of its standards are aligned to provide for these architectures. The test: if these are done well, to the end user, they do not exist. Len Bullard From: jeffrey@hks.com (Glenda Jeffrey) Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Date: 7 Jun 1995 15:18:47 GMT Organization: Hibbitt, Karlsson & Sorensen, Inc. Message-ID: <3r4g0n$1bs@argo.hks.com> References: <3r2h8t$9gn@pubxfer4.news.psi.net> X-Newsreader: TIN [version 1.2 PL2] W. Eliot Kimber (kimber@passage.com) wrote: : I attended the recent SGML Europe '95 conference. On display were : a number of soon-to-be-released SGML tools, primarily editors, : that will, I think, dramatically change the economics (and thus : acceptance) of SGML. : : \ Were any of these new products capable of dealing with equations in a sensible way? And what about tables? -- Glenda Jeffrey Email: jeffrey@hks.com Hibbitt, Karlsson & Sorensen, Inc Phone: 401-727-4200 1080 Main St. Fax: 401-727-4208 Pawtucket, RI 02860 From: Russell Thomas \ Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Date: 7 Jun 1995 17:53:49 GMT Organization: National Research Council Canada Message-ID: <3r4p3d$kev@nrcnet0.nrc.ca> References: <3r2h8t$9gn@pubxfer4.news.psi.net> <3r4g0n$1bs@argo.hks.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Macintosh; I; 68K) X-URL: news:3r4g0n$1bs@argo.hks.com jeffrey@hks.com (Glenda Jeffrey) wrote: >Were any of these new products capable of dealing with equations >in a sensible way? > >And what about tables? >-- I am not sure about the Wordperfect product but the Microstar product currently supports the Math components as defined in the ISO TR 9573 and I understand that it will be able, in future releases, to support one or more of the other Math fragments (ISO 12083). The Microstar product is also integrated with their Near & Far product so that you can view the DTD at the same time that you are editing the document instance - neat! If you are authoring in Word there is also the TagWizard product which I understand currently supports the ISO 12083 Math component (this is not surprising given that it was developed by Eric van Herwijnen who is the technical editor for the ISO Standard). Russ. ------------------------------------------------------------------------- Dr J Russell Thomas, Phone: (613) 993 - 0817 National Research Council of Canada, Fax: (613) 952 - 4040 I.R.C. - CCC, email: thomas@irc.lan.nrc.ca Building M-24, : russ@erie.irc.nrc.ca Montreal Road, Ottawa, Ontario K1A 0R6 Canada. Newsgroups: comp.text.sgml From: mlt@netcom.com (Marcy Thompson) Subject: Re: Purpose of SGML in 3 paragraph Message-ID: \ Organization: Netcom Online Communications Services (408-241-9760 login: guest) References: <3qt75q$po8@news.globalx.net> Date: Wed, 7 Jun 1995 19:53:29 GMT Sender: mlt@netcom8.netcom.com Andre J. Emmell writes: >In what way is an SGMl environment better than a well organized >Word Perfect (MsWord, or mainframe) environment? Who do you want to own your data? Remember when Word Perfect went from version 5.0 to version 5.1 and the file format changed? If you wanted to use the new features, you had to convert your data, and the conversion was not seamless. With SGML, if someone comes out with a new SGML tool with features you want to use, you can just use the new tool without having to worry about converting the data first. If it would be handy to give a tool with one set of features to your authors and a different tool with a different set of features to your editors, you don't have to hope that you can find a single tool with both sets, because you can buy different tools from different vendors to act on the same data. PostScript is a good case to study. Adobe owns the stnadard and they change it from time to time. So if you invested in expensive tools for printing PostScript documents and then someone sends you a PostScript level 2 document, you now have to spend money to upgrade your tools. Who made that expenditure necessary? Adobe did. The vendors did. I believe that information owners should control their data: its content, its format and its use. If the data is stored in a proprietary format then the format is controlled by some other organization with its own concerns. It will alter the format as it sees fit, without regard for your dependence on it. That's the paranoid theory of information control. And there's a lot of truth to it. SGML is a non-proprietary data format which allows you to haope for data longevity, even after the computers and tools you're using today are gathering dust at the Weird Stuff Warehouse. Well-designed MS Word environments depend on the good will, financial well-being and corporate marketing decisions made in Redmond. Well, that was five paragraphs. Marcy -- Marcy Thompson work: marcy@passage.com play: marcy@world.std.com mlt@netcom.com From: iccutah@cc.usu.edu (ICC Utah) Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Message-ID: <1995Jun7.180452.53602@cc.usu.edu> Date: 7 Jun 95 18:04:52 MDT References: <3r2h8t$9gn@pubxfer4.news.psi.net> <3r4g0n$1bs@argo.hks.com> Organization: Utah State University My news feed does not have the summary of new products written by Eliot Kimber. Would one of you mind e-mailing it to me. Thanks, Chris Chrysostom ICC utah > W. Eliot Kimber (kimber@passage.com) wrote: > : I attended the recent SGML Europe '95 conference. On display were > : a number of soon-to-be-released SGML tools, primarily editors, > : that will, I think, dramatically change the economics (and thus > : acceptance) of SGML. > : > : \ > Newsgroups: comp.text.sgml From: marq@world.std.com (Mark A Lilly) Subject: anyone using LivePage?? Message-ID: \ Organization: The World Public Access UNIX, Brookline, MA X-Newsreader: TIN [version 1.2 PL2] Date: Thu, 8 Jun 1995 00:34:49 GMT I just read about it in Data Based Advisor, and i'm wondering if anyone is using the software, LivePage? Thanks, Mark -- Newsgroups: comp.text.sgml From: papresco@calum.csclub.uwaterloo.ca (Paul Prescod) Subject: Re: public domain sgml parser (Mac/Windows/Dos)? Message-ID: <3r5hak$pof@calum.csclub.uwaterloo.ca> Keywords: parser, public domain Sender: news@undergrad.math.uwaterloo.ca (news spool owner) Organization: University of Waterloo Computer Science Club References: \ Date: Thu, 8 Jun 1995 00:47:16 GMT In article \, Rob Hoksbergen \ wrote: >Hi, >sorry for this question, which must have been asked before, but >are there any public domain sgml parsers running on either Macintosh or > Windows/Dos? http://www.sil.org/sgml/publicSW.html >I have been unable to find them. If they exist, do they read all public > domain dtd's? I would expect them to read all SGML conforming DTDs. SGML information is usually found at: http://www.sil.org/sgml/sgml.html Paul Prescod From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: DTD for resume Date: 8 Jun 1995 02:21:43 GMT Organization: \ Message-ID: <3r5mrn$khv@news1.delphi.com> References: \ X-Newsreader: Galahad 1.1b Marcy Thompson wrote: >So your great TeX macros won't do for me. I was actually amazed at the amount of mail asking me for the macros. So some people like the idea. >... However, It doesn't really have enough that I could specify my >personal preference for presenting myself in a resume, and it surely >won't work for my friend Tamsen-the-graphic-designer whose resume involves >very complex make-it-look-easy layout processing. Graphic designers are a special case. As are TV News anchors. All news anchors include video tapes of themselves as part of the resume; since that is a major part of what they do. Graphic designers play with graphics. That is their job. But what about a run-of-the-mill data entry person? They do not have the skills to create elaborate layouts. If they did, they would be looking for a job other than data entry. I work with a lot of "production staff". They have a wide variety of skills; but most of their experience is centered around data-entry, proofreading, and editing. Those type of skills don't lend themselves to "flashy" resumes. If some one applies for a proofreading job and hands you a resume with a complex layout you have to wonder: did they create the resume themselves? Or did they have someone else do it? If the did it themselves why are they applying for a proofreading job? Or consider the situation of hiring a computer programmer. What kind of resume should they have? Do they really need a flashy resume? Wouldn't code samples be a better way to show off their skills? We recently lost one of our production staff. He decided to become an automotive technician. What kind of resume do you need to apply for a job as an automotive technician? Does the layout of the resume have that much affect for a job like that? There are exceptions. But many jobs do not lend themselves to flashy resumes. All the flash means is the person had someone else do their resume for them. ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: Gordon_Kent@postoffice.brown.edu (Gordon Kent) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: Thu, 08 Jun 1995 08:42:53 -0400 Organization: The Education Alliance @ Brown University Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> <3mkm90$1q0@everest.pinn.net> \ \ <3q0673$qnc@frame.frame.com> <3q9i94$eeo_002@huji.ac.il> \ <3r4ad9$t4t@imp.demon.co.uk> X-Newsreader: Value-Added NewsWatcher 2.0b24.0+ In article <3r4ad9$t4t@imp.demon.co.uk>, Mark@scotts.demon.co.uk (Mark Scott) wrote: > In \, boo@primenet.com (Walter Ian Kaye) writes: > >For line art, sure. But what about pixel/halftone/etc graphics -- how does > >pdf compare with gif in that area? > > pdf uses a variety of compression schemes, the default for pictures is a form > of jpeg compression. So it should be as good if not better than GIF. There's been a lot of discussion about the issue of PDF hogging bandwidth. I recently performed an interesting experiment to exclusively test the relative bandwidth (or file sizes) of PDF files. The purpose of the experiment was to compare web page load size (in k) versus an exact copy of the web page in PDF (in k). Here's what I did. I used Navigator to link to our organization's home page (URL listed in the signature of this email)... a fairly typical web page. 1) Using the "print to file" option in our laserwriter driver, I printed the entire page, including graphics and default font to postscript 2) I distilled the postscript through Acrobat's Distiller using JPEG Medium (default) compression. The images were 72 dpi, so no downsampling of the graphics took place. I didn't embed any fonts. The text was compressed. The result was a two page PDF file, size: 26k. Home page size (in k) index.html: 9k GIF Header: 108k Misc. GIFS: 45k,36k,50k Total:248k Reduction of approx. 9 to 1 The result: A very small, exact copy (minus the default gray Navigator background) of our home page. With the addition of links and URLs in Exchange, I'd imagine the file size would grow by a maximum of 5k. GIFs included (at 72 dpi), text included, formatting exact. Again, this experiment was to compare only relative file sizes. It does not illustrate what PDF offers net publishers. It is not reasonable means of reducing HTML bandwidth (HTML flexibility is something the resulting, semi-static PDF form doesn¹t have). There's no telling (yet) how Navigator will handle the direct read-in of PDF files. Perhaps the download time will be greater (for reason of decompression?) But this initial test seems to suggest that bandwidth of traditionally W3-styled documents will not increase, but decrease. Lets hope people don't take this literally and convert their home pages to PDF to save file size. -- Gordon Kent Media & Technology Coordinator The Education Alliance / Brown University 144 Wayland Avenue Providence RI 02906 401.274.9548 http://www.brown.edu/Research/The-Education_Alliance/ From: chah@jet.uk (Colin Hogben) Newsgroups: comp.text.sgml Subject: Re: Question about using Sp Date: 08 Jun 1995 12:48:51 GMT Organization: Joint European Torus, Abingdon, Oxon, U.K. Message-ID: \ References: <3pdc5c$dnb@csgrad.cs.vt.edu> <19950517T190555Z.enag@naggum.no> In-reply-to: erik@naggum.no's message of 3 Jun 95 19:08:41 GMT In article <19950517T190555Z.enag@naggum.no> erik@naggum.no (Erik Naggum) writes: > [Fred L. Drake, Jr.] > > | Named entities are present in the document instance, and are being > | replaced in the output by the replacement text from the entity > | declaration. > | > | My intention is to be able to intercept references for named characters > | and replace them with the equivelent ANSEL codes or an acceptable > | transliteration. Is this a reasonable approach? > > have you tried changing the entity declarations? remember, if they weren't > intended to be changed to suit your needs, there wouldn't be any point in > having them. just because the ISO standard contains junk definitions that > were only intended to provide a recommended set of names, doesn't mean you > are forced to use them. Forgive my newbie ignorance, but it seems uncomfortable to me that I should need to re-parse the document instance in a different environment, i.e. different sets of entity declarations, depending on what application is going to use the parsed output. (I'm using sgmls and then perl scripts on the output.) In the case of named characters, defining a replacement text for some applications would mean that essentially out-of-band data is converted into in-band data, which can lose information. E.g. suppose I want to convert a document instance into LaTeX. If I do \ etc. then "flamb\é" gets converted to "flamb\\'e" in the parsed output; OK. But if the original document instance also contains "\\" characters in its own right, then the application (back end) cannot distinguish between these and the replacement character texts. The application needs to transform any characters problematical to the LaTeX output (e.g. "\\" and "%"), but it can't tell that the eacute has already been escaped. Waiting to be flamb\éed... :-) -- Colin Hogben =============================================================================== The above article is the personal view of the poster and should not be considered as an official comment from the JET Joint Undertaking =============================================================================== From: stickler@iag.net (Patrick Stickler) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Followup-To: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Date: 8 Jun 1995 13:53:17 GMT Organization: Internet Access Group, Florida's Internet Provider (407)786-1145 Message-ID: <3r6vcd$jem@news.iag.net> References: <3mgo02$mtp@doc.jmu.edu> <3mkm90$1q0@everest.pinn.net> \ \ <3q0673$qnc@frame.frame.com> <3q9i94$eeo_002@huji.ac.il> \ <3r4ad9$t4t@imp.demon.co.uk> \ X-Newsreader: TIN [version 1.2 PL2] In article \, Gordon Kent \ wrote: >... >The result was a two page PDF file, size: 26k. >... >[HTML Formats] >Total:248k > >Reduction of approx. 9 to 1 Your test compares apples to oranges. Nearly all current Web technology does not use compression, so of course any format that does will almost certainly use less bandwidth. You should have compressed the GIF images and HTML instances and compared those sizes -- presuming that, if compression is determined in the future to be beneficial for the Web, it will be used pretty much everywhere such benefit is to be gained. Comparing a compressed format to a non-compressed format is meaningless. Also, as most PDF and Netscape Extensions proponents constantly point out, HTML formatting and page layout is simple, even boring. Why don't you take a sample PDF file representing what *you* would publish on the Web, with 10+ fonts, sidebars, graphic headings, footnotes, etc., etc. etc. and consider the size of that document. Would a user even want to download such a document without compression, and even when compressed...? What is the size of your test PDF document without compression (presuming compression is not a manditory feature of PDF). In any case, even *if* there really is a 9 to 1 bandwidth reduction in favor of PDF over current Web formats (which I doubt), the loss of meaningful/useful markup by going with PDF would not IMO be worth many times the saved bandwidth. As with computer memory debates years back, technology will increase, and what seems obese today will be normal or even tiny tomorrow (cf. full video on demand via fiber optic). I agree that there is certainly a place for pretty much any useful format on the Web, as many uses might not be for on-screen consumption (e.g. printing out tax forms in PDF or Postscript format to fill out and mail, etc.), but this does not mean that any format is suitable as the primary vehicle for Web navigation, regardless of its other virtues. HTML is for information navigation and dissemination on disparate and even presently unknown platforms, *not* for page layout, publishing, traditional advertising, etc.... Yes, HTML needs more layout functionality (even perhaps being replaced by SGML proper and DSSSL Lite style sheets), but people wanting to do full page layout and have total control over the representation of HTML documents will *always* be dissapointed because of the constraints necessary for a flexible world-wide and platform independent medium. =========================================================================== Patrick Stickler KC4YYY Email: stickler@iag.net Senior Information Systems Engineer Phone: (407) 356-9852 Office Lockheed Martin Automated Logistic Systems 356-6094 Lab MP1270, 12506 Lake Underhill Rd. Fax: (407) 826-6261 Orlando, Florida 32825 U.S.A. 356-8949 --------------------------------------------------------------------------- Don't put off for tomorrow what you can do today; because if you enjoy it today, you can do it again tomorrow... =========================================================================== Newsgroups: comp.text.sgml From: jjr@panini.att.com (Jeff Rocca) Subject: An enhancement to NSGMLS Message-ID: \ Sender: news@nntpa.cb.att.com (Netnews Administration) Organization: AT\&T Bell Laboratories Date: Thu, 8 Jun 1995 15:41:42 GMT I am using version 0.4 of James Clark's SGML parser, NSGMLS, to build an application. I need to enhance NSGMLS and add an option to specify that entities should not be expanded in the output. Has anyone already made this enhancement? Any advice/comments? Jeff Rocca jjr@pez.att.com From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: Question about using Sp Date: 8 Jun 1995 09:14:48 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3r77lo$eb6@crl7.crl.com> References: <3pdc5c$dnb@csgrad.cs.vt.edu> <19950517T190555Z.enag@naggum.no> \ Colin Hogben \ wrote: >In the case of named characters, defining a replacement text for some >applications would mean that essentially out-of-band data is converted >into in-band data, which can lose information. > >E.g. suppose I want to convert a document instance into LaTeX. If I do > \ >etc. then "flamb\é" gets converted to "flamb\\'e" in the parsed >output; OK. But if the original document instance also contains "\\" >characters in its own right, then the application (back end) cannot >distinguish between these and the replacement character texts. The >application needs to transform any characters problematical to the >LaTeX output (e.g. "\\" and "%"), but it can't tell that the eacute has >already been escaped. You should declare the entity as: \ \ ESIS output distinguishes character data from system data, so your application could pass the SDATA parts through unchanged and only escape the CDATA parts. Or, if you want to convert the same parsed output into multiple formats, you might as well stick with the ISO definitions \ and have the application do the translation. This is more work for the application, of course, but it might be necessary anyway: For example, when converting to LaTeX, a single entity might need to have different replacement texts depending on the current output state: \ is appropriate in math mode, but if '\α' appears in body text it should be replaced with \ --Joe English jenglish@crl.com From: pdavis@copernicus.bbn.com (Peter Davis) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 8 Jun 1995 17:37:03 GMT Organization: Bolt, Beranek and Newman Inc. Message-ID: <3r7cfv$ekk@info-server.bbn.com> References: <3mgo02$mtp@doc.jmu.edu> <3mkm90$1q0@everest.pinn.net> \ \ <3q0673$qnc@frame.frame.com> <3q9i94$eeo_002@huji.ac.il> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3r6vcd$jem@news.iag.net> Reply-To: pdavis@bbn.com X-Newsreader: TIN [version 1.2 PL2] Patrick Stickler (stickler@iag.net) wrote: : In article \, : Gordon Kent \ wrote: : >... : >The result was a two page PDF file, size: 26k. : >... : >[HTML Formats] : >Total:248k : > : >Reduction of approx. 9 to 1 : Your test compares apples to oranges. Nearly all current Web technology : does not use compression, so of course any format that does will almost : certainly use less bandwidth. You should have compressed the GIF images : and HTML instances and compared those sizes -- presuming that, if : compression is determined in the future to be beneficial for the Web, : it will be used pretty much everywhere such benefit is to be gained. : Comparing a compressed format to a non-compressed format is meaningless. GIF images are already compressed via LZW compression. If the PDF in question was using JPEG compression, that may account for some of the difference in size. JPEG is capable of much higher compression ratios than GIF, but at the expense of image quality. : Also, as most PDF and Netscape Extensions proponents constantly point : out, HTML formatting and page layout is simple, even boring. Why : don't you take a sample PDF file representing what *you* would publish : on the Web, with 10+ fonts, sidebars, graphic headings, footnotes, : etc., etc. etc. and consider the size of that document. Would a user : even want to download such a document without compression, and even : when compressed...? What is the size of your test PDF document : without compression (presuming compression is not a manditory feature : of PDF). I don't understand your point. It sounds as if you're saying: "OK, so PDF does everything HTML does more compactly, but how compact is it for things that HTML can't do?" If these are the only two choices, then you HAVE to use PDF for stuff that HTML can't do. -pd -- Peter Davis "Education is not the 617/873-4145 BBN Educational Technologies filling of a pail, but FAX: 617/873-2455 150 Cambridge Park Drive the lighting of a fire." pdavis@bbn.com Cambridge, MA 02140 -- W. B. Yeats URL: http://copernicus.bbn.com/0/people/PDavis/homepage.html From: good@aur.alcatel.com (Charles Good) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 8 Jun 1995 17:57:13 GMT Organization: Alcatel Network Systems, Raleigh, NC. Message-ID: <3r7dlp$6f8@aurns1.aur.alcatel.com> References: <3qt75q$po8@news.globalx.net> Reply-To: good@aur.alcatel.com I think you must avoid comparing SGML with word processing, desktop publishing systems, and other end-user forms of documentation tools. SGML is also not an elegant authoring language. Wordsmiths (writers) and managers of contemporary documentation groups typically think in terms of the limitations and features of their current authoring and production platforms. However, information systems folks think in terms of reducing overhead and open systems structures which advocate ease of access and empowering the user. This is where SGML can really shine! Most documentation folks want to think of SGML as merely a method of eliminating proprietary [authoring/publishing] product overhead and replacing it with ASCII equivalent tags which allow cross-indexing and cross-refencing in a dynamic document management hierarchy. This reduces the burden on support staff, the burden on compute resources, and insures better interoperability in a distributed information product interchange environment (i.e., send documents to different companies that use different platforms and authoring tools). I know people who can argue language theory and document analysis all day long relative to DTD design and conflicting ISO standards. However, wordsmiths and their managers do not care (and often, do not want to know). They primary want results and they like their WYSIWYG platforms. Only the converts from from first generation Wordstar, or the UNIX nroff and troff clans have a real passion for writing with SGML. In most cases, writers want to write and print on their preferred platform and then push a button to generate SGML-compliant files (which they send to someone else to deal with). The more transparent and automatic this process, the better; the more intervention, the more conflict and frustration arises. From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: An enhancement to NSGMLS Date: 8 Jun 1995 11:22:02 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3r7f4a$h4h@crl7.crl.com> References: \ Jeff Rocca \ wrote: >I am using version 0.4 of James Clark's SGML parser, NSGMLS, to build an >application. I need to enhance NSGMLS and add an option to specify that >entities should not be expanded in the output. Has anyone already made >this enhancement? Any advice/comments? That sounds more like "introducing a bug" than "making an enhancement" :-) SGML parsers are *supposed* to expand entities; that's part of their job. In general, it's not possible to generate the element structure without expanding entities first, since entities can contain generic markup (start- and end-tags) that will affect the element hierarchy. Why do you want to do this, BTW? You might want to try using NDATA entities; for those, SGMLS reports the entity reference but doesn't expand them (since they are declared to contain non-SGML data). --Joe English jenglish@crl.com From: stickler@iag.net (Patrick Stickler) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Followup-To: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Date: 8 Jun 1995 19:02:17 GMT Organization: Internet Access Group, Florida's Internet Provider (407)786-1145 Message-ID: <3r7hfp$noq@news.iag.net> References: <3mgo02$mtp@doc.jmu.edu> <3mkm90$1q0@everest.pinn.net> \ \ <3q0673$qnc@frame.frame.com> <3q9i94$eeo_002@huji.ac.il> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3r6vcd$jem@news.iag.net> <3r7cfv$ekk@info-server.bbn.com> X-Newsreader: TIN [version 1.2 PL2] Peter Davis (pdavis@copernicus.bbn.com) wrote: : I don't understand your point. It sounds as if you're saying: "OK, so : PDF does everything HTML does more compactly, but how compact is it : for things that HTML can't do?" If these are the only two choices, : then you HAVE to use PDF for stuff that HTML can't do. PDF does *NOT* do everything HTML can do, and much of what it can do in the area of page layout, HTML *shouldn't* do. What I was saying was that although PDF may have better compression than existing Web formats, that does not make it a better choice than HTML as the navigational vehicle of choice for the Web. If images in GIF format are larger than another format which has better graphics compression at the expense of quality, then one will have to choose which is better for that application. PDF lacks sorely in many ways which make it a far inferior format for *WEB* documents than HTML, and better compression (at the present) does not nullify that fact. =========================================================================== Patrick Stickler KC4YYY Email: stickler@iag.net Senior Information Systems Engineer Phone: (407) 356-9852 Office Lockheed Martin Automated Logistic Systems 356-6094 Lab MP1270, 12506 Lake Underhill Rd. Fax: (407) 826-6261 Orlando, Florida 32825 U.S.A. 356-8949 --------------------------------------------------------------------------- Don't put off for tomorrow what you can do today; because if you enjoy it today, you can do it again tomorrow... =========================================================================== From: suthers+@pitt.edu (Daniel D Suthers) Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Date: 8 Jun 1995 19:11:17 GMT Organization: LRDC, University of Pittsburgh Message-ID: <3r7i0l$ske@usenet.srv.cis.pitt.edu> References: <3r2h8t$9gn@pubxfer4.news.psi.net> Reply-To: suthers+@pitt.edu X-Newsreader: TIN [version 1.2 PL2] [W. Eliot Kimber (kimber@passage.com) 6 Jun 1995 21:27:57 GMT]: | This makes SGML, and | particularly, native SGML editing, much more attractive, which is | always good for SGML consultants. Huh? If SGML-based tools become more attractive, wouldn't that mean less need for consultants? To paraphrase a post on another thread, "SGML: the less you see it, the more successful it is." -- ================================================================ Dan Suthers | Learning Research & Development Center suthers+@pitt.edu | University of Pittsburgh (412) 624-7036 voice | 3939 O'Hara Street (412) 624-9149 fax | Pittsburgh, PA 15260 (412) 363-3992 home | http://www.pitt.edu/~suthers/ ================================================================ From: Mark Thomas \ Newsgroups: comp.text.sgml Subject: Need info on COSE, Davenport, and AAP Date: 8 Jun 1995 19:14:42 GMT Organization: IPC Technologies, Inc. Message-ID: <3r7i72$b31@lucy.infi.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Newsreader: Nuntius 2.0.3_PPC X-XXMessage-ID: \ X-XXDate: Thu, 8 Jun 1995 19:11:44 GMT Hello, I am working for a team that will be using SGML to create online help, user documentation, and policies/procedure manuals for an IS dept of a large company. This company has recently decided to standardize on SGML but has not yet created or decided upon any DTDs. I am looking for more info on DTD creation and/or selection. I have heard of three related SGML initiatives: COSE, Davenport, and AAP. Where can I find out more about these? Also, does this group have a FAQ? I was unable to locate one in the usual places. .------------------------------------------------. | Mark Thomas \ | | Systems Analyst / Network Integration Specialist | | IPC Technologies Inc. http://www.ipctech.com/ | `------------------------------------------------' Newsgroups: comp.text.sgml From: pgb@arbortext.com (PG Bartlett) Subject: Re: A New Era of Affordable Tools Message-ID: <1995Jun8.191959.26284@arbortext.com> Organization: ArborText, Inc. 1000 Victors Way Ann Arbor, Michigan 48108 Date: Thu, 8 Jun 1995 19:19:59 GMT Elliot Kimber writes: Both Novell/Wordperfect and Microstar Software, Ltd. showed true SGML editors that will be priced in the 300 dollar US range and that provide functions comparable to or better than the current slate of editors (Adept Edit, Author/Editor, Incontext, Grif) available on Windows platforms. Long-time readers of this newsgroup know that we do not use it to publicize our products. However, since our $1,350 ADEPT*Editor product, as well as products from SoftQuad, InContext and Grif, have been described as no better than, and possibly worse than, a set of $300 products, the need for accuracy compels us to respond. WordPerfect and its fierce competitor, Microsoft Word, are the market leaders in short document word processing. There is a role for SGML in short documents, and here at ArborText, we applaud their decisions to support SGML. There is a different set of companies who lead the market in long document processing: Frame and Interleaf. And there is a role for SGML in long documents, too. However, there is a third, rapidly growing set of software companies creating products that do not try to support either short document or long document processes. Instead, these products support processes that not only separate content from its presentation, but also liberate content from the boundaries of documents themselves. All of these products share superficial similarities, but thoughtful inspection will expose major differences. It's not fair or rational to compare the apparent perfection of unreleased products, which have not endured close inspection, to the strengths and weaknesses that characterize every real product. And then there is the issue of price vs. value. FrameMaker justifiably costs more than Word because it offers more. Similarly, we charge more for our products, but because they do a lot more, they still represent a bargain to our customers. John Ford Executive Vice President ArborText, Inc. Newsgroups: comp.text.sgml From: abeyg@corel.ca (Abey George) Subject: Looking for Beta Sites for SGML and Corel VENTURA 6.0 Message-ID: \ Sender: nobody@gatekeeper.corel.ca Mime-Version: 1.0 Organization: Corel Corporation Date: Thu, 8 Jun 1995 19:29:21 GMT X-Newsreader: WinVN 0.99.2 Corel VENTURA 6.0 will include SGML functionality. We are currently recruiting knowledgeable sites interested in testing this area of the program. We hope to start the beta cycle for this sometime in September. Interested parties should contact Kelly Greig / Corel's Beta Coordinator - and provide a fax number so we may forward the beta application to you. Kelly Greig CIS: 73364,1466 Internet: kellyg@corel.ca Fax: 613-728-4091 From: "Andre J. Emmell" \ Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 8 Jun 1995 20:07:45 GMT Organization: Global-X-Change Message-ID: <3r7lah$mk@news.globalx.net> References: <3qt75q$po8@news.globalx.net> "Andre J. Emmell" \ wrote: > >(snip)... > > In what way is an SGMl environment better than a well organized > Word Perfect (MsWord, or mainframe) environment? > > >(snip)... Thank you for the responses and the maturity of ideas submitted. I have placed this question at several places. I will post a follow up "simple phrase" in a short while. In the mean time, I am accumulating the responses and if you are interested in the content (as some have requested), I will be glad to send it to you. Cheers, Andre J. Emmell sgl@globalx.net From: andyo@ruby.ora.com (Andy Oram) Newsgroups: comp.text.sgml Subject: Adaptive documentation (different documents for different users)? Date: 8 Jun 1995 17:17:57 -0400 Organization: O'Reilly & Associates, Inc. Message-ID: <3r7pe5$l5d@ruby.ora.com> I'm interested in the possibility of producing adaptive online documentation. That is, if you choose a "novice" button you get one series of links, if you choose "intermediate" you get another, and so on. (Of course, there are many other criteria you could choose in creating different documents.) I've heard of this thing in principle, but I don't know any documents about it. Does anybody on this list have experience with it? One person here is working on such a project. It's tricky to generate different "Next" links for different users, since they move through the document in different ways. Working in HTML, he is solving the problem by dynamically generating new pages through a CGI script. (This is not related to my job at O'Reilly & Associates.) ---------------------------------------------------------------------- Andy Oram O'Reilly & Associates, Inc. andyo@ora.com 90 Sherman Street http://jasper.ora.com/andyo/ Cambridge, MA 02140-3244 phone: (617) 499-7479 USA fax:(617) 661-1116 ---------------------------------------------------------------------- Newsgroups: comp.text.sgml From: waysys@acm.org (William A. Shaffer) Subject: SGML Viewers Organization: Waysys, Inc. Date: Fri, 9 Jun 1995 04:16:41 GMT Message-ID: \ X-Newsreader: Trumpet for Windows [Version 1.0 Rev B] Sender: news@crash.cts.com (news subsystem) Are there any other SGML viewers on the market with capabilities like Dynatext by EBT? From: votetaker@f10.facts.uky.edu (Kenneth Herron) Newsgroups: news.announce.newgroups,news.groups,comp.text,comp.text.desktop,comp.text.sgml,comp.text.tex,comp.text.interleaf,comp.text.frame,comp.publish.cdrom.software,comp.publish.cdrom.multimedia,comp.multimedia Subject: 2nd CFV: comp.publish.electronic.{misc,developer,end-user} Supersedes: <3qaumu$r91@rodan.UU.NET> Followup-To: poster Date: 9 Jun 1995 01:44:14 -0400 Organization: UUNET Technologies Inc, Fairfax VA USA Sender: tale@uunet.uu.net Approved: tale@uunet.uu.net Expires: 19 Jun 1995 00:00:00 GMT Message-ID: \ References: <3qaumu$r91@rodan.UU.NET> Reply-To: voting@f10.facts.uky.edu LAST CALL FOR VOTES (of 2) unmoderated group comp.publish.electronic.misc unmoderated group comp.publish.electronic.developer unmoderated group comp.publish.electronic.end-user Newsgroups line: comp.publish.electronic.misc General electronic publishing issues. comp.publish.electronic.developer Electronic publishing developer tools. comp.publish.electronic.end-user Electronic publishing end-user tools. Votes must be received by 23:59:59 UTC, 18 June 1995. This vote is being conducted by a neutral third party. For voting questions only contact votetaker@f10.facts.uky.edu. For questions about the proposed groups contact Mike Gilpin \. RATIONALE The existing comp.text hierarchy does not include a suitable forum for this fast growing area of publishing. Desktop publishing is only a specific subset of Electronic Publishing, and should remain separate. The current comp.publish.cdrom.* hierarchy typically discusses games software, and suffers from a considerable amount of noise such as "how do I get past this level in Dungeons of Zorph?", or "Which double speed cdrom player should I buy for my ZX81". This proposal exists for serious Electronic Publishing applications and developers. CHARTERS These three groups are intended for those interested in the creation of "serious" publications in electronic form, the tools available, and format issues. comp.publish.electronic.misc comp.publish.electronic.misc is an unmoderated newsgroup created with the purpose of posting general information and discussions items regarding Electronic Publishing issues. Electronic Publishing is taken to mean the publication of documents, images (still and moving), and sound in a digital format for use with computer based viewing applications. comp.publish.electronic.developer comp.publish.electronic.developer is an unmoderated newsgroup created with the purpose of posting information and discussion items relating to tools used to create and develop electronic publishing data collections. This can include discussion on tools used to create the media, as well as tools used to insert and manage hyperlinked data. Current examples are Director, Adobe Distiller, CD-ROM formatters, Authorware, Toolbook, etc. comp.publish.electronic.end-user comp.publish.electronic.end-user is an unmoderated newsgroup created with the purpose of posting information and discussion items relating to applications used by end-users to view collections of electronic publishing data. This typically means viewing software. Current examples are Common Ground, Acrobat, WorldView, etc. Also, this group should be a forum for discussion of end-user products, particularly where they are technologically advanced. For example Microsoft Encarta. However, games discussions should be kept to comp.publish.cdrom.software and discouraged here. HOW TO VOTE Erase everything above the top "-=-=-=-" line and erase everything below the bottom "-=-=-=-" line. Do not erase anything between these lines and do not change the group names. Give your name on the line that asks for it. For each group, place a YES or NO in the brackets next to it to vote for or against it. If you don't want to vote on a particular group, just leave the space blank. Don't worry about spacing of the columns or any quote characters (">") that your reply inserts. Then mail the ballot to: voting@f10.facts.uky.edu Just Replying to this message should work, but check the "To:" line. -=-=-=-=-=- Don't Delete Anything Between These Lines =-=-=-=-=-=-=-=- comp.publish.electronic.{misc,developer,end-user} Ballot \ (Don't remove this marker) Give your real name here-> If you do not give a real name your vote may be rejected. [Your Vote] Group ----------------------------------------------------------------------- [ ] comp.publish.electronic.misc [ ] comp.publish.electronic.developer [ ] comp.publish.electronic.end-user -=-=-=-=-=- Don't Delete Anything Between These Lines =-=-=-=-=-=-=-=- Anything else may be rejected by the automatic vote counting program. The votetaker will respond to your received ballots with a personal acknowledge- ment by mail - if you do not receive one within several days, try again. It's your responsibility to make sure your vote is registered correctly. Only one vote per person, no more than one vote per account. Addresses and votes of all voters will be published in the final voting results list. Bounce List - No need to revote ------------------------------------------------------------------------------ altmann@slspyr Monika Altmann bducharme@riasmtp.riatax.com Bob DuCharme Votes in error ------------------------------------------------------------------------------ hal@enterprise.advance.com Hal Hammond ! No vote statement in message hmiller@rosie.Read.TASC.COM Hal Miller ! No vote statement in message keie@cs.vu.nl EG Keizer ! No ballot tb@HTWM.DE Thomas Beierlein ! No vote statement in message From: Jean-Luc Sanson \ Newsgroups: comp.text.sgml Subject: Re: Adaptive documentation (different documents for different users)? Date: 9 Jun 1995 08:30:25 GMT Organization: Direction des Etudes et Recherches EDF Message-ID: <3r90r1$mf1@edf3.der.edf.fr> References: <3r7pe5$l5d@ruby.ora.com> hello andy, andyo@ruby.ora.com (Andy Oram) wrote: > > I'm interested in the possibility of producing adaptive online > documentation. That is, if you choose a "novice" button you get one > series of links, if you choose "intermediate" you get another, and so > on. (Of course, there are many other criteria you could choose in > creating different documents.) > > I've heard of this thing in principle, but I don't know any documents > about it. Does anybody on this list have experience with it? > It seems to me that the problem you try to solve could be well described using HyTime, and in particular using the Topic, Topic link and Topic Map concepts which are discussed by the CApH group. CApH stands for Convention on the Application of HyTime. This group is sponsored by GCARI. To know more on CApH and its works contact Steve Newcombe (Technoteacher) who chairs the group or Elliot Kimber (Passage Systems) Steve could be reached through email at Hyminder@techno.com and Elliot at kimber@passage.com Jean-Luc Sanson Electricite de France Direction des Etudes et Recherche 1 avenue du General de Gaulle 92141 Clamat France tel : 33-1-47654135 email : jean-luc.sanson@der.edf.fr ---------------------------------------------------------------------- From: mh@batman (Martin Hollender) Newsgroups: comp.text.sgml Subject: wais comp.text.sgml problem Date: 9 Jun 1995 09:09:55 GMT Organization: University of Kassel - Germany Message-ID: <3r9353$hcp@hrz-ws11.hrz.uni-kassel.de> Reply-To: mh@imat.maschinenbau.uni-kassel.de X-Newsreader: TIN [version 1.2 PL1] hello, I would like to access the comp.text.sgml archive via wais With Netscape both wais://ftp.ifi.uio.no/comp.text.sgml and a search via http://wais.wais.com --> comp.text.sgml did not work: A query like "CALS" produced no results (it should, I think) Any ideas what is going wrong? Thank you in advance, martin -- Martin Hollender From: Chris Lada \ Newsgroups: comp.text.sgml Subject: SGMLS for VMS Date: 9 Jun 1995 12:02:28 GMT Organization: Clark Boardman Callaghan Message-ID: <3r9d8k$14q@news.tpp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) Can anyone tell me if there is a version of SGMLS for the VMS operating system. Chris Lada clada@cbc.tpp.com From: jeffrey@hks.com (Glenda Jeffrey) Newsgroups: comp.text.sgml Subject: Re: SGML Viewers Date: 9 Jun 1995 12:23:27 GMT Organization: Hibbitt, Karlsson & Sorensen, Inc. Message-ID: <3r9efv$oic@argo.hks.com> References: \ X-Newsreader: TIN [version 1.2 PL2] William A. Shaffer (waysys@acm.org) wrote: : Are there any other SGML viewers on the market with capabilities like Dynatext : by EBT? Which capabilities do you consider important? There are several SGML viewers (Panorama, Olias, KnowledgeSet). If you like EBT's side by side Document/TOC, I believe Panorama is your only choice. (Somebody please correct me if I'm wrong here; I only saw a quick demo of Panorama at Seybold.) Currently, Panorama runs only on PCs, though they are promising Unix "soon". -- Glenda Jeffrey Email: jeffrey@hks.com Hibbitt, Karlsson & Sorensen, Inc Phone: 401-727-4200 1080 Main St. Fax: 401-727-4208 Pawtucket, RI 02860 Newsgroups: comp.text.sgml From: ers@exoterica.com (Eric R. Skinner) Subject: Re: Re : Omnimark From Exoterica Message-ID: <1995Jun9.143331.18608@exoterica.com> Organization: Exoterica Corporation References: <3r3ufg$7kr@soap.pipex.net> Date: Fri, 9 Jun 1995 14:33:31 GMT In article <3r3ufg$7kr@soap.pipex.net>, Neil Hurn \ wrote: >I am looking for UK training in this product. > >Can anyone help. > >Thanks in advance for answering. Exoterica itself offers training at its offices in Ottawa (Canada) and Paris (France). Courses are also taught at customer facilities around the world. The language of instruction is English. Our resellers may also offer training programmes. For more information on the European scene, contact our Europe office at info@europe.exoterica.com. Regards, -- Eric R. Skinner ers@exoterica.com Exoterica Corporation Tel +1 613 722 1700 Ottawa, Canada Fax +1 613 722 5706 Product information: info@exoterica.com From: mvulpe@i4i.org (Michel Vulpe) Newsgroups: comp.text.sgml Subject: Re: SGML Application Server (Product Announcement) Date: 9 Jun 1995 15:28:42 GMT Organization: Infrastructures for Information Inc. Message-ID: \ References: \ <802460124snz@sgml-cen.demon.co.uk> > What is the SGML System Declaration for SAS? What character sets does > it support? Martin, The SGML Application Server uses the concrete reference syntax by default but it reads standard declarations if they are supplied. Regards, Infrastructures for Information Inc. From: gordo@auntie.bbcnc.org.uk (Gordon Joly) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Followup-To: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Date: 09 Jun 1995 15:41:06 GMT Organization: BBC Networking Club Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> <3mkm90$1q0@everest.pinn.net> \ \ <3q0673$qnc@frame.frame.com> <3q9i94$eeo_002@huji.ac.il> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3r6vcd$jem@news.iag.net> <3r7cfv$ekk@info-server.bbn.com> <3r7hfp$noq@news.iag.net> In-reply-to: stickler@iag.net's message of 8 Jun 1995 19:02:17 GMT Compression? I can set MIME types for all sort of things! file.txt or file.txt.gz other.ps or other.ps.gz index.html or index.html.gz In other worlds all files can be sent GNU Zipped! -- Gordon Joly || gordo@bbcnc.org.uk || +44 181 225 7414 BBC Networking Club, Rm S401, Sulgrave House, Woodger Road, LONDON W12 8QT, UK. "Nation Shall Speak Peace Unto Nation" \BBC Networking Club\ From: tad@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: SGML Application Server (Product Announcement) Date: 9 Jun 1995 15:59:41 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <3r9r5d$7nb@cliffy.lfwc.lockheed.com> References: \ <802460124snz@sgml-cen.demon.co.uk> \ X-Newsreader: TIN [version 1.2 PL1] Michel Vulpe (mvulpe@i4i.org) wrote: : > What is the SGML System Declaration for SAS? What character sets does : > it support? : The SGML Application Server uses the concrete reference syntax by default : but it reads standard declarations if they are supplied. You didn't actually answer the question. Note that >>> SGML System Declaration <<< is a term refering to the limits that a 'system' can handle. This is different from 'SGML Declaration'. a SGML System Declaration is not transmitted with the document, therefore cannot be 'read if supplied'. It is supplied by the system creator to describe the scope of SGML documents that the 'system' can handle. -- --------------------------------------------------------------------- | | Tad McClellan, Logistics Specialist (IETMs and SGML guy) | Lockheed Martin Tactical Aircraft Systems | P.O. Box 748 MZ 1025 | Fort Worth, TX 76101 | | voice: (817) 763-6314 | email: mcclellantj@lfwc.lockheed.com | --------------------------------------------------------------------- | | Money isn't everything ... | | but it's way ahead of whatever's in second place. | --------------------------------------------------------------------- From: nnavarro@gomez.sc.intel.com (Nancy Navarro ~) Newsgroups: comp.text.sgml Subject: SGML-to-Windows online help? Date: 9 Jun 1995 19:17:15 GMT Organization: Intel Architecture Labs, Santa Clara CA Message-ID: <3ra6nr$7pf@bard.sc.intel.com> X-Newsreader: TIN [version 1.2 PL2] Does anybody know of a tool to get from SGML files to Microsoft Windows online help files? The reason is we want to start with SGML source and "down-translate" to multiple formats, among them MS Windows help, others being html, and also to hard-copy. Any clues ANYONE has, your reply will be MOST appreciated! (I have done some preliminary asking around and no one so far knows of anything direct. I am aware that a series of conversions using Bristol Technologies' Hyperhelp might be a rather arduous solution, but was hoping for something more straightforward.) Nancy Navarro Intel Architecture Labs nnavarro@gomez.intel.com Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 9 Jun 1995 16:43:40 -0400 From: "Claude L. Bullard" \ Message-ID: <9506092043.AA37374@source.asset.com> Subject: Blessings and Curses It seems I spend more of my days than I ought explaining the relationship of SGML and HTML to inquiring minds. Take it or leave it, I think HTML to be "Yet Another SGML App". Use it if it works for you. But the weirdness of the questions, the misunderstandings, all serve to make me queasy about this shotgun marriage. I submit the following quote from a piece of strange journalism from "Internetwork: The Manager's Magazine of Standards and Interoperability". The author is Brad Harrison. "Like Bill Joy's NFS and Jeffrey Case's SNMP version 1, HTML does just enough to get the job done - just enough to get the standard implemented and accepted to the point where there is no turning back. We have probably reached the point of no return with HTML. It's here to stay. The concept was never discussed; never debated. Not a government official, not even a representative of industry spent a day in discussion of the matter. Instead, a lone college student and some friends dictated our HTML future. The information highway, despite all the rhetoric and enthusiasm, is nothing more than a collection of aging technologies that are being extended far beyond what they were initially designed to do. HTML certainly shows that you can make an old dog do new tricks - keep in mind that it was born at IBM in the 60's - but the dog remains old and he will eventually die. Not the government, not the academics and certainly not private industry is doing anything to pave the way to enhanced technology that will keep the Information Highway ball rolling. At this rate, the global network will become an illusion, a collection of independent networks somewhat held together by old technology - a technology that was originally designed to help us fight communism." For anyone that wonders why some think the ISO standards, no matter how slow they are to come into being, are better, I say this: The best bread is made from cultivated wheat, not wild grass. Len Bullard From: Ray Chenier \ Newsgroups: comp.text.sgml Subject: Technical Writing student---SGML ?Where should I start??? Date: 9 Jun 1995 20:46:51 GMT Organization: Synapse Internet [Gatineau, Quebec, Canada] Message-ID: <3rabvr$6on@sentinel.synapse.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) Hi: My name is Ray Chenier and I am a student in technical writing. I know SGML is something I should learn about. Any suggestions as to where I should start? I think I should problaby DL some FAQs or something before I think about DL SGML software. No? Any suggestions greatly appreciated. Thanks Ray Chenier (rchenier@writer.synapse.net) 523-4314 Ottawa, Ontario Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 9 Jun 1995 17:41:21 -0400 From: "Claude L. Bullard" \ Message-ID: <9506092141.AA19355@source.asset.com> References: <3qt75q$po8@news.globalx.net> <3r7dlp$6f8@aurns1.aur.alcatel.com> Subject: Re: Purpose of SGML in 3 paragraph [Charles Good] | I know people who can argue language theory and document analysis all | day long relative to DTD design and conflicting ISO standards. | However, wordsmiths and their managers do not care (and often, do not | want to know). They primary want results and they like their WYSIWYG | platforms. Only the converts from from first generation Wordstar, or | the UNIX nroff and troff clans have a real passion for writing with | SGML. In most cases, writers want to write and print on their | preferred platform and then push a button to generate SGML-compliant | files (which they send to someone else to deal with). The more | transparent and automatic this process, the better; the more | intervention, the more conflict and frustration arises. I agree. I express this opinion, however: In many projects, writers have no control over the look and style of their documents; nor should they have any. SGML drives the process. It need not only be its byproduct after printing. Heresy? Stodginess? Old thinking? No: a fact of the contract. We can debate any theory or analysis, the fact remains that for very large long life cycle documents, formal contracting eliminates any creative options. Doubt that? You never did government work. Don't like it? Tough. Get over it. They get what they pay for and they pay OK. But that isn't why. SGML makes it possible to set up the system, tailor the system, and execute the system according to local constraints tailored from global requirements. We can make the constraints invisible, and test that it indeed was done according to the requirements of the Contract Deliverables Requirements List. Try that with RTF of PDF or whatever is being hacked this week. That still isn't why. Try to support very large complex systems. Integrate five databases with five different pieces of software and try not to make a mistake because each vendor swore his format was the best. Try to pick up a manual written ten years ago and reuse the instructions because the original authors are gone or dead and try to not to make a mistake because somewhere on some godforsaken mountainside, an 18 year old is trying to fix a 25 year old piece of hardware and trying not to make a mistake because he's in a hot LZ. Just try. While you are sitting at home arguing about fonts, he is sweating your mistakes. That's why. SGML isn't for memos or for *sexy* user manuals. Good systems and good training save lives. Both depend on accurate information, not creative layouts or way cool web pages. SGML helps make this possible in one of the most complex contracting and procurement systems in the world. If giving up WYSIWYG is the cost, junk it. I've used systems that allow both. They can rock together. As for me, I use SGML for one and only one reason: Results. That is where SGML shines. Len Bullard From: ses@tipper.oit.unc.edu (Simon E Spero) Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Date: 9 Jun 1995 23:04:49 GMT Organization: University of North Carolina, Chapel Hill Message-ID: <3rak2h$17lj@bigblue.oit.unc.edu> References: <3r2h8t$9gn@pubxfer4.news.psi.net> <3r7i0l$ske@usenet.srv.cis.pitt.edu> In article <3r7i0l$ske@usenet.srv.cis.pitt.edu>, Daniel D Suthers \ wrote: > >To paraphrase a post on another thread, "SGML: the less you see it, the more >successful it is." > The most important part about having 'SGML Inside' (tm) is just knowing it's there. There are times when you need to do things that the author of the document never dreamed of. When those moments happen, it's good to know that you can drop out of your text processor, knock off a quick script, and process the document your way, right away, without having to do that wonderful combination of AI and Cryptography you need to do with non-standard texts. What better and cheaper tools mean for the SGML consultant is less time spent training people to do things that should have been simple to do in the first place , and more time doing the fun stuff. Simon -- Contract with America - Explained! |Phone: +44-81-500-3000 Contract: verb |Mail: ses@unc.edu 1) To shrink or reduce in size - the economy contracted +----------------------- 2) To become infected -My baby contracted pneumonia when they stopped my welfare From: melhaven@mail.rcinet.com Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Date: 10 Jun 1995 03:25:58 GMT Organization: Dayton Microcomputer Association; Dayton, Ohio, USA Message-ID: <3rb3c6$1so@sally.dma.org> References: <3r2h8t$9gn@pubxfer4.news.psi.net> <9506071502.AA31001@source.asset.com> Reply-To: melhaven@mail.rcinet.com X-Newsreader: IBM NewsReader/2 v1.09 In <9506071502.AA31001@source.asset.com>, "Claude L. Bullard" \ writes: >Thanks for the review, Eliot. It seems that the early pieces for a >complete SGML-centered architecture are starting to appear. > >Tools that use the SGML to drive the Word or WordPerfect are, IMO, the >better of the breed as that satisfies the requirement to use SGML to design >the catalog or schema types independent of the styling systems, then enable >the local site to tailor the "look". If DSSSL proves adequate, then some >reuse of the style descriptions among the media should be possible. >Exciting stuff but still not quite there yet. > >You are quite right that soon, the vendors will have to strive to >distinguish themselves which is not so easy in standards-based systems. >Expect a fierce competition to create more intuitive user interfaces. This >points even more clearly to the integration of SGML tools into complete >architectures in which neither the standards or the technology, but the >applications themselves are sold. The SGML will be hidden from the user >who doesn't need to know more, but available to those who do. Whereas, I >am one who always "hacks the SGML", some years of CALS experience makes me >sympathize with the writers whose plates are already too full and regard >these systems as tools, not vocations. > >The MID research in its second phase bears out that architectures will >separate look, feel, content, and push and pull interactivity. One type >will support widget behavior, another will do the stylesheets, content >types become schemata and catalogs, and another will define the >communications layer. The stylesheets will take on prominence as the >content types become more abstract. The fixed mosaic look will become a >historical artifact. > >We still need the communications layers for document databases that provide >interoperability among different intelligent object notations (hmm.. CGM, >SGML, VRML, etc). The SGML communications layer might be based on notation >registration and HyTime types and property definitions. Similar to the >ODBC for relational systems, this layer would provide standard function >types (i.e, message passing) that all multimedia systems (SGML-based, PDF, >VRML, etc) would be expected to handle in a distributed client-server >environment with appropriate levels of conformance (all do level 1, most do >level 2, some do level 3, and so on). Hopefully, the work on SMSL which >appears to favor a CORBA/DSOM like solution will move swiftly to provide >this for SGML. > >All of this points to the hard work ahead for WG8 to ensure that all of its >standards are aligned to provide for these architectures. The test: if >these are done well, to the end user, they do not exist. > >Len Bullard - - The use of SGML in it's original form in opinion will be restricted to individual industries and companies. They will develop their unique DTDs for internal use. The public documents will still use SGML, but it will follow the HTML format. - I find the need for DTDs for every document I receive to be a problem I would rather not address. Give me a format, i.e., HTML version 1,2, n or whatever and forget the need for a unique DTD for each document. - One problem I see happening is the version of HTML. The standard is evolving faster than most of us can keep up with. Perhaps we need to setup some basic requirements as how we want to use HTML and stay with a version for a while. The tools will come quicker if we have some stability. Regards, Ken From: melhaven@mail.rcinet.com Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 10 Jun 1995 03:33:23 GMT Organization: Dayton Microcomputer Association; Dayton, Ohio, USA Message-ID: <3rb3q3$1so@sally.dma.org> References: <3qt75q$po8@news.globalx.net> \ Reply-To: melhaven@mail.rcinet.com X-Newsreader: IBM NewsReader/2 v1.09 In \, mlt@netcom.com (Marcy Thompson) writes: >Andre J. Emmell writes: > >>In what way is an SGMl environment better than a well organized >>Word Perfect (MsWord, or mainframe) environment? > >Who do you want to own your data? Remember when Word Perfect went >from version 5.0 to version 5.1 and the file format changed? If you >wanted to use the new features, you had to convert your data, and >the conversion was not seamless. > >With SGML, if someone comes out with a new SGML tool with features >you want to use, you can just use the new tool without having to >worry about converting the data first. If it would be handy to give a >tool with one set of features to your authors and a different tool with >a different set of features to your editors, you don't have to hope >that you can find a single tool with both sets, because you can buy >different tools from different vendors to act on the same data. > >PostScript is a good case to study. Adobe owns the stnadard and they >change it from time to time. So if you invested in expensive tools for >printing PostScript documents and then someone sends you a PostScript >level 2 document, you now have to spend money to upgrade your tools. Who >made that expenditure necessary? Adobe did. The vendors did. > >I believe that information owners should control their data: its content, >its format and its use. If the data is stored in a proprietary format >then the format is controlled by some other organization with its own >concerns. It will alter the format as it sees fit, without regard for >your dependence on it. > >That's the paranoid theory of information control. And there's a lot of >truth to it. SGML is a non-proprietary data format which allows you to >haope for data longevity, even after the computers and tools you're using >today are gathering dust at the Weird Stuff Warehouse. Well-designed >MS Word environments depend on the good will, financial well-being >and corporate marketing decisions made in Redmond. > >Well, that was five paragraphs. > >Marcy >-- >Marcy Thompson >work: marcy@passage.com >play: marcy@world.std.com mlt@netcom.com - - I agree with Marcy. Another issue is the maintenance of documents in the electronic format. If we start keeping our documents in an electronic form, and have to keep them for several years. How will you be able to read them when the documents you have in your records are not compatible with the reader or word processor you are using. For some reason document converters don't seem to do a very good job. A good example is Word for Windows version 2, and version 6. Regards, Ken From: Paul Montgomery \ Newsgroups: comp.text.sgml Subject: SGML parser to run on a Mac? Date: 10 Jun 1995 06:46:00 +0100 Organization: Demon Internet News Service Sender: news@news.demon.co.uk Message-ID: <802733987snz@paulmont.demon.co.uk> Reply-To: facteur@paulmont.demon.co.uk X-Newsreader: Demon Internet Simple News v1.29 X-Posting-Host: paulmont.demon.co.uk I am trying to find an SGML parser to run on a Mac. Any details of commercial, shareware, public domain packages, would be appreciated. We have experience of using SGMLS on a PC platform - is it available for the Mac? Thanks for any help that you can offer. -- Paul Montgomery From: gtn@ebt-inc (Gavin Nicol) Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Date: 10 Jun 1995 06:27:44 GMT Organization: Electronic Book Technologies, Inc. Message-ID: \ References: <3r2h8t$9gn@pubxfer4.news.psi.net> <9506071502.AA31001@source.asset.com> <3rb3c6$1so@sally.dma.org> In-reply-to: melhaven@mail.rcinet.com's message of 10 Jun 1995 03:25:58 GMT >The use of SGML in it's original form in opinion will be restricted >to individual industries and companies. They will develop their >unique DTDs for internal use. The public documents will still use >SGML, but it will follow the HTML format. I very much doubt that this will prove to be the case in the long run. For the first 2 years or so, you may be absolutely correct, but eventually, HTML will be used primarily for publishing home pages and whatnot, and (SGML|RTF|PDF|whatever) will be used for everything else. Large documents with long lifespans will almost certainly be in SGML (or something similar). >I find the need for DTDs for every document I receive to be a problem >I would rather not address. Give me a format, i.e., HTML version >1,2, n or whatever Only because you have to manage them. Once retreival of DTD's and stylesheets, and all other such mundane tasks becomes automatic, I doubt you, as an end-user, could distinguish between HTML, RTF, or SGML. On the other hand, I also doubt that documents, as we use the term today, will be around in 10-15 years, so most of the SGML vs RTF vs HTML vs PDF etc. wars are, in the long run, meaningless. -- Gavin Thomas Nicol | He who will not reason, is a bigot; Electronic Book Technologies | he who cannot is a fool; Email: gtn@ebt.com | and he who dares not is a slave. Phone/FAX: +81-3-3706-7351 | --- Sir William Drummond From: harvey \ Newsgroups: comp.text.sgml Subject: Re: SGML Viewers Date: 10 Jun 1995 15:05:18 GMT Organization: Electronic Commerce Connection Message-ID: <3rccbe$38r@news.cais.com> References: \ <3r9efv$oic@argo.hks.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) jeffrey@hks.com (Glenda Jeffrey) wrote: >William A. Shaffer (waysys@acm.org) wrote: >: Are there any other SGML viewers on the market with capabilities like Dynatext >: by EBT? > >Which capabilities do you consider important? There are several >SGML viewers (Panorama, Olias, KnowledgeSet). If you like EBT's >side by side Document/TOC, I believe Panorama is your only choice. >(Somebody please correct me if I'm wrong here; I only saw a quick >demo of Panorama at Seybold.) Currently, Panorama runs only on PCs, >though they are promising Unix "soon". >-- >Glenda Jeffrey There is also Explorer from Softquad. <><><><><><><><><><><><><><><><><><><><><><><><><><><><><> Betty Harvey | PERSONAL David Taylor Model Basin | harvey@eccnet.com harvey@navysgml.dt.navy.mil | ---------------------------------------------------------- Navy CALS WWW Server: http://navysgml.dt.navy.mil <><><><><><><><><><><><><><><><><><><><><><><><><><><><><> From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 10 Jun 1995 18:26:35 GMT Organization: Naggum Software; +47 2295 0313 Expires: 24 Jun 1995 18:26:34 GMT Message-ID: \ References: <19950603T204836Z@naggum.no> <3r3epn$c7i@news1.delphi.com> [Erik Naggum] | how would you like to see PS, RE, and RS interact? this was actually intended as an invitation to answer a technical question that might provide useful input to the ISO 8879 review process. [Jeffrey McArthur] | First I would like to forbid the use of RE and RS as short reference | tags. _that_ you would like to do so is not relevant. _why_ could have been. | I dislike short refs in general. not an argument. | But the examples on page 76-78 of The SGML Handbook give me heartburn. not an argument. | My simple solution would be to bannish RS and RE. Remove then entire | concept from SGML. the concept is in SGML because it is useful in dealing with text that has been prepared by existing people on existing systems. many would argue that it is precisely the existence of "lines" of input that define a format to be "text". do you wish to remove this property of SGML, too? I don't think I understand your position. if I understood what you have "proposed" so far, I cannot possibly agree with it. | I also dispise shorttag. Rank is worthless. these are not arguments. | The ISO standards process is slow. Any changes to SGML could follow | the same course that HTML is trying to follow: define some practices as | depricated. The idea is that some currently used methods will continue | to work, but there is an easy way to test your documents to see if they | will conform to the new, stricter standard. these are called application conventions, and you can apply them today. | Take the approach the N. Writh has taken to programming languages: cut | out all the needless crap. If in doubt take it out. Then if a case | can be made that the construct is really useful, you can always put it | back in. this reflects the actual ISO process. unlike what you appear to think, ISO committees do not bloat standards for fun. there have usually been people who argue fiercely against any addition, as well as those to argue fiercely against any deletion. additions _are_ more easily accepted than deletions, however. putting things back in doesn't actually work the way you think. the only way to make sure that these things are addressed is to make a solid proposal that will address the issues that gave rise to the things you don't like that are there today, or show that these issues are not worth addressing. your article is useless in both respects. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: public domain sgml parser (Mac/Windows/Dos)? Date: 10 Jun 1995 18:36:35 GMT Organization: Naggum Software; +47 2295 0313 Expires: 24 Jun 1995 18:36:35 GMT Message-ID: \ References: \ [Rob Hoksbergen] | sorry for this question, which must have been asked before, but are | there any public domain sgml parsers running on either Macintosh or | Windows/Dos? | | I have been unable to find them. If they exist, do they read all public | domain dtd's? | | In general: where can i find information on this? the SGML Repository at ftp.ifi.uio.no:/pub/SGML is worth visiting. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: public domain dtd's? Date: 10 Jun 1995 18:37:21 GMT Organization: Naggum Software; +47 2295 0313 Expires: 24 Jun 1995 18:37:21 GMT Message-ID: \ References: \ [Rob Hoksbergen] | I am looking for the ISO-standard DTD's. I do have ISO12083:1993 for | article, but it calls for several entity sets that i cannot trace. | | Also, where can i find other public domain dtd's, like the AAP ones | (article & maths). the ISO entity sets are in ftp.ifi.uio.no:/pub/SGML/ENTITIES. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: Question about using Sp Date: 10 Jun 1995 19:04:46 GMT Organization: Naggum Software; +47 2295 0313 Expires: 24 Jun 1995 19:04:44 GMT Message-ID: \ References: <3pdc5c$dnb@csgrad.cs.vt.edu> <19950517T190555Z.enag@naggum.no> \ [Colin Hogben] | Forgive my newbie ignorance, but it seems uncomfortable to me that I | should need to re-parse the document instance in a different | environment, i.e., different sets of entity declarations, depending on | what application is going to use the parsed output. (I'm using sgmls | and then perl scripts on the output.) you're going to have to reparse _something_. why not stay as close to the original source as possible? why invent yet another intermediate format? the format that is parsed for free does not exist. | In the case of named characters, defining a replacement text for some | applications would mean that essentially out-of-band data is converted | into in-band data, which can lose information. if it can lose information, don't do it. | E.g. suppose I want to convert a document instance into LaTeX. If I do | \ | etc. then "flamb\é" gets converted to "flamb\\'e" in the parsed | output; OK. But if the original document instance also contains "\\" | characters in its own right, then the application (back end) cannot | distinguish between these and the replacement character texts. if so, it is silly to confuse it. instead, use the following construct. \ the idea with SDATA is that it represents system-specific data. if your document contains backslashes, you will need to process them specially, as they are certainly "data" and not "TeX commands". your \\'e is a TeX command, i.e., a different beast altogether. if you wish to confuse the two, it is not SGML's problem. a system that doesn't give you the rope to hang yourself isn't worth using. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Date: 10 Jun 1995 20:02:38 GMT Organization: Naggum Software; +47 2295 0313 Expires: 24 Jun 1995 20:02:35 GMT Message-ID: <19950610T200238Z@naggum.no> References: <3r2h8t$9gn@pubxfer4.news.psi.net> <3r4g0n$1bs@argo.hks.com> <1995Jun7.180452.53602@cc.usu.edu> [Chris Chrysostom] | My news feed does not have the summary of new products written by Eliot | Kimber. Would one of you mind e-mailing it to me. this seems to be a good occasion to remind people of the comp.text.sgml archive at ftp.ifi.uio.no in /pub/SGML/comp.text.sgml, sorted by date and author. e.g., if you use a reasonably user-friendly FTP client, you can go to the archive, and list the files that match the pattern: 1995/199506*/*.Kimber and come up with this one file: 1995/19950606/212757.Kimber the file can then be retrieved with whatever file transfer tools are available. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: stevmorgan@aol.com (StevMorgan) Newsgroups: comp.text.sgml Subject: Re: sqml2asc? Is there such a biest? Date: 10 Jun 1995 16:57:27 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3rd0vn$ms1@newsbf02.news.aol.com> References: \ Reply-To: stevmorgan@aol.com (StevMorgan) I have been working with Tools and Techniques of Austin, TX. on a particular SGML to just about anything conversion, ODBC, Asci, and around 30 other types of files. My application was big enough that they built the specific conversion for me, however, there will be a kit available to map your own tag structure pretty simply for a conversion. From: stevmorgan@aol.com (StevMorgan) Newsgroups: comp.text.sgml Subject: Re: Looking for Beta Sites for SGML and Corel VENTURA 6.0 Date: 10 Jun 1995 18:03:05 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3rd4qp$nmn@newsbf02.news.aol.com> References: \ Reply-To: stevmorgan@aol.com (StevMorgan) Kelly, I specifically purchased Ventura 5 for the upgrade to Ver. 6 with SGML from InContext. I daily download reams of SGML and can routinely convert SGML to Ascii, ODBC and about 30 other formats. What I need however is to publish the information in an excellent DTP app. I am aware of Ventura 6's ambitions, I also run PageMaker 5.0 Win, and QuarkXpress 3.31r5 and have inquired about SoftQuad's SGML enabler for Quark. I am looking at FrameMaker and Folio products as well as LivePage by Inforium. This is being posted thru AOL's new ver. 2.5 which has the best web browser I have seen, but I do not use AOL on a regular basis, so please reply to my Internet address which is: StevenL@erinet.com My fax number is (513) 426-2801. My business numbers are (513) 426-7321 and (513) 477-8450. From: jang@xs4all.nl (Jan Grootenhuis) Newsgroups: comp.text.sgml Subject: Re: Question about using Sp Date: 11 Jun 1995 12:28:59 GMT Organization: XS4ALL, networking for the masses Message-ID: <3renib$ia2@news.xs4all.nl> References: <3pdc5c$dnb@csgrad.cs.vt.edu> <19950517T190555Z.enag@naggum.no> \ <3r77lo$eb6@crl7.crl.com> X-Newsreader: NN version 6.5.0 #666 (NOV) X-Posted-By: jang@xs1.xs4all.nl jenglish@crl.com (Joe English) writes: >[...] For example, when converting to LaTeX, a single entity >might need to have different replacement texts depending on the >current output state: > \ >is appropriate in math mode, but if '\α' appears in >body text it should be replaced with > \ Maybe, in body text the 'agr' from Greek alphabetic characters ISOgrk1 should be used rather than the Greek symbol alpha from ISOgrk3? With USEMAP, it is rather straightforward to achieve different translations like this at the SGML end. Cheers, Jan From: haitto@synex.se (Hasse Haitto) Newsgroups: comp.text.sgml Subject: Re: SGML Viewers Date: 11 Jun 1995 20:00:06 GMT Organization: Synex Information AB Sender: haitto@nic.pi.se Message-ID: <3rfi06$pur@news.pi.se> References: \ Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Posted-From: InterNews 1.0.5@klkmac046.nada.kth.se X-Authenticated: haitto on POP host nic.pi.se In article \ waysys@acm.org (William A. Shaffer) writes: > Are there any other SGML viewers on the market[...]? To read more about SGML browsing technology that lets you go directly from SGML source to viewing - without an intermediate, proprietary, binary format - check out the URLs http://www.pi.se/synex/viewport.html and http://www.sq.com/products/panorama/panor-fe.htm _______________________________________________________________________ Hasse Haitto haitto@synex.se Synex Information AB Fax: + 46 8 751 5907 Kallforsv. 24, 124 32 Bandhagen SWEDEN ----------------------------------------------------------------------- Newsgroups: comp.text.sgml Message-ID: <2005@leadingedg.win.net> Reply-To: mgraves@leadingedg.win.net (Michael Graves) From: mgraves@leadingedg.win.net (Michael Graves) Date: Mon, 12 Jun 1995 03:34:14 GMT Subject: Compilers? Having made a significant effort to build a WWW site I would also like to use it as an off-line electronic publication. To that end, does anyone know of an HTML "compiler" that would allow me to distribute an executable containing the pages in question? Michael Graves From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: Blessings and Curses Date: 12 Jun 1995 05:08:46 GMT Organization: \ Message-ID: <3rgi4u$c9m@news1.delphi.com> References: <9506092043.AA37374@source.asset.com> X-Newsreader: Galahad 1.1b Claude L. Bullard wrote: >"Like Bill Joy's NFS and Jeffrey Case's SNMP version 1, HTML does just >enough to get the job done - just enough to get the standard implemented >and accepted to the point where there is no turning back. We have probably >reached the point of no return with HTML. It's here to stay. The concept >was never discussed; never debated. Not a government official, not even a >representative of industry spent a day in discussion of the matter. >Instead, a lone college student and some friends dictated our HTML future. I like that. Just like two guys in a garage changed the way the world thought of computers when they invented the Apple II. The Apple II had some serious flaws. But it did inspire others. Where would we be today without people who get up and do something rather than wait for government officials and industry representatives to do the work? ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 12 Jun 1995 05:08:52 GMT Organization: \ Message-ID: <3rgi54$c9m@news1.delphi.com> References: <9506092141.AA19355@source.asset.com> X-Newsreader: Galahad 1.1b Claude L. Bullard wrote: >In many projects, writers have no control over the look and style of their >documents; nor should they have any. SGML drives the process. It need not >only be its byproduct after printing. SGML does not drive this process. What drives this process is the publisher and/or copyeditor. They are the ones who care about the format. It does not matter what the author says, the publisher is the one paying the bills. If the author says "make this bold", but the publisher says, "make it italic" want to guess who wins? The author can make suggestions on what the document should look like. But they are only suggestions. Unless you want to go to a vanity press, or self publish, the author does not have the final say. Authors tend to forget that their work may be only a small peice of a much larger body of work. Most publishers want to have a consistant style. They want all their books and articles to look somewhat similar. They enforce this in printed books. Electronic documents should also follow a similar style. ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: kam@physics.ox.ac.uk (Kaylene Murdoch) Newsgroups: comp.text.sgml Subject: Re: Microstar location needed. Help Date: 12 Jun 1995 10:46:53 GMT Organization: Starlink Message-ID: <3rh5ut$97b@news.ox.ac.uk> References: \ X-Newsreader: TIN [version 1.2 PL2] Tim Wayne (twayne@julian.uwo.ca) wrote: : I would like to know how to contact Microstar, a company involved with various : SGML utilities located in Ottawa Ontario. An e-mail address or URL would be : great. On my 'Near and Far' manual I have: Microstar Software Ltd 34 Colonnade Road North Ontario Canada K2E 7J6 Tel: (613) 727 5696 Fax: (613) 727 9491 Internet: cade@microstar.com Cheers Kaylene murdochk@oup.co.uk Newsgroups: comp.text.sgml From: meh@world.std.com (M. E. Hunter) Subject: Print vs. E-Doc Message-ID: \ Organization: Interconsult, Inc. Date: Mon, 12 Jun 1995 13:38:31 GMT Can anyone tell me if any companies have started charging for printed documentation, while providing the electronic versions free of charge? For instance, is there anywhere that will give you the CD-ROM/disks/download but charges if you want them to send you the book? Please reply via email (meh@world.std.com) Thanks for your help. -- M. Elizabeth Hunter v: 617-354-0400 Assistant to the President f: 617-354-0940 InterConsult, Inc. e: meh@world.std.com The Brickyard, 84 Sherman Street, Cambridge, MA 02140 From: digitome@iol.ie (Digitome Ltd) Newsgroups: comp.text.sgml Subject: Irish SGML Users Group Date: 12 Jun 1995 14:42:04 +0100 Organization: The Pipeline Message-ID: <3rhg7c$c4p@wilde.iol.ie> An SGML Users Group has been set up in Ireland. First meeting is scheduled for September. We are assembling a list of people who may be interested in receiving notification about the inaugural meeting. For more information contact :- Sean Mc Grath Paul Mc Keon Digitome Ltd. Professional Accounting Software Ltd. digitome@iol.ie pasltd@iol.ie Tel : +353 96 72092 +353 1 6613742 Fax : +353 96 72093 +353 1 6621056 From: mckee@smiley.mitre.org (Hugh McKee) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 12 Jun 1995 13:59:17 GMT Organization: The MITRE Corporation, McLean, VA Message-ID: <3rhh7l$2ru@reuters2.mitre.org> >From: suthers+@pitt.edu (Daniel D Suthers) > >Your argument will be stronger if you can identify a need >to treat your documents as a database -- e.g., to select portions in response >to a query, automatically generate specialized documents, etc. For example, >suppose your company suddenly needed to make portions of the documents >available on the internet for use by customers or service personnel at other >sites. BUT, there is an essential preceding step, the writing of the DTD. If the authors of the DTD did not know that portions of the documents would be used by customers or service personnel then I think there will be difficulties. As I understand it (probably imperfectly), to revise a DTD, and reprocess all the documents, can be a daunting task. From: claird@Starbase.NeoSoft.COM (Cameron Laird) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 12 Jun 1995 10:34:10 -0500 Organization: NeoSoft Internet Services +1 713 968 5800 Message-ID: <3rhmpi$no@Starbase.NeoSoft.COM> References: <3qt75q$po8@news.globalx.net> <3r7dlp$6f8@aurns1.aur.alcatel.com> <9506092141.AA19355@source.asset.com> In article <9506092141.AA19355@source.asset.com>, Claude L. Bullard \ wrote: . . . >Try to support very large complex systems. Integrate five databases with >five different pieces of software and try not to make a mistake because >each vendor swore his format was the best. Try to pick up a manual written . . . >SGML isn't for memos or for *sexy* user manuals. Good systems and good >training save lives. Both depend on accurate information, not creative >layouts or way cool web pages. SGML helps make this possible in one of the >most complex contracting and procurement systems in the world. If giving >up WYSIWYG is the cost, junk it. I've used systems that allow both. They >can rock together. As for me, I use SGML for one and only one reason: > >Results. That is where SGML shines. . . . I have a related way of thinking about this that brings some of the appeal back to those not-so-large and not-so-complex (more typical of my customers): SGML facilitates concurrent engineering. SGML allows/obliges us to modularize presenta- tion, content, ... issues, and therefore permits division of labor to work its magic. As Claude and Eliot and others have related, we're also now in the fortunate position of having tools that permit "synthesis of labor", so that an SGML en- vironment can also be WYSIWYG (among other qualities). -- Cameron Laird http://starbase.neosoft.com/~claird/home.html claird@Neosoft.com +1 713 267 7966 claird@litwin.com +1 713 996 8546 From: claird@Starbase.NeoSoft.COM (Cameron Laird) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 12 Jun 1995 10:38:07 -0500 Organization: NeoSoft Internet Services +1 713 968 5800 Message-ID: <3rhn0v$10e@Starbase.NeoSoft.COM> References: <3rhh7l$2ru@reuters2.mitre.org> In article <3rhh7l$2ru@reuters2.mitre.org>, Hugh McKee \ wrote: > >>From: suthers+@pitt.edu (Daniel D Suthers) . . . >>to a query, automatically generate specialized documents, etc. For example, >>suppose your company suddenly needed to make portions of the documents >>available on the internet for use by customers or service personnel at other >>sites. > >BUT, there is an essential preceding step, the writing of the DTD. If >the authors of the DTD did not know that portions of the documents would >be used by customers or service personnel then I think there will be >difficulties. As I understand it (probably imperfectly), to revise a >DTD, and reprocess all the documents, can be a daunting task. > Yes. It's less daunting than any of the alternatives, though. The problem--migrating data to new uses and environments--is one that faces many organizations, but few of them realize that's what they're about with the clarity that a focus on DTDs provides. -- Cameron Laird http://starbase.neosoft.com/~claird/home.html claird@Neosoft.com +1 713 267 7966 claird@litwin.com +1 713 996 8546 From: reedts9@wfu.edu (thomas sykes reed) Newsgroups: comp.text.sgml Subject: SGML Toolkits? Date: 12 Jun 1995 18:08:44 GMT Organization: Wake Forest University Message-ID: <3rhvrc$rfa@eis.wfunet.wfu.edu> X-Newsreader: TIN [version 1.2 PL2] I'm creating an application that will contain a SGML browser as part of it. I'm looking for some SGML Toolkits to use in this current development. I am aware of EBT's DynaText System Integrator Toolkit. I understand SoftQuad is releasing a SGML Browser toolkit sometime (hopefully). Are there any others out there that supports a Windows platform? Please reply to Tom Reed, reedts9@wfunet.wfu.edu. From: streich@austin.wireline.slb.com (Robert Streich) Newsgroups: comp.text.sgml Subject: Re: SGML-to-Windows online help? Date: 12 Jun 1995 21:16:20 GMT Organization: "Schlumberger Austin Research" Sender: streich@odie (Robert Streich) Message-ID: <3riar4$492@maggie.austin.wireline.slb.com> References: <3ra6nr$7pf@bard.sc.intel.com> nnavarro@gomez.sc.intel.com (Nancy Navarro ~) wrote: > Does anybody know of a tool to get from SGML files to Microsoft > Windows online help files? The reason is we want to start with > SGML source and "down-translate" to multiple formats, among them > MS Windows help, others being html, and also to hard-copy. > > Any clues ANYONE has, your reply will be MOST appreciated! (I > have done some preliminary asking around and no one so far > knows of anything direct. I am aware that a series of conversions > using Bristol Technologies' Hyperhelp might be a rather arduous > solution, but was hoping for something more straightforward.) Not a direct answer to your post, Nancy. But you may want to give Exoterica a call. Talk to Eric Skinner, 1-800-565-9465. They have a lot of experience at down translations to RTF. It shouldn't be a big deal to add the Help RTF extensions. I doubt that you will find anything off-the-shelf because of the difficulty of generalizing any tool to work with any DTD. Having used OmniMark for a couple of years now, and having had nothing but good experiences with Exoterica as a company, I can recommend the company and the product highly. As a matter of fact, I'm in the final testing stage of a down-translate program to Interleaf ASCII that we wrote in OmniMark. It made for a nice, clean, straightforward program. Second choice, if you have time to wait, is a product that should be out shortly called DynaPage. It's being developed by AIS, sold through EBT. Using DynaText sheets and Balise, it will transform your doc to Interleaf ASCII, MIF, or RTF. I'm not sure what state the RTF stuff is in, but once complete, you may have enough control to put in the extra RTF to support Help. AIS seems to have some pretty good ideas about their product. Personally, I'd still go for the first option. This way you're assured of control over what is happening. I've grown _extremely_ wary of any solution where I wind up depending on the capabilities of some product in the processing chain especially a brand new one. I've been bit too many times lately. I have to ask the obvious question, however. Why use Windows Help? Do you have a choice? -- Robert Streich streich@austin.sar.slb.com Schlumberger 512-331-3318 (voice) Austin Research 512-331-3760 (fax) From: topher@ERC.MsState.Edu (John Christopher Lakey) Newsgroups: comp.text.sgml Subject: creating my own application Date: 12 Jun 1995 23:40:32 GMT Organization: Mississippi State University Message-ID: <3rij9g$bqh@NNTP.MsState.Edu> X-Newsreader: TIN [version 1.2 PL2] I have a question about how I should go about creating an application. Lets say I have a document or set of document I want to markup. Lets pick the resume mentioned earlier as an example. A. I design my DTD - hopefully I put a lot of thought into this part. B. I write applications which use my documents I am going to create. C. I try andsell this system to someone or give it away and become well known. (atleast to a few people on some newsgroup somewhere). By the way, resumes is not what I have in mind, even though I do have a friend who works for a career services center and they use a resume program which prompts you for all kind of information, but doesn't use any of it in the actualy layout of the resume. I was most unpleased to have entered all of this information only to find out it was used strictly for searching and I had to create a resume from scratch (inside the same application). Anyhow ... Part B is what I am concerned with. What are my options now. Let's say I want to create some resumes using an editor I write (maybe a FORMS based CGI script).Are these the steps I take 1. Verify the document using sgmls or sometheng else 2. Parse the output of sgmls using my own parser 3. Do whatever to the parsed output a-create a searchable database b-format is using one of several resume style sheets c-convert the document to HTML or something else (actually, same as 3.b, right). I guess I said all of this to say - are there tools that I can use after sgmls to make my job easier? Actually, I am thinking of creating a DTD for the Bible. I could use this DTD to create a humongous index and other cool stuff. Browsing is not really a goal right now. I just want to mark it up to be able to extract information. If anyone knows of a marked up Bible that is PD please let me know. Also, if anyone is interested in helping me design a DTD send me some mail. Here are some tags Iam thinking about INFORMATION ABOUT TEXTUAL DIVISONS ------------------- title author time period tag editor TEXTUAL DIVISONS ------------------- TESTAMENT BOOK CHAPTER VERSE PARAGRAPH PSALM POEM SONG WORD QUOTE REFERENCE CLASSIFICATION TAGS ------------------- CHARACTER (TYPES = man,woman,angel,animal)(NAME=)(ID=) MEASURE (TYPE = wieght,length,volume...)(AMOUNT=)(UNIT=cubit,foot,...) LOCATION (TYPE = city,village,well,mountain,country,...)(NAME=)(ID=) DATE (TYPE = whatever calender being used) OBJECT (???) MATERIAL (TYPE = GOLD,SILVER,ACHAI WOOD,...) (more?,less?) I only started reading about SGML yesterday, but so far I think it's pretty nice. sgmls compiled on my system and I wrote a simple DTD and I like the ideas mentioned on this newsgroup. I usually get suckered in on things that sounds cool. I learned TeX (which is still cool, but more of a hastles than I first realized), Tcl (still on my good side), etc. Hopefully SGML will do everything everyone is saying it will do. Sorry to be so long winded. If you have any advice or are interested in talking about a Bible DTD send me some mail. -topher From: tlehr@clark.net (Tom Lehr) Newsgroups: comp.text.sgml Subject: Anyone know of Synth\æsis by \Æsthetex? Date: 13 Jun 1995 02:06:46 GMT Organization: Clark Internet Services, Inc. Message-ID: <3rirrm$nh0@clarknet.clark.net> Mime-Version: 1.0 X-Newsreader: WinVN 0.93.11 Some months ago I attended a demonstration by a Scottish company, AEsthetex, of their SGML publishing software, Synthaesis. The software appeared to be a native SGML editor that operated like a database and could produce high-quality typeset output. It seemed to address many of the areas of traditional print publishing that have been ignored by other companies producing SGML applications, for example, tracking & managing changes to documents and providing formatted output. However, I have not found any reference to this company or its products in any of the SGML areas of the Internet. Does anyone have any knowledge or experience of this software? It seems odd to me that there would be SGML software on the market that nobody was discussing here. From: mwm@contessa.phone.net (Mike Meyer) Newsgroups: comp.text.sgml Subject: Re: Blessings and Curses Date: Mon, 12 Jun 95 21:39:25 PST Organization: Missionaria Phonibalonica Message-ID: <19950612.7668BC0.1357F@contessa.phone.net> References: <9506092043.AA37374@source.asset.com> <3rgi4u$c9m@news1.delphi.com> X-NewsReader: Amiga Yarn 3.9, 1995/05/09 10:42:03 In <3rgi4u$c9m@news1.delphi.com>, j_mcarthur@BIX.com (Jeffrey McArthur) wrote: > I like that. Just like two guys in a garage changed the way the world > thought of computers when they invented the Apple II. The Apple II had > some serious flaws. But it did inspire others. Where would we be today > without people who get up and do something rather than wait for government > officials and industry representatives to do the work? Hard to see how it could be worse than the present situation, with a standard for desktop computing that's two or three levels of kludge stacked on top of kludge, three different kinds of memories including limits that date from 70s hardware design, incomplete protection of tasks from each other, half-assed multitasking, and a collection of mistakes that the mainframe designers had made in the first place, doing a bad job of imitating what was done on less expensive hardware over 10 years ago. With real luck, we can get history to repeat itself with PDF instead of HTML playing the role of MSDOS, so that transition to a web based on real SGML won't be nightmare. \ From: gelpo@news.jyu.fi (DigPub) Newsgroups: comp.text.sgml Subject: digital publishing model Date: 13 Jun 1995 07:33:50 GMT Organization: University of Jyvaskyla, Finland Message-ID: <3rjf0u$2hu@mordred.cc.jyu.fi> X-Newsreader: TIN [version 1.2 PL2] Hello! We are studying digital publishing at University of Jyvaskyla and our aim is to develop a suitable publishing model for all kinds of publishing stuff at our university. We would like to as far as possible automate transformation of documents first to SGML standard and then with different conversion programs to HTML. We are working in unix-environment and our intend is to use public domain programs. If you in your organization are using or studying corresponding model, we are anxious to hear from methods and programs you use. Also other information relating to subject is appreciated. t. DigPub-project -- ========================================================================= Riitta Kunnari of. : MaE 112 Virpi Lyytikäinen tel : 941-603050 Tiina Pietinen email: gelpo@jytko.jyu.fi University of Jyväskylä, PL 35, 40351 JYVÄSKYLÄ, FINLAND ========================================================================= From: bruce@sgml.dircon.co.uk (Bruce Hunter) Newsgroups: comp.text.sgml Subject: CALS table processing Date: 13 Jun 1995 11:06:19 GMT Organization: SGML Systems Engineering Message-ID: <3rjrfb$n65@newsgate.dircon.co.uk> X-Newsreader: WinVN version 0.81 Does anyone know if there exists a rigourous, authorative description of the intended CALS table processing model? I have MIL-M-28001B and the SGML Handbook, but they are not specific enough about many areas. I want enough detail to know how to write a CALS table handler for an SGML Viewer, such that it can correctly handle all the various permutations possible in the CALS table DTD. Without a detailed descrition of the implied semantics this will be impossible. Does such a thing exist? Or does an "official" FOSI fragment exist which shows completely how to process tables? Product vendors who claim support for CALS tables all currently support restricted subsets. Try moving the same moderately complex CALS table between Softquad, Incontext, Arbortext and Grif products, and watch the results. I want to create something which will support CALS tables in the way in which the authors of the DTD intended, but that is impossible without a very detailed description of what was intended. If such a description of a processing model does not exist, is anyone willing and able to try to answer specfic questions as they arise? There are many things which are unclear or undefined. Some of the ones currently bugging me are : 1) Is there only one name space for colnames and spannames throughout a table, or are they specific to a tgroup, thead, tfoot, entrytbl, etc? I would guess that they all share the same name space, as spanspecs referenced from an entry in a thead are subelements of the parent tgroup and are not local to the thead. But should it then be an error to have a colspec in an entrytbl, say, with the same name as a colspec or spanspec in the surrounding table? Or are entrytbls something different again, with their own local namespace? 2) Is this valid : \ \ \ \ \ \ If so, to which column does the second colspec relate? Or is it an error if at least one colspec has a colnum specified and others do not? Would re-arranging the above to : \ \ \ \ \ \ be valid? Should colspecs with colnums appear in the order of their colnums or is this an error? 3) Why does the valign attribute on entry have a default value? This makes the valign attribute on thead, tfoot, tbody redundant. I have lots more questions for anyone willing to try to take them on. Thanks to anybody who can assist. regards, Bruce Newsgroups: comp.text.sgml From: Steve Pepper \ Subject: Re: Anyone know of Synth\æsis by \Æsthetex? Message-ID: <1995Jun13.143157.18605@falch.no> Organization: Falch Infotek, Oslo, Norway References: <3rirrm$nh0@clarknet.clark.net> Date: Tue, 13 Jun 1995 14:31:57 GMT tlehr@clark.net (Tom Lehr) writes: >Some months ago I attended a demonstration by a Scottish company, >AEsthetex, of their SGML publishing software, Synthaesis. The software >appeared to be a native SGML editor that operated like a database and >could produce high-quality typeset output. It seemed to address many of >the areas of traditional print publishing that have been ignored by other >companies producing SGML applications, for example, tracking & managing >changes to documents and providing formatted output. However, I have not >found any reference to this company or its products in any of the SGML >areas of the Internet. Does anyone have any knowledge or experience of >this software? It seems odd to me that there would be SGML software on >the market that nobody was discussing here. As far as I know, the software is still some way from completion (at least the SGML part of it; Synthaesis is also a general typesetting system). I haven't had the chance to look at it yet but it certainly sounds like a very interesting product. For more details, get in touch with AEsthetex 626 Lanark Road Juniper Green Edinburgh EH14 5EW UK Tel. +44-131-458 3933 Fax +44-131-458 3903 email: angus@aesthetex.co.uk Best regards, Steve -- pepper@falch.no -------------------------------------------------------------- falch infotek a.s, postboks 130 kalbakken, n-0902 oslo, norway tel +47 2290 2733 fax +47 2290 2599 http://www.falch.no/ From: jad \ Newsgroups: comp.text.sgml Subject: Re: A New Era of Afforable Tools Date: 13 Jun 1995 14:34:40 GMT Organization: EUnet Belgium, Leuven, Belgium Message-ID: <3rk7m0$g4i@news.Belgium.EU.net> References: <3r2h8t$9gn@pubxfer4.news.psi.net> <3r4g0n$1bs@argo.hks.com> jeffrey@hks.com (Glenda Jeffrey) wrote: > > W. Eliot Kimber (kimber@passage.com) wrote: > : I attended the recent SGML Europe '95 conference. On display were > : a number of soon-to-be-released SGML tools, primarily editors, > : that will, I think, dramatically change the economics (and thus > : acceptance) of SGML. > : > : \ > > Were any of these new products capable of dealing with equations > in a sensible way? > > And what about tables? > -- We just received a Beta C version of WP SGML Edition. Nothing documented about equations (so, probably no support ?), although WordPerfect's equation "language" is very eqn-like. As regards tables, there is built-in support for : CALS AAP and ISO 12083 WordPerfect tables (a DTD fragment matching WP tables) Conversion of tables is completely transparent during Save As SGM ... or Open SGM ... operations. A more detailed account will follow soon. --------------------------------------------------- Jacques Deseyne \ Sema Group Belgium Stallestraat 96 B-1180 Brussels Belgium --------------------------------------------------- From: sujoy@sol.uconn.edu (Sujoy Sen) Newsgroups: comp.text.sgml Subject: SGML grammar and/or parser Date: 13 Jun 1995 15:37:41 GMT Organization: Systems Optimization Lab, UCONN Message-ID: <3rkbc5$4sm@bellboy.ucc.uconn.edu> References: <9506061732.AA6148@notes.microstar.com> Reply-To: sujoy@sol.uconn.edu Keywords: parser X-Newsreader: xvnews 2.3 Beta PL5, use only for testing please. Hi I need to write a SGML parser for a very specific application. I was wondering whether I could get hold of the grammar somewhere and/or a unix based parser. I did download a parser (sgmls) but cannot figure out its output. Any help would be greatly appreciated. -- Sujoy From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: SGML grammar and/or parser Date: 13 Jun 1995 09:43:19 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3rkf77$lk1@crl2.crl.com> References: <3rkbc5$4sm@bellboy.ucc.uconn.edu> Keywords: parser Sujoy Sen \ wrote: >I need to write a SGML parser for a very specific application. I was wonderin >I could get hold of the grammar somewhere and/or a unix based parser. I did >a parser (sgmls) but cannot figure out its output. The output format is thoroughly documented in the man page; what don't you understand? --Joe English jenglish@crl.com From: grmeyer@mcs.com (Gordon R. Meyer) Newsgroups: comp.text.sgml Subject: Opinions about DynaText? Date: Tue, 13 Jun 1995 11:54:41 -0600 Organization: Our Lady of Chaos and Silicon, inc. Message-ID: \ If anyone has some feedback/opinions/stories they'd like to share about EBT's DynaText online documenation system I'm all ears. :-) I'm currently involved in an initial evaluation of the product. We are not an SGML-based shop (we use Word for Windows) so if anyone has any conversion-related notes those would be of interest too. Thanks! --Gordon -- Gordon Meyer grmeyer@mcs.com Friends Don't Let Friends Use Windows (TM) From: tprodin@vp2247.pms.ford.com (Timothy R Prodin) Newsgroups: comp.text.sgml Subject: Re: Blessings and Curses Date: 13 Jun 1995 18:46:56 GMT Organization: EE CAE, Ford Motor Company Message-ID: <3rkmf0$8oq@eccdb1.pms.ford.com> References: <9506092043.AA37374@source.asset.com> Reply-To: tprodin@cadcam.pms.ford.com In article <9506092043.AA37374@source.asset.com>, Claude L. Bullard \ wrote: >It seems I spend more of my days than I ought explaining the relationship >of SGML and HTML to inquiring minds. Take it or leave it, I think HTML to >be "Yet Another SGML App". Use it if it works for you. But the weirdness >of the questions, the misunderstandings, all serve to make me queasy about >this shotgun marriage. I submit the following quote from a piece of >strange journalism from "Internetwork: The Manager's Magazine of Standards >and Interoperability". The author is Brad Harrison. > ...... > >The information highway, despite all the rhetoric and enthusiasm, is >nothing more than a collection of aging technologies that are being >extended far beyond what they were initially designed to do. HTML >certainly shows that you can make an old dog do new tricks - keep in mind >that it was born at IBM in the 60's - but the dog remains old and he will >eventually die. Hm....So far, considering he used the dreaded ih phrase, I don't think this writer really knows what he is talking about. The claim that HTML (really SGML) was born at IBM in the 60's is somewhat wrong. True, it can be traced back to IBM's Document Composition Facility, but it has been changed quite a bit from then to now. .... > At this rate, the >global network will become an illusion, a collection of independent >networks somewhat held together by old technology - a technology that was >originally designed to help us fight communism." And here is the clincher. The Mr. Harrison clearly doesn't understand: 1) just because something is old, doesn't mean it doesn't work well. 2) technologies like TCP/IP and SGML are designed to facilitate growth and reuse by being extensable to modern technology. From: fazal@karp.cs.albany.edu (Fazal Ilahi) Newsgroups: comp.text.sgml Subject: Newbie question on Qwertz Date: 13 Jun 1995 18:52:57 GMT Organization: cs Message-ID: <3rkmq9$jkk@rebecca.albany.edu> Hi I have just begun using qwertz .The dtd's that I am using parse through without errors using sgmls but when I use the format program of qwertz it brings up sgmls errors.I understand that the dtds that I am using have to be of the qwertz document type.how do I ensure that. Please help, Fazal From: nik@dreamland.tcp.co.uk (Nick Barron) Newsgroups: comp.text.sgml Subject: Re: Technical Writing student---SGML ?Where should I start??? Date: Tue, 13 Jun 95 19:52:24 GMT Organization: None Message-ID: <19950613.2F66D8.11C12@dreamland.tcp.co.uk> References: <3rabvr$6on@sentinel.synapse.net> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-NewsSoftware: GRn 2.1 Feb 19, 1994 In article <3rabvr$6on@sentinel.synapse.net> Ray Chenier \ writes: > Hi: My name is Ray Chenier and I am a student in technical writing. I know SGML is The comp.text.sgml FAQ is a good starting point, and there's also a good bibliography floating around. Mail me if you have trouble finding these. Have fun! Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 13 Jun 1995 16:14:24 -0400 From: "Claude L. Bullard" \ Message-ID: <9506132014.AA08491@source.asset.com> References: <9506092043.AA37374@source.asset.com> <3rgi4u$c9m@news1.delphi.com> Subject: Re: Blessings and Curses [Jeffrey McArthur] | Where would we be today without people who get up and do something | rather than wait for government officials and industry representatives | to do the work? That's the blessing part. The curse was when those same gentlemen claimed to own parts of a design that they lifted from Xerox and Xerox lifted from Douglas Englebart. Englebart worked at Stanford and served on the committee that created the Dexter Reference Model, which is still the best design for hypermedia. Government and industry paid to support the standards from which the Web is derived, they paid the salaries at CERN and at U of Illionis. They paid for the wires, repeaters, routers and most of the other technology in which you thrive. Industry and government work. Inspiring. Oh, and if you know the history of Apple, you know that the man who made the Apple happen was a retired business executive, not a garage hacker. Myths are inspiring also, but terrible history. Note that the author of the article I am quoting considers HTML to be a disaster forced on users by a raging media and kids too inexperienced to know the difference between good design and flashy interfaces. For the downside of the same scenario, check out the Wired article describing the disaster that was Xanadu. Trying to do it *right* can often lose to those who simply choose to do the piece that gets the most glory. It isn't sad; just a lesson about quality and the mass market. Sometimes they combine and we get a George Gershwin. Sometimes they don't and we get Madonna. Both sold big. Both had impact. Gershwin had genius and advanced the state of the art. The selling of Madonna is undeniable; the contribution to art is debatable. I think the same can be safely said of the Web. Its impact is a fait accompli, but its contribution to the state of the art is negligible. I know about the "glory" scores, the millions of users flocking to the Net, the commercial applications growing like roadside fireworks stands, the disempowered voices that now flame brightly where once only dull researchers sent quaint unadorned email. That batter has run the bases and is basking in the bullpen. The next batter is just warming up. Choosing to implement half a solution, the Internet changed. Now the Congress and the service providers are going to clamp down hard. Parents are looking for children in bus stations. Rights never questioned before are going to disappear from the lawbooks as if they never existed. There is a strong movement toward neo-Ludditism. There are infrastructure owners trying to figure out how to get the most money from services that used to be free. Lawsuits in colleges over sex discrimination have forced some online services to be closed. Very inspiring. No its not fair to lay all of that at the feet of the Web. It simply popularized the Internet and made it profitable. Those of us who are engineers and must plan for the next generation of Webs should give some thought to this and not simply attempt to colonize for our own glory or sense of *fighting the system*. We did that in the Sixties. What did we learn? Some learned the awesome and frustrating need to be responsible, informed individuals. No crowd need apply. A generation of young engineers is feeding on hype and pizza, believing as we once believed they are better than their parents, more moral, and smarter. We discovered we weren't different, we were just as corruptible and corrupt, and only a bit better educated. In some ways it was "a lesson too late for the learning... made of sand." In other ways, it was splendor and ecstacy. We became what we wanted to be: free of the illusion of freedom and responsible to the freedom to choose among illusions to which we would enslave our lives. Willingly. C'est un beau geste, n'est pas? Some say you are creating freeware to save the world. You aren't. You are doing free research to be harvested by Netscape or Silicon Graphics or me. Thank you. There is a natural tendancy to fight the system. That is a thrill but at some point, one still has to get up and make a profit to have a decent life. Beware the egoism of thinking that nothing old has merit. Everything the Web does was done before. Only the scale changed. The new and the old are synthesized. We can direct evolution, but we are changed by it. We can make a difference, but we don't always get to choose the difference it makes. So, revel in your time of "splendor in the grass", then change the thing that you dependably can: your mind. Then, you can be where you want to be. Len Bullard Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 13 Jun 1995 17:45:17 -0400 From: "Claude L. Bullard" \ Message-ID: <9506132145.AA14857@source.asset.com> References: <9506092141.AA19355@source.asset.com> <3rgi54$c9m@news1.delphi.com> Subject: Re: Purpose of SGML in 3 paragraph [Jeffrey McKern] | SGML does not drive this process. What drives this process is the | publisher and/or copyeditor. They are the ones who care about the | format. It does not matter what the author says, the publisher is the | one paying the bills. If the author says "make this bold", but the | publisher says, "make it italic" want to guess who wins? The money wins or he who shouts loudest or gets to the boss first. Certainly, for some organizations, this is true. For others, no such copy editor exists and the author does have a variety of choices. I didn't make my point clear. I should have said the DTD can be used to drive the software that drives the editing process. The idea is that SGML is not always only a downstream format for transport. In more powerful systems, the DTD is used to tell the editing system exactly what information is to be entered. If that DTD is HTML, one is specifying in the main, that the look and structures of mosaic display applications are intended. If one uses a DTD such as MIL-D-87269, one intends that a database of logistics information is to be created. For some applicationss, the look is dominant as they are not selling the information, they are using the look to attract the consumer. Perfectly legitimate. The problem arises when someone in the food chain can't make the right decision about which requirement should dominate, or is denied by the local system the ability to express that intention. As to consistent look for large bodies of information, that is a local decision for the organization. IMO, SGML isn't where that decision should be enforced. SGML is better applied to enforcing content. The marriage of SGML element types to presentation types robs the author or the organization of the power to make choices. They shouldn't marry; they should date and change partners when they become boring or unreproductive. One can use the techniques of DSSSL or FOSI or the Explorer and IADS systems which use SGML to specify style information and output that if the need arises. I want to interactively specify styles because I don't enjoy layout design. So, yes, I like to have WYSIWYG and SGML, but if I have to choose, I choose SGML because the complexity of organizing and testing the content is more important to me. If an organization-wide stylesheet exists, I prefer using it over reading a specification to divine what is required and raising a herd of macros to do the layout. Most use SGML precisely because we don't want to do that. The copy editor or publisher can control all of that if they will leave the content alone. Subject matter experts write the DTDs for the content. This of course, has everything to do with what one is writing. Tech manuals, advertisements and e.e. cummings just don't have much in common but the glyphs. SGML can account for the intentions of each where those intentions can be expressed as SGML applications. Screen layouts have requirements that printed pages never have. You can't resize a page, change its fonts, choose a novice or expert level, etc. Frankly, as pressed byopia overtakes me, the ability to reset fonts is much appreciated. We've been over this ground a lot on the CTS. Applicationss are starting to appear which do what is required. It is important to preserve and express intent: of the author, of the user, of the customer, of the copyeditor and the subject matter. SGML can be used to express intent, but often not everyone's intent can be clearly expressed in the same DTD. That is why some separate them and control the order in which intentions are bound to the presentation or the database. That is also why the new generation of SGML systems and standards are welcome and a bit overdue. Len From: haim@cc.bellcore.com (kilov,haim) Newsgroups: comp.text.sgml Subject: Re: Adaptive documentation (different documents for different users)? Date: 13 Jun 1995 19:09:43 -0400 Organization: Bell Communications Research (Bellcore) Message-ID: <3rl5rn$dk0@dancer.cc.bellcore.com> References: <3r7pe5$l5d@ruby.ora.com> Andy Oram: "I'm interested in the possibility of producing adaptive online documentation. That is, if you choose a "novice" button you get one series of links, if you choose "intermediate" you get another, and so on. (Of course, there are many other criteria you could choose in creating different documents.) " Yes, indeed. But probably different users will need different content fragments of a document (or a document collection), not just different links. Thus, you will need a model (specification) of the intellectual content of a document; and then you will be able to use it as a road map, or, to be more precise, as a content map. A paper on these issues "A document management model" will be published in "Computer Communications", I think, in June 1995. There were also other papers, e.g., in Proceedings of the OOPSLA94 workshop on behavioral semantics ("Understanding (hypermedia) document collections"). These papers are based on the idea of using the same precisely defined concepts and constructs for modeling information and documents that describe this information. Hope this helps. -Haim Kilov haim@cc.bellcore.com From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 14 Jun 1995 00:17:46 GMT Organization: \ Message-ID: <3rl9ra$18v@news1.delphi.com> References: \ X-Newsreader: Galahad 1.1b RS and RE: a simple proposal The SGML declaration defines many capacity values. Capacity points allows applications working with SGML data to have a rough indication of the storage required. In the current standard there are no indications about the size of records. So some SGML editing applications create files that is one large record and others produce files that are roughly 80 characters or less. Many existing text processing tools have finite input limits. TeX for example, normally has its input buffer limited to 1024 characters (some implementations allow more, some less). Lex is another text processing tool that has a input buffer limit (this varies wildly). Trying to use any of these existing tools on SGML data can be a difficult problem because there is no indication how large a record the data may have. The only limit on the record size is the TOTALCAP. One solution to this problem would be to add a new capacity to the SGML declartion providing a limit on the record size. RECORDCAP would be one possible name. Backward compatability can easilly be achieved by setting the default value of RECORDCAP to be equal to TOTALCAP. However this practice would be depricated. Ideally, RECORDCAP would be some moderately small value to encourage the use of standard off-the-shelf processing tools. ================================================ RS and RE: a more radical proposal One of the common problems with SGML are DTD's with mixed content models. The problem with mixed content model is that in some cases the exact syntax can be confusing to someone editing the data. In the concrete reference syntax RS is defined to be 10 and RE is 13. Tab (9) is a SEPCHAR. Many computer languages defined today make no distinction between line-break characters (10 and 13), tabs (9) and spaces (32). All of these are treated as white space. SGML already has the concept of SPACE and SEPCHAR. I propose that (10) and (13) be defined as SEPCHAR's and not as RS and RE. This would mean that line breaks and tabs would be interchangable. This does away with the complicated RS/RE rules in mixed content models. This would simplify processing of SGML data by non-sgml aware tools. The current rules for processing of data using RS/RE make it difficult to properly process SGML data with tools not specifically designed for SGML. Backward compatiblity can be assured by just defining RS and RE as they currently are. However that practice would be depricated. (For those interrested in seeing what this would mean, just replace the line-breaks in an SGML file with tabs.) ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 14 Jun 1995 00:17:52 GMT Organization: \ Message-ID: <3rl9rg$18v@news1.delphi.com> References: \ X-Newsreader: Galahad 1.1b Erik Naggum wrote: >| Take the approach the N. Writh has taken to programming languages: cut >| out all the needless crap. If in doubt take it out. Then if a case >| can be made that the construct is really useful, you can always put it >| back in. > >this reflects the actual ISO process. unlike what you appear to think, ISO >committees do not bloat standards for fun. there have usually been people >who argue fiercely against any addition, as well as those to argue fiercely >against any deletion. additions _are_ more easily accepted than deletions, >however. putting things back in doesn't actually work the way you think. I see a serious contradiction here. The approach Wirth has "if in doubt take it out", but you said "additions _are_ more easily accepted than deletions". Those statements are contradictory. I would make it almost impossible to have anything added, but very, very easy to have constructs removed. Have you read the article by Wirth, "From Modula to Oberon"? For-loop constructs have been part of computer languages for decades. But in Oberon-1, Wirth took them out. The case that For-loops were necessary was not strong enough. For-loops were added back to Oberon-2 (eventually a strong enough case was made). ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: rieger@bse.de (Wolfgang Rieger) Newsgroups: comp.text.sgml Subject: Re: CALS table processing Date: Wed, 14 Jun 1995 01:27:34 GMT Organization: MUC.DE e.v -- private Internet access Message-ID: <3rldpp$hkl@marvin.muc.de> References: <3rjrfb$n65@newsgate.dircon.co.uk> X-Newsreader: Forte Free Agent v0.46 bruce@sgml.dircon.co.uk (Bruce Hunter) wrote: >Does anyone know if there exists a rigourous, authorative description of >the intended CALS table processing model? There is a documet with the title "A recommended navy interpretation of the MIL-M28001B table model - revision 1" dated June 1994 which may answer some of your questions. I received that document in postscript format from Betty Harvey at harvey@navysgml.dt.navy.mil. Regards Wolfgang Rieger Wolfgang Rieger Email: rieger@bse.de bse - Buero fuer Software-Entwicklung WWW : http://www.bse.de/ Frankfurter Ring 193a Phone: +49 89 323 19 93 80807 Munich, Germany Fax : +49 89 323 19 93 From: amanda@intercon.com (Amanda Walker) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: Wed, 14 Jun 1995 00:33:53 -0400 Organization: InterCon Systems Corporation, Engineering Dept. Message-ID: <9506140033.AA53791@amandap.dial.intercon.com> References: \ Reply-To: amanda@intercon.com (Amanda Walker) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII X-Newsreader: InterCon TCP/Connect II 2.1.4b2 boo@primenet.com (Walter Ian Kaye) writes: > For line art, sure. But what about pixel/halftone/etc graphics -- how > does pdf compare with gif in that area? About the same. Images in PDF files can be either LZW compressed (same algorithm as GIF uses) or JPEG compressed. Amanda Walker InterCon Systems Corporation Newsgroups: comp.text.sgml From: ubacr63@pluto.dcs.bbk.ac.uk (Dan Smith) Subject: SGML based databases Message-ID: <1995Jun14.114905.55821@ucl.ac.uk> Date: Wed, 14 Jun 1995 11:49:05 GMT Organization: University College London X-Newsreader: TIN [version 1.2 PL2] I am currently researching textual databases and am wondering if there are any extant SGML based (experimental / non-experimental) database systems in existence. As there seems to be a lot of data being produced in SGML format it seems sensible to explore this approach. If anyone has any information I would be grateful so that I know what to and what not to follow up! TIA Dan From: Nils ten Broeke \ Newsgroups: comp.text.sgml Subject: My Home Page - searching for a new job Date: 14 Jun 1995 12:32:24 GMT Organization: Faculty of Arts, Utrecht University, NL Message-ID: <3rmkso$3f8@ruulch.let.ruu.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1b3 (Windows; I; 16bit) To: comp.text.sgml Utrecht, June 14th 1995 To every reader of this newsgroup. With this message I would like to invite everybody to have a look at my home page: it's new. At URL: http://www.let.ruu.nl/C+L/broeke/home.htm/ you can discover my fields of interest and information about my searching a new job. Right now I'm working at the Utrecht University (The Netherlands). Because my present job will end on August 1st 1995 I'm looking for a new job in which I would very much like to be able to work with markup languages (SGML, HTML) and hypertext applications. For more information about this and more please see my WEB pages. Thank you very much. From: Nils ten Broeke \ Newsgroups: comp.text.sgml Subject: My Home Page - searching for a new job Date: 14 Jun 1995 12:32:39 GMT Organization: Faculty of Arts, Utrecht University, NL Message-ID: <3rmkt7$3f8@ruulch.let.ruu.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1b3 (Windows; I; 16bit) Utrecht, June 14th 1995 To every reader of this newsgroup. With this message I would like to invite everybody to have a look at my home page: it's new. At URL: http://www.let.ruu.nl/C+L/broeke/home.htm/ you can discover my fields of interest and information about my searching a new job. Right now I'm working at the Utrecht University (The Netherlands). Because my present job will end on August 1st 1995 I'm looking for a new job in which I would very much like to be able to work with markup languages (SGML, HTML) and hypertext applications. For more information about this and more please see my WEB pages. Thank you very much. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: Purpose of SGML in 3 paragraph Date: 14 Jun 1995 13:21:45 GMT Organization: Naggum Software; +47 2295 0313 Expires: 28 Jun 1995 13:21:44 GMT Message-ID: <19950614T132145Z@naggum.no> References: <3rhh7l$2ru@reuters2.mitre.org> [Hugh McKee] | BUT, there is an essential preceding step, the writing of the DTD. If | the authors of the DTD did not know that portions of the documents | would be used by customers or service personnel then I think there will | be difficulties. As I understand it (probably imperfectly), to revise | a DTD, and reprocess all the documents, can be a daunting task. this is unfortunately true with most existing systems. (actually, I don't know _any_ existing system which helps this process, but I don't know all of them.) revising the data dictionary in a data base is considered to be an essential part of a functionally complete DBMS, and the support software will take care of maintaining the consistency of the database and hide the change from any existing software. I would have imagined that some of the knowledge and experience from this world could have been applied to the SGML world without entering the database business. the processing necessary to affect a universal implementation of a delta has the option of being very simple, as SGML allows "optional" elements, but we find that it is necessary to have degrees of enforcement of content models. this probably means that several DTDs will have to be maintained since we don't have such mechanisms in SGML today. i.e., an element may be "deprecated", such that it should cause an error in new documents, but should not cause an error in existing documents. i.e., one would use different DTDs for validation and processing. after all, if a document has been validated, it need not be validated again, and a looser processing DTD may solve a number of "compatibility problems". most SGML education today stresses "the" DTD, as opposed to "a" DTD that fits the data (and vice versa) according to some set of premises (out of many possible sets). I think this DTD "singularity" is very unfortunate, as it effectively constrains not only the uses of the documents, but also the uses of SGML. however, to make this work, we need software that can analyze two DTDs to decide whether conforming instances relative to one of them are also conforming instances relative to the other, but not vice versa, thus defining a less-than relation for DTDs. in the current software climate, we can just about forget such theoretical work if it doesn't come with a glitzy graphical user interface, so it probably won't be commercially viable, especially not if it isn't written in the language du jour, and if it is made commercially available, it will only be usable with a particular glitzy graphical user interface. *sigh* feel free to prove me wrong. #\ -- don't trust anyone under 30. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 14 Jun 1995 13:39:21 GMT Organization: Naggum Software; +47 2295 0313 Expires: 28 Jun 1995 13:39:20 GMT Message-ID: <19950614T133921Z@naggum.no> References: \ <3rl9ra$18v@news1.delphi.com> [Jeffrey McArthur] | The only limit on the record size is the TOTALCAP. how did you reach this conclusion? TOTALCAP only applies to the sum total of the other capacities such that the system capacity may be exceeded even if no other capacity limit is exceeded if their sum exceeds TOTALCAP. | One solution to this problem would be to add a new capacity to the SGML | declartion providing a limit on the record size. RECORDCAP would be | one possible name. Backward compatability can easilly be achieved by | setting the default value of RECORDCAP to be equal to TOTALCAP. | However this practice would be depricated. Ideally, RECORDCAP would be | some moderately small value to encourage the use of standard | off-the-shelf processing tools. insofar as we still want capacities, this is a good proposal. however, capacities should be reworked to be descriptive, instead of normative. i.e., an SGML document should include the actual capacity points needed, and the capacity point calculations should be broken up into more useful units. the present design is almost completely useless, and ignored. | One of the common problems with SGML are DTD's with mixed content | models. The problem with mixed content model is that in some cases the | exact syntax can be confusing to someone editing the data. definitely true. | In the concrete reference syntax RS is defined to be 10 and RE is 13. | Tab (9) is a SEPCHAR. Many computer languages defined today make no | distinction between line-break characters (10 and 13), tabs (9) and | spaces (32). All of these are treated as white space. SGML already | has the concept of SPACE and SEPCHAR. I propose that (10) and (13) be | defined as SEPCHAR's and not as RS and RE. this has the unfortunate side effect of changing or invalidating existing SGML documents. as long as we tolerate that, how about ignoring _all_ whitespace after a start-tag and before an end-tag, regardless of this (annoying) inclusion business? or is this too radical a change? #\ -- don't trust anyone under 30. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 14 Jun 1995 13:48:12 GMT Organization: Naggum Software; +47 2295 0313 Expires: 28 Jun 1995 13:48:11 GMT Message-ID: <19950614T134812Z@naggum.no> References: \ <3rl9rg$18v@news1.delphi.com> [Jeffrey McArthur] | I see a serious contradiction here. The approach Wirth has "if in | doubt take it out", but you said "additions _are_ more easily accepted | than deletions". Those statements are contradictory. I would make it | almost impossible to have anything added, but very, very easy to have | constructs removed. the approach taken by Common Lisp is to create a macro facility strong enough that users can create their own constructs, instead of waiting for Wirth (or someone else) to add necessary things back. I'm a strong favor of abstraction, and I think "if in doubt take it out" is only going to hinder abstraction. if SGML had had a meta-facility (such as a macro facility), I don't think we would have had to suffer HyTime's cancerous syntax. a consequence of SGML's non-existing ability to talk about itself, is that there needs to be a special engine on top of or in addition to the parser that processes the first layer of meta-information. recursively applied meta-information is impossible, especially meta-information that would apply to a lower layer. this is a _severe_ limitation in SGML today. with it, we wouldn't have to take things out, we could only tell SGML itself to ignore (or deprecate) something we don't like at a lower layer, such as _parts_ of features. #\ -- don't trust anyone under 30. From: medined@earth.planet.net (David Medinets) Newsgroups: comp.text.sgml Subject: SGML --> Xerox Metacode Date: 14 Jun 1995 15:59:31 GMT Organization: Planet Access Networks - Stanhope, NJ Message-ID: <3rn113$cvj@jupiter.planet.net> X-Newsreader: TIN [version 1.2 PL2] I have a need to convert some SGML documents into Xerox Metacode to drive Xerox 9700 printers. Would anybody have some pointers as to how this could be accomplished? thanks, david medinets Newsgroups: comp.text.sgml From: mark3@dcs.gla.ac.uk (Mark Magennis) Subject: Re: SGML based databases X-Nntp-Posting-Host: mac_f091_181.dcs.gla.ac.uk Message-ID: \ Sender: news@dcs.gla.ac.uk (Network News) Organization: Dept. of Computing Science, University of Glasgow References: <1995Jun14.114905.55821@ucl.ac.uk> Date: Wed, 14 Jun 1995 16:12:35 GMT As part of the e-journal project I'm involved in our German partners are using an SGML-aware database. They say: "We consider two possibilities for storing structured SGML documents in a database : The object-oriented database management system VODAK, developed at GMD-IPSI and the commercial database management system BASIS plus from Information Dimensions Inc. The VODAK prototype provides a complete database system kernel. It allows the user to write schemas and applications in the VODAK native language VML (VODAK Modeling Language). Using VQL (VODAK Query Language) declarative SQL-like queries can be posed against the database, either interactively or within methods and applications. The SGML support offered by VODAK is an application framework which partially consists itself of a schema written in VML. The document management system BASIS plus offers a set of software tools for developing applications that handle SGML documents. Our main focus will be on VODAK since, by its object-oriented approach, VODAK provides a better concept in handling additional semantics, e.g., the administration of the associated PDF version of the document." If you'd like to follow this up, try contacting Klaus Suellow at GMD-IPSI (suellow}@darmstadt.gmd.de). I'm sure he'd be willing to help. Mark From: kjs@ebt.com (Kent Summers) Newsgroups: comp.text.sgml Subject: Re: Adaptive documentation (different documents for different users)? Date: 14 Jun 1995 16:12:54 GMT Organization: Electronic Book Technologies, Inc. Sender: -Not-Authenticated-[3804] Message-ID: <3rn1q6$e31@spock.ebt.com> References: <3r7pe5$l5d@ruby.ora.com> <3r90r1$mf1@edf3.der.edf.fr> X-Posted-From: InterNews 1.0@spock.ebt.com. Xdisclaimer: No attempt was made to authenticate the sender's name. such behavior can actually be achieved in a much simpler manner than using hytime to solve this requirement (hytime topic link/map is however very well suited to this task). you could simply have different stylesheets which expose or hide content, as required. appropriate links would be embedded in the different skill level specific content types. this would allow you to test your theory against your end-user base without boiling the ocean. if you find that you need more control over how you links behave, independent of the text, then you might want to look into a hytime implementation. In article <3r90r1$mf1@edf3.der.edf.fr> Jean-Luc Sanson \ writes: > andyo@ruby.ora.com (Andy Oram) wrote: > > > > I'm interested in the possibility of producing adaptive online > > documentation. That is, if you choose a "novice" button you get one > > series of links, if you choose "intermediate" you get another, and so > > on. (Of course, there are many other criteria you could choose in > > creating different documents.) > > > > I've heard of this thing in principle, but I don't know any documents > > about it. Does anybody on this list have experience with it? > > > It seems to me that the problem you try to solve could be well described > using HyTime, and in particular using the Topic, Topic link and Topic > Map concepts which are discussed by the CApH group. CApH stands for Convention > on the Application of HyTime. This group is sponsored by GCARI. To know > more on CApH and its works contact Steve Newcombe (Technoteacher) > who chairs the group or Elliot Kimber (Passage Systems) > Steve could be reached through email at Hyminder@techno.com > and Elliot at kimber@passage.com > > Jean-Luc Sanson From: qis_fleiss@earthlink.net (Joel Fleiss) Newsgroups: comp.text.sgml Subject: HTML specifications by FTP? Date: 14 Jun 1995 16:17:39 GMT Organization: QIS Inc. Message-ID: <3rn233$31i@mars.earthlink.net> X-Newsreader: WinVN 0.90.3 We were wondering if there is a place to pick up an ascii file describing the latest HTML specification. We know there is hypertext information which can be browsed using Mosaic or Netscape, but just a complete printout that we can eyeball would make understanding HTML much simpler. Any suggestions on anonymous FTP sites or other net newsgroups where we can get more info on this? Since this is an SGML newsgroup, please e-mail us instead of posting. Thanks. QIS Inc. qis_fleiss@earthlink.net Newsgroups: comp.text.sgml From: Phil.Kime@ed.ac.uk (Phil Kime) Subject: A general question for people with SGML experience Message-ID: \ X-Attribution: Phil X-Www: http://www.cogsci.ed.ac.uk/~philkime/home.html Sender: usenet@cogsci.ed.ac.uk (C News Software) Reply-To: Phil.Kime@ed.ac.uk (Phil Kime) Organization: Centre for Cognitive Science X-Comments: MIME/PGP welcome X-Newsreader: Emacs 19.27.2 (GNUS 4.1) Date: Wed, 14 Jun 1995 16:42:16 GMT I'm posting to get a rough idea of the sort of time scales and costs associated with SGMLing large publications. I have a colleage who is interested in the possibility of SGMLing a large document (not dissimilar from, say, a University Calendar) from scratch. He's really appreciate an estimate of the sort of time it would take to get someone to do this. We are talking probably around 500,000 words. I was thinking along the lines of 2 or 3 weeks to write a DTD, around 2 months for markup and 2 months for a decent primary LPD. Plus a month for general hacks and filter writing. Is this the right sort of time scale for an SGML consultant? Also, what about cost? What does your average consultant charge nowadays? Is SGML even the right thing to do? The requirement is roughly that the document should be available on the Web and in printed form without unecessary duplication of time marking up into HTML from scratch or from the terrible output from the HTML export of most WPs and DTP systems. There is possibly even interest in interfacing with a database ... it sounds on the face of it as thought SGML is ideal for this but they really have no idea about the kind of time or cost involved and this seems like an ideal place to get rough idea ... p -- = Phil Kime (Phil.Kime@ed.ac.uk) = = Centre for Cognitive Science/Dept of Philosophy = = Edinburgh University = Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 14 Jun 1995 14:30:45 -0400 From: "Claude L. Bullard" \ Message-ID: <9506141830.AA45783@source.asset.com> Subject: Web Notations A question for the Web experts: how does a Web document indicate that it contains or references multiple notations? I guess in other words, can HTTP indicate multiple types within one document? Further, if it can, how does it indicate which version or flavor of the type is referenced or contained? For example, one may need a document instance in which HTML has been used for some parts of the text, another SGML application was used to create queriable tables, VRML is used to create alternative visualizations for data in the tables, and PDF is used to create non-editable images of copywritten material. These are just examples of course. Let us not debate the suitability of each notation for the application, please. In SGML, of course, we have notation declarations that indicate the notation and version which can be used as a registry for the servers required to support the document. The DTD in effect, plays the role of the catalog for types that must be supported and any foreign notations. Presumably, in due course, a standardized layer will evolve that should allow these notations to interoperate via a shared protocol. Is that protocol HTTP? As this is an issue of some importance to the evolution of the document database communications layer standardization for hypermedia, I am curious how the Web achieves this and what if any plans others have for supporting it. I ask this question on CTS because there is a fair chance that others may be dealing with this. For HTML or VRML only users, it may not be an issue yet. Len Bullard From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 14 Jun 1995 12:18:47 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3rncmn$cn8@crl9.crl.com> References: \ <3rl9ra$18v@news1.delphi.com> <19950614T133921Z@naggum.no> Erik Naggum \ wrote: >[Jeffrey McArthur] >| In the concrete reference syntax RS is defined to be 10 and RE is 13. >| Tab (9) is a SEPCHAR. Many computer languages defined today make no >| distinction between line-break characters (10 and 13), tabs (9) and >| spaces (32). All of these are treated as white space. SGML already >| has the concept of SPACE and SEPCHAR. I propose that (10) and (13) be >| defined as SEPCHAR's and not as RS and RE. > >this has the unfortunate side effect of changing or invalidating existing >SGML documents. as long as we tolerate that, how about ignoring _all_ >whitespace after a start-tag and before an end-tag, regardless of this >(annoying) inclusion business? or is this too radical a change? It would certainly be useful in some circumstances -- this exact behavior was an application convention in one of my projects (this was to avoid \ certain markup problems \ that kept showing up in documents) -- but I think that on the whole it would be too radical a change. It would be very useful to ignore all initial REs and SEPCHARs in elements with *mixed* content; unfortunately that would only be 99.8% backward-compatible. --Joe English jenglish@crl.com Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 14 Jun 1995 18:23:16 -0400 From: "Claude L. Bullard" \ Message-ID: <9506142223.AA44920@source.asset.com> References: <3rhh7l$2ru@reuters2.mitre.org> <19950614T132145Z@naggum.no> Subject: Re: Purpose of SGML in 3 paragraph DTD revision and instance update, is as Erik points out, a daunting task. The same can be said of most database design. Good design is one way to keep this from happening too often. That means apply proven methods or tools to the design process. Not enough is published on this subject for SGML as it is for other notations and database types. There are many methodologies for designing relational databases (e.g., IDEF, etc). These methodologies usually have some period of iterative refinement before the final product is loosed on the unwitting but suspicious end-user community. Tools to support such design tasks are appearing for SGML just as they did for relational systems. Extending these to more abstract concepts such as HyTime semantic networks (in one incarnation, topic maps) is desirable. As more information goes online, it becomes more difficult to manage the relationships among the different document types, but it is just as necessary as maintaining the relationships within them. Hypermedia link validation (does the target exist/is it a valid target for this semantic) is not only useful, it is mandatory. Macro facilities would help as would classing mechanisms with more than one level of subclassing. Erik is right that the need for external engines to do this is awkward as they are seldom shareable and this hampers SGML systems development owing to the need to constantly recreate that layer. Let us hope that HyTime improves enough to help more. Some standard solution is better than all of the ad hoc ones out there. I hope that some of the toolkits being advertised address this as well. What became of POEM? DTD creation and management is difficult because, as Erik points out correctly, we are trained to think in terms of a DTD singularity (Good concept. Copyright it, Erik). Writing multiple DTDs for the same information usually requires that we store the information multiple times (undesirable for all the *normalizing* reasons), or store the information once and use external independent links (semantic networks) to impose alternative views much as connector records in relational systems do, or specify transforms which rewrite the information. We may discover the need to do all of these. Optional element types are another way, but that is in a sense only putting alternative DTDs into the same DTD. It is convenient, but when one uses tools that are driven by the DTD, the looseness can be self-defeating. We get the tool to aid the author, use the DTD and the tool to express the intent of the information designer, then give the author a lot of conflicting choices that they have to sort out which eventually weakens the ability of the software to do the job we bought it for. It is inevitable in that information environments are messy, but that is small solace to the author who thinks the information designer didn't do a good job, or the SGML just isn't capable. This is where the HTML design is most aggravating. It is rather flat and that gives one few clues about using it. It is the problem with most attempts to write universal DTDs and another example of the singularity issue. Declaration subsets are one approach to tightening a DTD by forcing the editing software to only "see" designated subtrees. Dave Petersen pointed this out some months ago on CTS. As Kent Summers notes, one can use a stylesheet to impose other organizations, but this appears on the surface to be an external link in a proprietary language to tie in processing. That processing may be similar to the DSSSL concept in which an instance written to one DTD is transformed into another to provide a new semantic. The only problem with the approach is lack of standarization. Otherwise, it and the HyTime independent link which can point to the processing construct are essentially the same concept. The link is either pointing to the information and to the program that rewrites it, or it is pointing to the information and subsets it for the display. Please, correct me if I am wrong about that. Using the link to point to a query is a similar concept with the advantage of dynamism and the potential to use the DTD much as the relational system uses a catalog. Two obstacles in HyTime are the fixed anchorroles in the links and the fixed lists in the FCSs. One use of a function or query could be to allow the system to ask the catalog/schema/DTD if and which generic identifiers or other properties exist to be link targets based on the evaluation of the query or function parameters. Again, standardization is an issue and one hopes the work in WG8 this summer gets this in better shape. As the query issues are settled, definitions for passing these across the server interfaces is ripe for creative discussion. I hope this becomes a subject worth some hot debates on CTS. The ODBC equivalent for document databases is the next frontier. As to glitszy interfaces, I like 'em where they don't interfere with the work. I like 'em when they use the SGML that expresses the intent of the information designer to make it easier for an author to satisfy the intentions. I dislike them when they only serve to make the finger fat so that the wedding ring offered by the vendor courting the customer can't be removed. Like kids, information is a poor hostage for a failed marriage. Len Bullard -- Don't trust anyone under 5 foot 3. From: suthers+@pitt.edu (Daniel D Suthers) Newsgroups: comp.text.sgml Subject: SGML --> Database for statistical analysis Date: 14 Jun 1995 22:41:56 GMT Organization: LRDC, University of Pittsburgh Message-ID: <3rnojk$o6f@usenet.srv.cis.pitt.edu> Reply-To: suthers+@pitt.edu X-Newsreader: TIN [version 1.2 PL2] Greetings, I'd like to hear from anyone with experience in extracting data from an SGML-tagged document for statistical analysis, or for storage in a relational database system for future queries and statistical analysis. Intended application: statistical analysis of coded discourse corpora, for the sake of testing linguistic hypotheses. (My colleagues currently enter the coded data by hand; we want to have a GUI for sgml tagging and automatic dumping into a database.) To give you an idea of the units of interest, queries might ask thing like "what percentage of occurences of the marker 'however' is before statements of type X?". Interested in how you did it, what to avoid, etc. Especially interested in those who used databases or statistical software that is likely to be available to academic researchers (e.g., public domain or commonly site-licensed software). n.b. (1) I know about n.b. (2) I'm NOT asking about how to store an SGML document itself in a database, though we will need to do that too. (Any public-domain support for that?) The source document need not be recoverable from the database. The closest post I could find in recent comp.text.sgml was ... ---------- Article: 7162 of comp.text.sgml Newsgroups: comp.text.sgml Date: 14 Dec 1994 23:35:42 UT From: Jeffrey McArthur \ Organization: ATLIS Publishing Message-ID: <3dd113$2eb@news1.delphi.com> Subject: Unusual use of SGML ... I had three documents that were in a proprietary typesetting format. I needed to convert the documents into a relational database. I did this by way of SGML. ... ---------- -- ================================================================ Dan Suthers | Learning Research & Development Center suthers+@pitt.edu | University of Pittsburgh (412) 624-7036 voice | 3939 O'Hara Street (412) 624-9149 fax | Pittsburgh, PA 15260 (412) 363-3992 home | http://www.pitt.edu/~suthers/ ================================================================ Newsgroups: comp.text.sgml From: jjr@panini.att.com (Jeff Rocca) Subject: SGML conversion utilities Message-ID: \ Sender: news@nntpa.cb.att.com (Netnews Administration) Organization: AT\&T Bell Laboratories Date: Wed, 14 Jun 1995 22:47:15 GMT I'm looking for UNIX-based utilities available in the public domain to convert various formats to and from SGML (e.g., SGML-to-LaTeX and LaTeX-to-SGML). Any pointers? Thanks. Jeff Rocca jjr@pez.att.com From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 15 Jun 1995 00:00:59 GMT Organization: Naggum Software; +47 2295 0313 Expires: 29 Jun 1995 00:00:59 GMT Message-ID: <19950615T000059Z@naggum.no> References: \ <3rl9ra$18v@news1.delphi.com> <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> [Joe English] | It would be very useful to ignore all initial REs and SEPCHARs in | elements with *mixed* content; unfortunately that would only be 99.8% | backward-compatible. do you mean mixed as in _both_ elements and PCDATA ("mixed content" is an unfortunate term in the standard that means "has PCDATA in it somewhere"), or in contrast to "declared content" like CDATA and RCDATA (which, by the way, I urge people not to use, but that's not quite the same as treating them as non-existent :). #\ -- don't trust anyone under 30. From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 14 Jun 1995 17:50:23 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3ro04f$ql5@crl.crl.com> References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> Erik Naggum \ wrote: >[Joe English] >| It would be very useful to ignore all initial REs and SEPCHARs in >| elements with *mixed* content; unfortunately that would only be 99.8% >| backward-compatible. > >do you mean mixed as in _both_ elements and PCDATA ("mixed content" is an >unfortunate term in the standard that means "has PCDATA in it somewhere"), >or in contrast to "declared content" like CDATA and RCDATA (which, by the >way, I urge people not to use, but that's not quite the same as treating >them as non-existent :). What I meant was elements whose content model contains both #PCDATA and subelements, and is not a repeatable OR group... but now that I think about it, that wouldn't be all that useful after all since there are plenty of other reasons why such content models are a bad idea. Nevermind... --jenglish@crl.com From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: CALS table processing Date: 15 Jun 1995 02:12:38 GMT Organization: ACM Network Services Message-ID: <3ro4um$22p@hopper.acm.org> References: <3rjrfb$n65@newsgate.dircon.co.uk> Reply-To: davep@ACM.ORG In article <3rjrfb$n65@newsgate.dircon.co.uk>, bruce@sgml.dircon.co.uk (Bruce Hunter) writes: > >Does anyone know if there exists a rigourous, authorative description of >the intended CALS table processing model? I have MIL-M-28001B and the SGML >Handbook, but they are not specific enough about many areas. I want enough >detail to know how to write a CALS table handler for an SGML Viewer, such >that it can correctly handle all the various permutations possible in the >CALS table DTD. Without a detailed descrition of the implied semantics >this will be impossible. Does such a thing exist? Or does an "official" >FOSI fragment exist which shows completely how to process tables? No such luck. Volunteers were burned out and DOD wouldn't cough up $$$. >Product vendors who claim support for CALS tables all currently support >restricted subsets. Try moving the same moderately complex CALS table >between Softquad, Incontext, Arbortext and Grif products, and watch the >results. That's the price we pay. > I want to create something which will support CALS tables in the >way in which the authors of the DTD intended, but that is impossible >without a very detailed description of what was intended. If such a >description of a processing model does not exist, is anyone willing and >able to try to answer specfic questions as they arise? I'll give it a try. I was on the small subcommittee that originally came up with the model back in '89. >There are many things which are unclear or undefined. Some of the ones >currently bugging me are : > >1) Is there only one name space for colnames and spannames throughout a > table, or are they specific to a tgroup, thead, tfoot, entrytbl, etc? > I would guess that they all share the same name space, as spanspecs > referenced from an entry in a thead are subelements of the parent > tgroup and are not local to the thead. But should it then be an error > to have a colspec in an entrytbl, say, with the same name as a colspec > or spanspec in the surrounding table? Or are entrytbls something > different again, with their own local namespace? The original intent was to have the tgroup be the namespace container. Nobody thought much about entrytbl. Making each entrytbl a separate namespace seems consistent. >2) Is this valid : > > \ > \ > \ > \ > \ > \ > > If so, to which column does the second colspec relate? Or is it an > error if at least one colspec has a colnum specified and others do > not? Would re-arranging the above to : > > \ > \ > \ > \ > \ > \ > > be valid? Should colspecs with colnums appear in the order of their > colnums or is this an error? The original expectation was that most formatting would be specified by table styles rather than explicit overrides. The original use for colspecs was for column skipping and spanning; colnum was REQUIRED. We expected only a few tables would even need to have any colspecs. Later Somebody suggested that colnum could be inferred sequentially, so it was made IMPLIED; the implied column number would be one greater than the previous. Having two colspecs for the same column was intended to be an error, regardless of whether their column number was explicit or inferred. If you don't use tabular styles (shame on you!) then you need colspecs all over the place, and the implicit numbering becomes more of a convenience. >3) Why does the valign attribute on entry have a default value? This > makes the valign attribute on thead, tfoot, tbody redundant. All formatting attributes should have been IMPLIED; some were overlooked. If they're not IMPLIED, then they always override the specified style sheet; this is stupid. (Most applications rightly can't tell the difference between an explicitly specified attribute value and a default.) >I have lots more questions for anyone willing to try to take them on. >Thanks to anybody who can assist. Will get them when and as I have time. Dave Peterson SGMLWorks! davep@acm.org From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: CALS table processing Date: 15 Jun 1995 03:09:43 GMT Organization: \ Message-ID: <3ro89n$ek2@news1.delphi.com> References: <3rjrfb$n65@newsgate.dircon.co.uk> X-Newsreader: Galahad 1.1b Bruce Hunter wrote: > be valid? Should colspecs with colnums appear in the order of their > colnums or is this an error? Ah, the joy of tables. I have never seen a table spec that I really liked. I have done a lot with the SoftQuad table model (very similar to CALS) but it has its oddities. There was a discussion of this a while back. Perhaps you should browse the archive for some of the discussions in the past. But using your example as a basis, consider the following: \ \ \ \ \ \ This makes things really interresting. The number of columns is supposed to be three, but there are four specifications. Other interresting problems occur if you specify and define four columns and then in one of the rows have five separate cells of data. What is a formatting engine to do? What I did was write a simple little Lex program that scans through the table and checks for situations that can cause problems in the typesetting engine. I incorporate that into our editing application. ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: melhaven@mail.rcinet.com Newsgroups: comp.text.sgml Subject: Re: SGML --> Database for statistical analysis Date: 15 Jun 1995 03:52:50 GMT Organization: Dayton Microcomputer Association; Dayton, Ohio, USA Message-ID: <3roaqi$hkt@sally.dma.org> References: <3rnojk$o6f@usenet.srv.cis.pitt.edu> Reply-To: melhaven@mail.rcinet.com X-Newsreader: IBM NewsReader/2 v1.09 In <3rnojk$o6f@usenet.srv.cis.pitt.edu>, suthers+@pitt.edu (Daniel D Suthers) writes: >Greetings, > >I'd like to hear from anyone with experience in extracting data from an >SGML-tagged document for statistical analysis, or for storage in a relational >database system for future queries and statistical analysis. > >Intended application: statistical analysis of coded discourse corpora, for the >sake of testing linguistic hypotheses. (My colleagues currently enter the >coded data by hand; we want to have a GUI for sgml tagging and automatic >dumping into a database.) To give you an idea of the units of interest, >queries might ask thing like "what percentage of occurences of the marker >'however' is before statements of type X?". > >Interested in how you did it, what to avoid, etc. Especially interested in >those who used databases or statistical software that is likely to be >available to academic researchers (e.g., public domain or commonly >site-licensed software). > >n.b. (1) I know about n.b. (2) I'm NOT asking about how to store an SGML >document itself in a database, though we will need to do that too. (Any >public-domain support for that?) The source document need not be recoverable >from the database. > >The closest post I could find in recent comp.text.sgml was ... >---------- >Article: 7162 of comp.text.sgml >Newsgroups: comp.text.sgml >Date: 14 Dec 1994 23:35:42 UT >From: Jeffrey McArthur \ >Organization: ATLIS Publishing >Message-ID: <3dd113$2eb@news1.delphi.com> >Subject: Unusual use of SGML >.... >I had three documents that were in a proprietary typesetting format. I >needed to convert the documents into a relational database. I did this by >way of SGML. ... >---------- > >-- >================================================================ >Dan Suthers | Learning Research & Development Center >suthers+@pitt.edu | University of Pittsburgh >(412) 624-7036 voice | 3939 O'Hara Street >(412) 624-9149 fax | Pittsburgh, PA 15260 >(412) 363-3992 home | http://www.pitt.edu/~suthers/ >================================================================ - - If you have a specific document and you can create the DTD, tags can be incorporated into the document as a data feed. For example a contract written in SMGL could have tags for the line item, description, unit of issue, contractor, dollar value, etc. The information so tagged can be used to update a data base. The important thing is to have specific tags to use as identifiers for a data feed to other systems. Regards, Ken Newsgroups: comp.text.sgml From: marq@world.std.com (Mark A Lilly) Subject: Re: SGML --> Database for statistical analysis Message-ID: \ Organization: The World Public Access UNIX, Brookline, MA X-Newsreader: TIN [version 1.2 PL2] References: <3rnojk$o6f@usenet.srv.cis.pitt.edu> Date: Thu, 15 Jun 1995 05:01:26 GMT I read about a product called LivePage out of Ontario, written by the man who created the Watcom database engine, if that's important info to folks. It purports to me able to take SGML documents and store them, extremely well indexed, in a SQL database. I am awaiting info, and if anyone's interested, i can convey more. Cheers, Mark -- From: churchyh@doc.cc.utexas.edu (Henry Churchyard) Newsgroups: comp.text.sgml Subject: Re: Web Notations Date: 15 Jun 1995 01:10:49 -0500 Organization: The University of Texas at Austin; Austin, Texas Message-ID: <3roit9$ce6@doc.cc.utexas.edu> References: <9506141830.AA45783@source.asset.com> Yes, when you click on a link, the HTTP server which delivers the requested document also returns a header including a MIME "Content-Type: something/something" declaration, which the browser uses to decide which application will interpret the document (if the content type is "text/html", then the browser itself will interpret the incoming document, of course). -- _______ .------|_______ Churchyard/Orr || Henry Churchyard ------+ _______ Genealogy Web Site || University of Texas `------|_______ http://uts.cc.utexas.edu/~churchh/genealgy.html From: dmatthewson@cix.compulink.co.uk (David K Matthewson) Newsgroups: comp.text.sgml Subject: Browser for HTM or SGML files on a LOCAL PC Date: 15 Jun 1995 09:05:14 +0100 Organization: Matthewson Associates Sender: news@news.demon.co.uk Message-ID: <3ropjq$hmv@imp.demon.co.uk> Reply-To: dmatthewson@cix.compulink.co.uk X-Nntp-Posting-Host: spol-lon.demon.co.uk X-Newsreader: Forte Free Agent v0.55 Does anyone know of a browser for use in a non net environment - ie: for viewing local files of a disc? I know this can be done with Netscape etc by using Nullsock but then a user has lots of 'dead' controls. TiA David DMATTHEWSON@CIX.COMPULINK.CO.UK FAX: +44-(0)-1580-200493 ------------------------------- Newsgroups: comp.text.sgml From: Phil.Kime@ed.ac.uk (Phil Kime) Subject: Any comments on SGML Enabler extension for QuarkXPress? Message-ID: \ Followup-To: comp.text.sgml X-Attribution: Phil X-Www: http://www.cogsci.ed.ac.uk/~philkime/home.html Sender: usenet@cogsci.ed.ac.uk (C News Software) Reply-To: Phil.Kime@ed.ac.uk (Phil Kime) Organization: Centre for Cognitive Science X-Comments: MIME/PGP welcome X-Newsreader: Emacs 19.27.2 (GNUS 4.1) Date: Thu, 15 Jun 1995 09:02:27 GMT has anyone had any experience with the SGML Enabler extension for QuarkXPress? Any comments? Is it worth the $495? p -- = Phil Kime (Phil.Kime@ed.ac.uk) = = Centre for Cognitive Science/Dept of Philosophy = = Edinburgh University = From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 15 Jun 1995 10:47:45 GMT Organization: Naggum Software; +47 2295 0313 Expires: 29 Jun 1995 10:47:44 GMT Message-ID: <19950615T104745Z@naggum.no> References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> <3ro04f$ql5@crl.crl.com> [Joe English] | What I meant was elements whose content model contains both #PCDATA and | subelements, and is not a repeatable OR group... but now that I think | about it, that wouldn't be all that useful after all since there are | plenty of other reasons why such content models are a bad idea. well, what if #PCDATA was treated as an element in its own right, except that it had no start- or end-tag, such that the treatment that is now given to the entire content model would only be given to those situations where the content models explicitly allows #PCDATA? that is, if #PCDATA is not a possible "element" at an given point, the rules for element content (i.e., ignore white space) applies. as we learned from HyTime and DSSSL, SGML's undelimited #PCDATA is just a very convenient shorthand notation for an element that contains only data. HyTime calls it a "pseudo-element", and can address pseudo-elements instead of just elements. this shorthand is of course necessary in keyed-in text, but might as well be a separate element with machine generated SGML documents. hate to say this, but ODA actually got this one thing right. I think the reasons only repeatable OR groups with #PCDATA (including the singleton OR group :) are a good idea is that the RE/RS stuff is getting in the way, and once we remove the bad side effects of RE/RS, so does the restriction on repeatable OR groups. are there other reasons? #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: s.rahtz@elsevier.co.uk (Sebastian Rahtz) Newsgroups: comp.text.sgml Subject: Re: Blessings and Curses Date: Thu, 15 Jun 1995 10:51:32 Organization: Elsevier Science Ltd Message-ID: \ References: <9506092043.AA37374@source.asset.com> X-Newsreader: Trumpet for Windows [Version 1.0 Rev A] In article <9506092043.AA37374@source.asset.com> "Claude L. Bullard" \ writes: >Instead, a lone college student and some friends dictated our HTML future. I think this is historically a rather strange view. HTML came from CERN, from professional computer scientists employed by a very large international research organization. Marc Andreessen and NCSA may have made it popular and useable with Mosaic, but they didnt invent HTML, or substantially alter its character. Sebastian Sebastian Rahtz s.rahtz@elsevier.co.uk Production Methods Group +44 1865 843662 Elsevier Science Ltd Kidlington Oxford, UK From: harvey \ Newsgroups: comp.text.sgml Subject: Re: Web Notations Date: 15 Jun 1995 11:33:41 GMT Organization: Electronic Commerce Connection Message-ID: <3rp5ql$jj4@news.cais.com> References: <9506141830.AA45783@source.asset.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) "Claude L. Bullard" \ wrote: >In SGML, of course, we have notation declarations that indicate the >notation and version which can be used as a registry for the servers >required to support the document. The DTD in effect, plays the role of the >catalog for types that must be supported and any foreign notations. >Presumably, in due course, a standardized layer will evolve that should >allow these notations to interoperate via a shared protocol. Is that >protocol HTTP? > >As this is an issue of some importance to the evolution of the document >database communications layer standardization for hypermedia, I am curious >how the Web achieves this and what if any plans others have for supporting >it. I ask this question on CTS because there is a fair chance that others >may be dealing with this. For HTML or VRML only users, it may not be an >issue yet. > Len: Glad to see you back |-). I am not sure I understand your questions completely but what I think you are asking does HTML have the capability of referencing and delivering other types of data, such as PDF, sound, video, pure SGML. The answer is yes. MIME is used to tell the application what type of data it is receiving to allow it to handle arbitrary data types. I have been reading about HotJava which if I understand what it is meant to do, will take HTML/HTTP/Browser a step forward and allow applications within the browsers (can someone correct me if I am wrong). I have been hearing these wonderful things about HotJava but being a visual person, I am having a hard time conceptualizing exactly what it is supposed to do. Unfortunately, we are only runing SunOS4.1.3 instead of Solaris so I can't even try it for myself. Betty <><><><><><><><><><><><><><><><><><><><><><><><><><><><><> Betty Harvey | PERSONAL David Taylor Model Basin | harvey@eccnet.com harvey@navysgml.dt.navy.mil | ---------------------------------------------------------- Navy CALS WWW Server: http://navysgml.dt.navy.mil <><><><><><><><><><><><><><><><><><><><><><><><><><><><><> From: tad@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 15 Jun 1995 12:18:06 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <3rp8du$7in@cliffy.lfwc.lockheed.com> References: \ <3rl9ra$18v@news1.delphi.com> <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> X-Newsreader: TIN [version 1.2 PL1] Erik Naggum (erik@naggum.no) wrote: : do you mean mixed as in _both_ elements and PCDATA ("mixed content" is an : unfortunate term in the standard that means "has PCDATA in it somewhere"), My understanding of paraphrasing "mixed content" is "has PCDATA in it somewhere along with at least one other element". So, \ would not be "mixed content" in my interpretation. Do you agree? -- --------------------------------------------------------------------- | | Tad McClellan, Logistics Specialist (IETMs and SGML guy) | Lockheed Martin Tactical Aircraft Systems | P.O. Box 748 MZ 1025 | Fort Worth, TX 76101 | | voice: (817) 763-6314 | email: mcclellantj@lfwc.lockheed.com | --------------------------------------------------------------------- | | All I want, is a little more than I'll ever get. | --------------------------------------------------------------------- From: tad@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 15 Jun 1995 12:20:13 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <3rp8ht$7in@cliffy.lfwc.lockheed.com> References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> <3ro04f$ql5@crl.crl.com> <19950615T104745Z@naggum.no> X-Newsreader: TIN [version 1.2 PL1] Please ignore my last post. I was sleep reading and didn't check all the responses first. Sorry. -- --------------------------------------------------------------------- | | Tad McClellan, Logistics Specialist (IETMs and SGML guy) | Lockheed Martin Tactical Aircraft Systems | P.O. Box 748 MZ 1025 | Fort Worth, TX 76101 | | voice: (817) 763-6314 | email: mcclellantj@lfwc.lockheed.com | --------------------------------------------------------------------- | | Just be thankful we don't get all the government we pay for! | --------------------------------------------------------------------- From: blstewar@ingr.com (Bobbi Stewart) Newsgroups: comp.text.sgml Subject: Southeast SGML User's Group Meeting Date: 15 Jun 1995 13:15:24 GMT Organization: Intergraph Corporation Message-ID: <3rpbpc$2aj@b30news.b30.ingr.com> _____________________________________________ T H E S G M L B U F F E T Delivered by the Southeast SGML User's Group __________________________________________ A P P E T I Z E R SGML Overview 8 a.m. - 9 a.m. Presented by Beth Micksch, Intergraph C H O I C E O F E N T R E E S Management Track: "Making the SGML Decision" "The Impact of SGML on Your Organization" Technical Track: "Building Your SGML Team" "Exploring Document Analysis" An interactive session for beginning SGML developers D E S S E R T Wrap-up S E R V E D O N L Y 8 a.m. - 12 noon L O C A T I O N Intergraph Corporation, bldg. 20 Huntsville, Alabama D A T E Tuesday, July 11, 1995 _______________________________________________ Registration Information: Deadline July 7, 1995 Entry Fee: $4.00 members, $7.00 non-members Mail the following information and entry fee to: Beth Micksch Intergraph Corporation Huntsville, AL 35894 Mailstop: GD3006 OR FAX: (205) 730-3240, attention Beth Micksch Email: bmicksch@ingr.com Name: ________________________________________ Title: _________________________________________ Organization: __________________________________ Address: ______________________________________ Phone: ________________________________________ FAX: __________________________________________ For more information about the Southeast SGML User's Group, contact Sandra Marquez (205) 730-3408 or email samarque@ingr.com. ________________________________________________ Southeast SGML User's Group is a non-profit organization. Newsgroups: comp.text.sgml From: wwb@world.std.com (bill w bunnell) Subject: Synthaesis by AEsthetex - Tom Lehr's ? Message-ID: \ Organization: The World Public Access UNIX, Brookline, MA Date: Thu, 15 Jun 1995 13:38:34 GMT Tom Lehr, In an earlier posting to Newsgroups: comp.text.sgml you described and noted as follows: >Some months ago I attended a demonstration by a Scottish company,< >AEsthetex, of their SGML publishing software, Synthaesis. The software< >appeared to be a native SGML editor that operated like a database and< >could produce high-quality typeset output. It seemed to address many of< >the areas of traditional print publishing that have been ignored by other< >companies producing SGML applications, for example, tracking & managing< >changes to documents and providing formatted output. However, I have not< >found any reference to this company or its products in any of the SGML< >areas of the Internet.< and then asked, and commented as follow: >Does anyone have any knowledge or experience of< >this software? It seems odd to me that there would be SGML software on< >the market that nobody was discussing here.< Steve Pepper ( http://www.falch.no/) then responded as follow: >As far as I know, the software is still some way from completion (at< >least the SGML part of it; Synthaesis is also a general typesetting< >system). I haven't had the chance to look at it yet but it certainly< >sounds like a very interesting product.< For more details, get in touch with< >AEsthetex< >626 Lanark Road< >Juniper Green< >Edinburgh EH14 5EW< >UK< >Tel. +44-131-458 3933< >Fax +44-131-458 3903 < >email: angus@aesthetex.co.uk < Allow me to add the following: The development of Synthaesis does continue. Several US installations are in-process, and not yet able to be discussed in this forum. Commentary, earlier in time, did appear in the Seybold Reports. Not a reference often found here. Reflection on your question as to absence of commentary here suggests that others have not had the early opportunity you had to see the software. Probably because their needs/curiosity were more general, and not as well researched. Suffice it to say that while experience is still limited, it does, as you, Mr. Pepper, and others to whom it has been shown have noted, it does offer some interesting solutions to "problems" not as well addressed by products which receive more frequent comment here. Your comment brings to mind the assumptions underlying the early failures of telephone polls in predicting political elections. Also, in the US, for more details about Aethetex / Synthaesis can be found through E-Mail at: wwb@world.std.com or voice at (203) 245-2808 WWB From: suthers+@pitt.edu (Daniel D Suthers) Newsgroups: comp.text.sgml Subject: Re: Web Notations Date: 15 Jun 1995 14:14:19 GMT Organization: LRDC, University of Pittsburgh Message-ID: <3rpf7r$ru8@usenet.srv.cis.pitt.edu> References: <9506141830.AA45783@source.asset.com> <3rp5ql$jj4@news.cais.com> Reply-To: suthers+@pitt.edu X-Newsreader: TIN [version 1.2 PL2] If I understand him correctly, Claude is NOT asking whether HTTP can decide what document type it got and know what to do do with it. Sure it can do that, but each document must be of one uniform type. He is asking whether it can deal with MULTIPLE types (notations) within a SINGLE document. Knowing that Hyper-G has a "cluster" concept whereby documents of different types are clustered into a single document, I forwarded his query to the Hyper-G list \. Here is the reply (I made it clear that this was Claude's prose, not mine, but they quote it as if it were mine): -------------------- enclosure -------------------- Date: Thu, 15 Jun 1995 10:17:03 +0200 From: Gerald Mesaric \ Message-Id: <199506150817.AA13846@pluto> To: suthers+@pitt.edu, bullardc@source.asset.com In-Reply-To: <199506150249.WAA01068@trench.lrdc.pitt.edu> (suthers+@pitt.edu) Subject: Re: (fwd) Web Notations >>>>> "suthers" == suthers \ writes: suthers> A question for the Web experts: how does a Web document suthers> indicate that it contains or references multiple notations? suthers> I guess in other words, can HTTP indicate multiple types suthers> within one document? Further, if it can, how does it suthers> indicate which version or flavor of the type is referenced or suthers> contained? No, HTTP is NOT able to reference more than ONE document at the same time. The Hyper-G protocol can link to a "Cluster" of objects, supporting the simultanious display of multiple multimedia documents and als to provide multilinguality. suthers> For example, one may need a document instance in which HTML suthers> has been used for some parts of the text, another SGML suthers> application was used to create queriable tables, VRML is used suthers> to create alternative visualizations for data in the tables, suthers> and PDF is used to create non-editable images of copywritten suthers> material. These are just examples of course. Let us not suthers> debate the suitability of each notation for the application, suthers> please. In Hyper-G you just have to put all documents that you want to display at the same time, into one Cluster. A Cluster is a set of documents. With special flags that you can add to the documents, you are able to specify, what within the cluster should be displayed simultaniously, what language-independant and what not. suthers> In SGML, of course, we have notation declarations that suthers> indicate the notation and version which can be used as a suthers> registry for the servers required to support the document. suthers> The DTD in effect, plays the role of the catalog for types suthers> that must be supported and any foreign notations. suthers> Presumably, in due course, a standardized layer will evolve suthers> that should allow these notations to interoperate via a suthers> shared protocol. Is that protocol HTTP? We -- at the Hyper-G protocol -- use the standard "sgmls" Parser to parse HTF (a SGML compliant), therefore it would -- and will be -- possible to use any kind of SGML (and send the DTD over the network with the document). suthers> As this is an issue of some importance to the evolution of suthers> the document database communications layer standardization suthers> for hypermedia, I am curious how the Web achieves this and suthers> what if any plans others have for supporting it. I ask this suthers> question on CTS because there is a fair chance that others suthers> may be dealing with this. For HTML or VRML only users, it suthers> may not be an issue yet. BTW, Hyper-G will support VRML in a couple of weeks (we are now in BETA state) and HTML3 by the end of September. You can access online information about Hyper-G at following URLs: \ \ \ Note, that these URLs all point to the same object in the Hyper-G database, a WWW and a Gopher gateway convert all Hyper-G data automatically at "run-time" into W3 (HTML) and Gopher formats. P.S.: I will append our first "Hyper-G Status Report" to this mail. Yours, Gerald. ---------------------------------------------------------------------------- Hyper-G Status Report Volume 1/95 ================================= 0. Introduction Hyper-G represents the first of a new generation of Internet information systems -- it provides real hypermedia, supporting tools for structuring, maintaining and serving heterogeneous multimedia data including text, images, digital audio and video, PostScript and 3D scenes. Hyper-G guarantees automatic hyperlink consistency and supports hyperlinks to and from multimedia documents, full text retrieval, a UNIX-like security system with user management and user groups, and client gateways to Gopher and World Wide Web (WWW) browsers such as Netscape, Mosaic and Arena. Hyper-G is fully compatible with current Internet technology and includes seamless access to popular Internet server technologies such as WWW and Gopher. Hyper-G clients such as Harmony and Amadeus access Hyper-G servers across the Internet, allowing users to view and manipulate information in multiple ways. Their advanced navigational tools help users orient themselves and avoid becoming "lost in hyperspace". 1. Hyper-G Server Technology The Hyper-G server concept guarantees link consistency and provides real hypermedia. This is only possible because links in Hyper-G are not stored within documents, but in a separate hyperlink database called the Database Server (DBServer). The data itself is also contained in an object database called the Document Cache Server (DCServer). The third part of the server concept is the full text server, which is responsible for the full text index. 1.1. Database Server (DBServer) o Hyperlink consistency o Structuring of information (collections, clusters) o Advanced search facilities o Multilinguality o UNIX-like security system with hierarchical user groups o Hyperlinks in multimedia documents o Accounting and billing concepts o Distributed client-server concept 1.2. Document Cache Server (DCServer) o Caching facilities o Storing and handling of multimedia documents o Gateway to WWW and Gopher for native HG clients (FTP planned) 1.3. Full Text Server (FTServer) o Generation of full text indices o Full text search facilities 2. Hyper-G Client Products 2.1. Amadeus (PC/MS-Windows 3.1x, 95 & NT) Amadeus is the native 32 Bit Windows/NT/95 client for Hyper-G. It provides full integration of Hyper-G's rich features and is able to display nearly every common type of multimedia data. Amadeus supports interactive authoring and point-and-click hyperlink creation in multimedia documents such as images and PostScript files. The features of Amadeus include: o 32 bit application (native Windows NT/Windows 95 application, also runs under Windows 3.1x using Win32s) o Full Hyper-G functionality (including authoring capabilities). o Multi-Protocol client, supports World-Wide-Web and Gopher protocols. (Note: HTML 3.0 (for WWW), FTP and WAIS support planned for fall) o Easy-to-use graphical user interface. o Integrated RTF->HTF converter lets you import text generated by most popular word processors. o Integrated viewers for Text (SGML), images, movies and PostScript. o External 3D-Scene viewer (currently Windows NT only) o Extensive online help. o Comes complete with 35-page user manual. o Is absolutely free software, may be copied and used without restrictions, even for commercial applications. o Stand-alone PostScript Viewer included in package o Demo (no edit functions) of local-database-Amadeus included (alpha-version) 2.2. Harmony (UNIX/X11) Harmony is the Unix/X11 client for Hyper-G, the first "second generation", publicly available, networked hypermedia information system running over the Internet. Harmony takes advantage of Hyper-G's rich information structuring facilities to provide innovative two and three-dimensional visualisations of the information space. It has extensive facilities for interactive authoring and point-and-click hyperlink creation in heterogeneous media, including 3D models and PostScript files. Harmony is also multilingual: its user interface adjusts dynamically to the language of first choice and documents available in multiple languages are selected in order of language preference. Harmony's specific feature set includes: o Hierarchical browsing of Hyper-G collections. o Attribute and content search with user-definable scope. o Graphical local map of incoming *and* outgoing hyperlinks, as well as parent-child and other relationships. o Location feedback through tightly coupled search, hyperlink, and collection browsing. o History browser with timeline of past interactive waypoints. o 3D information landscape. o File upload and save facilities. o Interactive structuring, authoring, and annotation facilities. o Interactive hyperlink editing between arbitrary document types. 3. Hyper-G, Gopher and WWW URLs Additional online information about Hyper-G \ \ \ Amadeus \ \ \ Harmony \ \ \ 4. Software License Hyper-G client software, including Amadeus and Harmony, is distributed for free over the Internet. The Hyper-G server binaries are free for non-commercial use. The source code will also be available. For detailed information on source code licenses and commercial usage of the Hyper-G server software, please contact Gerhard Pail \ at the Institute for Information Processing and Computer Supported New Media (IICM). 5. Administrivia Editor: Gerald Mesaric Proofreading: Vanessa Mayrhofer Authors of this issue: Keith Andrews, Thomas Dietinger, Gerald Pani, Joerg Faschingbauer, Gerald Mesaric. -------------------------------------------------------------------------- IICM Institute for Information Processing and Computer Supported New Media \ \ \ ---------------------------------------------------------------------------- Gerald Mesaric gmesaric@iicm.tu-graz.ac.at Institute for Information Processing and Computer Supported New Media (IICM) Graz University of Technology - Austria/Europe/Earth Voice: ++43/316/832551-11 Fax: ++43/316/824394 ---------------------------------------------------------------------------- ______ __ __ ______ ______ \\ \\ \\ \\ |_ | | _| _ _ __ ___ _ __ / ____) / / / / \\ \\ \\ \\ | `-' | \\ / | '_ \\ / _ \\ '__) ,---. | | ___ / / / / / / / / | ,-. |\\ \\/ /| |_) | __/ | `---' | |__/ | \\ \\ \\ \\ /_/_/_/ |__| |__|\\ / | .__/ \\___|_| \\______| \\_\\_\\_\\ /_/ |_| For information on Hyper-G look under ftp://iicm.tu-graz.ac.at/pub/Hyper-G ! ---------------------------------------------------------------------------- -- ================================================================ Dan Suthers | Learning Research & Development Center suthers+@pitt.edu | University of Pittsburgh (412) 624-7036 voice | 3939 O'Hara Street (412) 624-9149 fax | Pittsburgh, PA 15260 (412) 363-3992 home | http://www.pitt.edu/~suthers/ ================================================================ From: suthers+@pitt.edu (Daniel D Suthers) Newsgroups: comp.text.sgml Subject: GUI for DTD editing? Date: 15 Jun 1995 14:36:45 GMT Organization: LRDC, University of Pittsburgh Message-ID: <3rpght$ru8@usenet.srv.cis.pitt.edu> Reply-To: suthers+@pitt.edu X-Newsreader: TIN [version 1.2 PL2] GUIs for editing a document under a given DTD are coming out, but is anyone offering a GUI for writing the DTD themselves? -- ================================================================ Dan Suthers | Learning Research & Development Center suthers+@pitt.edu | University of Pittsburgh (412) 624-7036 voice | 3939 O'Hara Street (412) 624-9149 fax | Pittsburgh, PA 15260 (412) 363-3992 home | http://www.pitt.edu/~suthers/ ================================================================ From: Hans Voorbij \ Newsgroups: comp.text.sgml Subject: Renewal of my Url Date: 15 Jun 1995 14:43:30 GMT Organization: Faculty of Arts, Utrecht University, NL Message-ID: <3rpgui$fpd@ruulch.let.ruu.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1b3 (Windows; I; 16bit) Hello. Yesterday I sent a mail with my URL. Today they changed it to: http://www.let.ruu.nl/C+L/broeke/home.htm Please, feel free to have a look at my Web pages, now also all in English. Thank you, Nils From: "Nils ten Broeke, instead of Hans Voorbij" \ Newsgroups: comp.text.sgml Subject: Renewal of my URL, part two. Date: 15 Jun 1995 14:51:34 GMT Organization: Faculty of Arts, Utrecht University, NL Message-ID: <3rphdm$fpd@ruulch.let.ruu.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1b3 (Windows; I; 16bit) Hello. Because of working on a different computer a mistake occured. My new URL is: http://www.let.ruu.nl/C+L/broeke/home.htm Please, feel free to have a look at my Web pages, now also all in English. By the way, my name is Nils ten Broeke and not Hans Voorbij. Thank you. From: ht@chestnut.ling.upenn.edu (Henry S. Thompson) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 15 Jun 1995 14:57:00 GMT Organization: University of Pennsylvania Message-ID: \ References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> <3ro04f$ql5@crl.crl.com> <19950615T104745Z@naggum.no> In-reply-to: erik@naggum.no's message of 15 Jun 1995 10:47:45 GMT I think this is pretty critical, and I want to be clear on what the consequences of Erik's suggestion are. Suppose I have something like \ \ . . . \

\ This is a test \ how many pseudo-elements in this \ p \ element? \ I do declare \ \

At the moment this actually works pretty well, and we get only the five pseudo-elements we expect. Will this still be the case under Erik's proposal? More importantly, will it fix the following infelicity, which consistently fools most observers, but (for better or worse) not most parsers: \ \ \ \

This is not valid! (The RE after the quote tag above forces the PCDATA branch of the disjunction). \

\ \\

But this is OK! (note you must keep the following to end tags on the same line as well) \

\ Getting this fixed would be a DEFINITE step forward, and would not invalidate previously valid documents, just make some perplexingly invalid documents valid. ht -- Henry S. Thompson, Human Communication Research Centre, University of Edinburgh Visiting Fellow at Institute for Research in Cognitive Science University of Pennsylvania, 3401 Walnut Street, Philadelphia, PA 19104-6228 Phone: (+1) 215 898-0366 Fax: (+1) 215 573-9247 Email: ht@unagi.cis.upenn.edu From: donturn@cc.gatech.edu (Don Turnbull) Newsgroups: comp.text.sgml Subject: Re: digital publishing model Date: 15 Jun 1995 18:01:42 GMT Organization: Georgia Tech Message-ID: \ References: <3rjf0u$2hu@mordred.cc.jyu.fi> X-Newsreader: Value-Added NewsWatcher 2.0b24.0+ I'm very interested in your digital publishing model search. I'm working on my Master's thesis related to taht subject with a methodology I'm calling Object Oriented Information Design- it will be a set of protocols and tools to help publish diverse sources (and types) of information. I'd like to hear more about what you're found and maybe we can start an ongoin exchange of information. Thanks, Don Turnbull Georgia Institute of Technology -- Don Turnbull donturn@cc.gatech.edu http://www.gatech.edu/lcc/idt/Students/donturn/donturnhome.html From: donturn@cc.gatech.edu (Don Turnbull) Newsgroups: comp.text.sgml Subject: Re: SGML --> Xerox Metacode Date: 15 Jun 1995 18:06:15 GMT Organization: Georgia Tech Message-ID: \ References: <3rn113$cvj@jupiter.planet.net> X-Newsreader: Value-Added NewsWatcher 2.0b24.0+ A company called Image Sciences in Dallas, Texas has a program called DCF/PLUS that does exactly that. I used to work on that product there, it's quite good. Call Image Sciences as (214)891-6500 tell them Don Turnbull recommended you (I don't get anything, but they'll get a kick out of hearing me referring people to them.) -- Don Turnbull donturn@cc.gatech.edu http://www.gatech.edu/lcc/idt/Students/donturn/donturnhome.html From: ewong@ctp.com (Eddy Wong) Newsgroups: comp.text.sgml Subject: Re: GUI for DTD editing? Date: 15 Jun 1995 19:27:52 GMT Organization: Cambridge Technology Partners Message-ID: <3rq1jo$rv0@toon.ctp.com> References: <3rpght$ru8@usenet.srv.cis.pitt.edu> In article <3rpght$ru8@usenet.srv.cis.pitt.edu>, suthers+@pitt.edu (Daniel D Suthers) writes: |> GUIs for editing a document under a given DTD are coming out, but is anyone |> offering a GUI for writing the DTD themselves? |> |> -- |> ================================================================ |> Dan Suthers | Learning Research & Development Center |> suthers+@pitt.edu | University of Pittsburgh |> (412) 624-7036 voice | 3939 O'Hara Street |> (412) 624-9149 fax | Pittsburgh, PA 15260 |> (412) 363-3992 home | http://www.pitt.edu/~suthers/ |> ================================================================ Have you checked Microstar's Near & Far? Eddy. -- /\\/\\ | Eddy Wong ewong@ctp.com | / /_.\\| Cambridge Technology Partners, Inc. (617) 374-9800 | \\ /./| 304 Vassar St, Cambridge, MA 02139 (617) 374-8300 (fax) | \\/\\/ | | From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 15 Jun 1995 12:29:44 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3rq1n8$o6j@crl2.crl.com> References: \ <19950615T000059Z@naggum.no> <3ro04f$ql5@crl.crl.com> <19950615T104745Z@naggum.no> Erik Naggum \ wrote: >well, what if #PCDATA was treated as an element in its own right, except >that it had no start- or end-tag, such that the treatment that is now given >to the entire content model would only be given to those situations where >the content models explicitly allows #PCDATA? that is, if #PCDATA is not a >possible "element" at an given point, the rules for element content (i.e., >ignore white space) applies. That might interact weirdly with delimiter recognition modes. The current rules work like USEMAP [*], where the parser uses the same mode for the whole element. The suggested rules would work more like DATATAG, where the mode changes at different points in the content model. The presence of DATATAG is part of the reason why the content model ambiguity rule is necessary, though I don't think this change would have the same effect since there are only two modes to choose from, not an arbitrary set of delimiters. [*] that is, the \ form that appears in the DTD, not the \ form that appears in the document instance. Let's see... \ \ \

\... \ text \

\

\...\ text \

\

\... \\... \\

\

\...\\... \ \

\

\... \ \... \\

\

\...\ \... \\

The last two lines would be illegal (currently all but the first and third are illegal); so I guess it would still be *possible* to create content models that cause inexplicable errors, just a little bit harder. >I think the reasons only repeatable OR groups with #PCDATA (including the >singleton OR group :) are a good idea is that the RE/RS stuff is getting in >the way, and once we remove the bad side effects of RE/RS, so does the >restriction on repeatable OR groups. are there other reasons? That's just a design principle that I try to stick to. For any elements with a structure more rigid than "text with some marked-up subphrases" (i.e. a repeatable OR group with #PCDATA), I figure that if character data can only appear at certain points in the content model, then the PCDATA elements at those points must have some contextual significance and deserve to have their own GI. (Of course, I only came up with that design principle after being perplexed by the effects of mixed-content delimiter recognition mode, but I think it's a good one nevertheless :-) --Joe English jenglish@crl.com From: joeclark@hookup.net (Joe Clark) Newsgroups: comp.fonts,comp.text.desktop,comp.text.frame,comp.text.sgml Subject: NEW! Typetalk mailing list on typography and graphic design Date: Thu, 15 Jun 1995 15:25:13 -0500 Organization: HookUp Communication Corporation, Oakville, Ontario, CANADA Message-ID: \ [Revised 950615] ANNOUNCING THE TYPETALK MAILING LIST ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There now exists an Internet mailing list dedicated to all manner of discussions related to typography and graphic design. Note that this list is interested primarily in the ARTISTIC aspects of graphic design. The list is unmoderated, and no one will censor what you say, but questions of the "Why won't my TrueType font print on my HP LaserJet?" and "Why won't SuperATM work with \?" sort are best directed to the newsgroup comp.fonts. Now that that's out of the way, feel free to post any kind of question or comment about any aspect of graphic design, past, present, or future. Examples: € Discussions of particularly successful or unsuccessful designs, designers, graphic movements, typefaces, and the like (don't hold your feelings back!) € Questions about the "historical accuracy" of various fonts € Commentary on the graphic-design press (the hardcopy kind-- _I.D._, _Print_, _Step-by-Step Graphics_, _Graphis_, etc.) € Chitchat about graphic design and typography seen in unusual places-- on bicycles, for example, or on TV, on pricetags, whatever Note that this Typetalk mailing list has nothing to do with the Typetalk relay service in the U.K. HOW TO SUBSCRIBE ~~~~~~~~~~~~~~~~ To receive the Typetalk list, any electronic-mail account that can exchange mail with the Internet will work. That includes the Internet per se, Bitnet, FidoNet, many FirstClass systems, CompuServe, America Online, eWorld, Prodigy, and Genie, while QuickMail, Microsoft Mail, CC:Mail, WordPerfect Office and other corporate E-mail systems will work *as long as they are connected to the Internet in some way*. To subscribe, send a message containing the words subscribe typetalk (and no other words or characters) in the body of a message to listmanager@hookup.net. (The subject line is ignored.) To unsubscribe, send a message containing the words unsubscribe typetalk (and no other words or characters) in the body of a message to listmanager@hookup.net. To send a message to the list SO THAT EVERYONE CAN READ IT, mail your message to typetalk@hookup.net NOTE THAT THE ADDRESSES FOR (UN)SUBSCRIBING AND POSTING TO THE LIST ARE DIFFERENT! If you want to receive the Typetalk list as a digest-- a few long messages a day containing a number of shorter messages instead of all the shorter messages as separate items-- send the words on typetalk digest in the body of a message to listmanager@hookup.net. (If you want to turn the digest feature off, send the words "on typetalk nodigest" in the body of a message to listmanager@hookup.net.) TO REPORT PROBLEMS ~~~~~~~~~~~~~~~~~~ If there are any problems, report them to me at owner-typetalk@hookup.net. Sign up today! -- Joe Clark joeclark@hookup.net ~~~~~~~~~~~~~~~~~~~ Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 15 Jun 1995 16:26:59 -0400 From: "Claude L. Bullard" \ Message-ID: <9506152026.AA17483@source.asset.com> References: <9506141830.AA45783@source.asset.com> <3rp5ql$jj4@news.cais.com> Subject: Re: Web Notations [Betty Harvey] | Glad to see you back |-). I am not sure I understand your questions | completely but what I think you are asking does HTML have the | capability of referencing and delivering other types of data, such as | PDF, sound, video, pure SGML. The answer is yes. MIME is used to tell | the application what type of data it is receiving to allow it to handle | arbitrary data types. Many thanks. Yes. I've been catching up with the MID work and other research related to it. With the Mosaic User's Guide in my hand, and some hints from others, I understand that MIME specifies types which by local system means such as a .mailcap, I can associate the local helper application to the MIME type and file plus add some command line options. This is very similar to the SGML Notation declarations with fixed attributes values. I've yet to look at Java yet, although the VRML board is quite excited about it. It is one of the technologies I will factor into the rest of the summer's assignment for providing technical recommendations for the MID. The original question has one extra facet that I'm not completely clear on. An SGML notation also indicates the flavor or version of the notation as well as the helper application which services it. Can the MIME types and the .mailcap conventions specify the flavor? Another issue that becomes important is the ability to discern in advance if the local system has the capacity to process the incoming file. For HTML types, which are conventionally broken into reasonably sized chunks, this may not be such a concern unless one has a very small cache. On the other hand, once one adds scripts or MID applications which define complete processing environments as well, having a few of these around in local memory could be problematic. So we must still define ways to indicate the processing environment required for the instance. Len Bullard Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 15 Jun 1995 16:27:53 -0400 From: "Claude L. Bullard" \ Message-ID: <9506152027.AA47346@source.asset.com> References: <3rpght$ru8@usenet.srv.cis.pitt.edu> Subject: Re: GUI for DTD editing? [Dan Suthers] | If I understand him correctly, Claude is NOT asking whether HTTP can | decide what document type it got and know what to do do with it. Sure | it can do that, but each document must be of one uniform type. He is | asking whether it can deal with MULTIPLE types (notations) within a | SINGLE document. You correctly interpret my original question. Others have answered that the multipart option is available but not well supported. Another issue that I don't understand is how or if MIME supports cooperative peers: i.e., applications passing commands to one another, a.k.a, push and pull. It may be that a mail protocol simply isn't up to this. I've much to learn about all of this. | GUIs for editing a document under a given DTD are coming out, but is | anyone offering a GUI for writing the DTD themselves? Yes. The Microstar Near & Far product is excellent for this. They combine it with a methodology for DTD development as well. Some quibble that they don't use proper SGML terminology, but in the early stages of becoming acquainted with DTD development, this isn't as useful as it is later. I've fussed at them about not providing a cross-reference for their users so that SGML to SGML communications can proceed on common terms. Microstar is also working on an add-in to Word that Eliot Kimber mentioned in which the DTD tree interface of their product also drives the Word editor. My feeling is that within the year to 18 months, we'll see complete shrink-wrapped and easily used architectures for using SGML to develop and launch applications for both print and online services. That leaves the next level which is the high-end complex applicationss that define not only the document, but sophisticated transactions under author-designed interactive control. This level will enable the definition of negotiating documents such as are possible in intelligent tutoring systems. This is the gateway to adaptive dynamism in documents in which we go from using the Internet as a Library from Hell to a formal collaborative environment for performing very complex distributed multi-agent tasks. When working in virtual enterprises, the formality of the communication is quite important since the physical/visual cues are absent unless one adds a video/audio interface. Even this is coming, but, driving a user through a formal negotiation is quite useful and resembles the same systems defined for IETMs and Training. Combining HTML with Java may be one approach to this, but it leaves only the fixed widget types and the HTTP addressing and location techniques. Loading it all up into the URL may lead to problems. The # hack doesn't seem robust to me but I'm just learning about all of this. Some will want to combine the scripts inside the presentation document type in the local language. The result is going to be more options for WWW navigation such as VRML and MID already point to. Again, a protocol for communications somewhat stronger than HTTP will be needed so that not only helper applications, but cooperative peer applications can be used in conjunction with a server DBMS. While applications are appearing that claim to do all of this, I'm uncertain about the status of their interoperability. We may well be staring into some combination of OLE and CORBA. Yes, many pound on OLE, but 75 million users are hard to ignore. Let's hope that Microsoft enters into a negotiation with the rest of the world soon on that issue. Len Bullard From: bruce@sgml.dircon.co.uk (Bruce Hunter) Newsgroups: comp.text.sgml Subject: Re: CALS table processing Date: 16 Jun 1995 00:23:46 GMT Organization: SGML Systems Engineering Message-ID: <3rqiui$h1i@newsgate.dircon.co.uk> References: <3rjrfb$n65@newsgate.dircon.co.uk> <3ro89n$ek2@news1.delphi.com> X-Newsreader: WinVN version 0.81 In article <3ro89n$ek2@news1.delphi.com>, j_mcarthur@BIX.com (Jeffrey McArthur) says: >What I did was write a simple little Lex program that scans through the >table and checks for situations that can cause problems in the typesetting >engine. I incorporate that into our editing application. And I'll bet it wasn't that simple. But where did you get the semantic knowledge in order to write the lex program? Is there some "official" description of this, or did you just roll your own? regards, Bruce From: Ben Jakob <73164.172@CompuServe.COM> Newsgroups: comp.text.sgml Subject: BRF ? do you know What it is? Date: 16 Jun 1995 00:48:07 GMT Organization: dvanced Standards Inc. Message-ID: <3rqkc7$9sk$1@mhadf.production.compuserve.com> I have anumber of text files that came from a News Papaer system and I want to be able to transfer them into Acrobat format PDF or any other presentable format . All I know that the software is called BRF. Do you know which software it is? Or is there is a way to convert it to PDF format? Or HTML? Here is a example of the text Thank Ben ..PD:Friday, June 9, 1995 .DT:950609 .PB:Daily. Page: 01 .LO: .SC:News .ME: .KW: Electric Corporation. Electricity. Finance. .IL: .HD:IEC announces electricity price rise From: sroggen@infinet.com (Steve Roggenkamp) Newsgroups: comp.text.sgml Subject: Re: A general question for people with SGML experience Date: 16 Jun 1995 03:45:19 GMT Organization: InfiNet Message-ID: <3rquof$he9@horus.infinet.com> References: \ X-Newsreader: TIN [version 1.2 PL2] Phil Kime (Phil.Kime@ed.ac.uk) wrote: : I'm posting to get a rough idea of the sort of time scales and costs : associated with SGMLing large publications. I have a colleage who is : interested in the possibility of SGMLing a large document (not dissimilar : from, say, a University Calendar) from scratch. He's really appreciate : an estimate of the sort of time it would take to get someone to do : this. We are talking probably around 500,000 words. I was thinking : along the lines of 2 or 3 weeks to write a DTD, around 2 months for : markup and 2 months for a decent primary LPD. Plus a month for general : hacks and filter writing. Is this the right sort of time scale for an : SGML consultant? Also, what about cost? What does your average : consultant charge nowadays? Is SGML even the right thing to do? The : requirement is roughly that the document should be available on the Web : and in printed form without unecessary duplication of time marking up : into HTML from scratch or from the terrible output from the HTML export : of most WPs and DTP systems. There is possibly even interest in : interfacing with a database ... it sounds on the face of it as thought : SGML is ideal for this but they really have no idea about the kind of : time or cost involved and this seems like an ideal place to get rough : idea ... : p : -- : = Phil Kime (Phil.Kime@ed.ac.uk) = : = Centre for Cognitive Science/Dept of Philosophy = : = Edinburgh University = If HTML meets your needs, you could markup your document using HTML. This would eliminate the effort to write yet another DTD and you could interface to the database using CGI. I have used a filter named html2latex to produce LaTeX output for a small user document we want to have available in an online as well as print. The only modifications I had to manually make after running it through the filter was to change the word 'TeX' to '\\TeX'. Now, I will be the first to say this document was NOT a good test for a large system like yours since it contained only 11 pages of simple stuff, but it shows the promise of this approach. I wish I could give you a reference to the source code for html2latex, but I found it on the March 1995 Linux cdrom by InfoMagic available here in the states. It appears to have been written by NCSA, so you might check there. I spent a while cruising the net without success looking for it before I found it on my cdrom. Hope this helps. -- Steve Roggenkamp sroggen@infinet.com Newsgroups: comp.text.sgml From: rim@dolphin.cs.adfa.oz.au (Bob McKay) Subject: Re: SGML based databases Message-ID: <1995Jun16.071004.26064@sserve.cc.adfa.oz.au> Sender: news@sserve.cc.adfa.oz.au Organization: Australian Defence Force Academy, Canberra, Australia References: <1995Jun14.114905.55821@ucl.ac.uk> Date: Fri, 16 Jun 1995 07:10:04 GMT ubacr63@pluto.dcs.bbk.ac.uk (Dan Smith) writes: >I am currently researching textual databases and am wondering if there >are any extant SGML based (experimental / non-experimental) database >systems in existence. >As there seems to be a lot of data being produced in SGML format it >seems sensible to explore this approach. >If anyone has any information I would be grateful so that I know what >to and what not to follow up! You might look at Structured Information Manager from Ferntree Computer/CITRI; I don't have an internet address for them, but you can contact John Cook of Ferntree by fax on 61 6 247 6773 I haven't used the system, but have been impressed by what I have seen and heard of it. It's being fairly heavily used by CSIRO, the main Australian government research organisation, to support their scientific databases. Disclaimer: I've no connection with either Ferntree, CITRI, or CSIRO, though I have friends in all three organisations... Cheers Bob McKay -- ,-_|\\ Bob McKay, CS, ADFA tel:+61 6 268 8169, fax: 8581 / = \\ / \\ ACT 2600 AUSTRALIA net:rim@cs.adfa.oz.au / :::::: \\ \\_,-._/ The ADFA CS Dept offers postgrad research & course \\ :::::: / v degrees in both CS and info sys. Reply for further info \\ __ / From: Jean-Luc Sanson \ Newsgroups: comp.text.sgml Subject: Re: Adaptive documentation (different documents for different Date: 16 Jun 1995 09:03:14 GMT Organization: Direction des Etudes et Recherches EDF Message-ID: <3rrhci$hb6@edf3.der.edf.fr> References: <3r7pe5$l5d@ruby.ora.com> <3r90r1$mf1@edf3.der.edf.fr> <3rn1q6$e31@spock.ebt.com> kjs@ebt.com (Kent Summers) wrote: > > such behavior can actually be achieved in a much simpler manner than > using hytime to solve this requirement (hytime topic link/map is > however very well suited to this task). > > you could simply have different stylesheets which expose or hide > content, as required. appropriate links would be embedded in the > different skill level specific content types. this would allow you to > test your theory against your end-user base without boiling the ocean. > > if you find that you need more control over how you links behave, > independent of the text, then you might want to look into a hytime > implementation. > Kent, you are right, it all depends if you want to have control on your links. Hytime is something one should look at for mid to long term. For those who are not aware of Hytime, what we could say is that Hytime implied a big time investement, but starting now at least at the intellectual level, I am sure, will pay off in the future. Some good starts would be to read Steven DeRose (EBT staff member) book on HyTime "Making Hypermedia Work" A user's guide to HyTime S. J. DeRose and D. G. Durand Kluwer Academic Publishers ISBN 0-7923-9432-1 Eliot has also a book in preparation (september 95?) "Practical hypermedia an introduction to HyTime" Charles Goldfarb on Open Information Management New-York, Prentice Hall Professional Technical Reference ISBN 0-13-309899-0 Jean-Luc> > In article <3r90r1$mf1@edf3.der.edf.fr> From: jeffrey@hks.com (Glenda Jeffrey) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 16 Jun 1995 12:54:36 GMT Organization: Hibbitt, Karlsson & Sorensen, Inc. Message-ID: <3rruuc$1kh@argo.hks.com> References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> <3ro04f$ql5@crl.crl.com> X-Newsreader: TIN [version 1.2 PL2] Joe English (jenglish@crl.com) wrote: : What I meant was elements whose content model contains : both #PCDATA and subelements, and is not a repeatable OR : group... but now that I think about it, that wouldn't : be all that useful after all since there are plenty of : other reasons why such content models are a bad idea. I know this is a newbie question, but can you tell me why such content models are a "bad idea"? I have read Van Herwijnen's book, and understand the problems when tag minimization is used. But if you are using any sort of front end to generate the SGML, minimization is generally not an issue. Is mixed content bad even in situations where minimization is not being used? -- Glenda Jeffrey Email: jeffrey@hks.com Hibbitt, Karlsson & Sorensen, Inc Phone: 401-727-4200 1080 Main St. Fax: 401-727-4208 Pawtucket, RI 02860 From: loeffen@let.ruu.nl (Arjan Loeffen) Newsgroups: comp.text.sgml Subject: Object-orientation: statements on objects in SGML Date: Fri, 16 Jun 1995 13:47:55 UNDEFINED Organization: C\&L Message-ID: \ X-Newsreader: Trumpet for Windows [Version 1.0 Rev B] Dear reader, I have put some statements about the object-oriented model of SGML structured documents on a web page. I'd like to hear what the reader thinks. I will add more sections of my PhD thesis on the web in due time. Interested readers are invited to link to http://www.let.ruu.nl/departments/C+L/loeffen/phdthes/statemen.htm Arjan. ---------------------------------------------------------------------- Arjan Loeffen Achter de Dom 22-24 ++31+30536417 voice work Faculty of Arts 3512JP Utrecht ++31+206656463 voice home University of Utrecht The Netherlands ++31+30539221 fax work ---------------------------------------------------------------------- LOEFFEN@LET.RUU.NL www.let.ruu.nl/departments/C+L/loeffen/home.htm From: Michal Young \ Newsgroups: comp.text.sgml Subject: Re: Web Notations Date: 16 Jun 1995 14:55:46 GMT Organization: Purdue University, Software Engineering Research Center Message-ID: <3rs61i$o08@ector.cs.purdue.edu> References: <9506141830.AA45783@source.asset.com> <3rp5ql$jj4@news.cais.com> <9506152026.AA17483@source.asset.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Macintosh; I; 68K) X-URL: news:9506152026.AA17483@source.asset.com There is a MIME type called multipart/mixed (or something close to that) which is typically used in mailers to enclose multiple attachments of various kinds. There is a nicely worked out scheme for specifying whether those parts follow each other sequentially, should be presented together (e.g., a picture and a sound track), or are alternatives (e.g., a postscript and plain-text version). So, in principle the MIME content type indicator in http could be used to transfer mixed documents. I would be very surprised if any current browsers support it. --Michal From: milor001@maroon.tc.umn.edu (R A Milowski) Newsgroups: comp.text.sgml Subject: Re: A general question for people with SGML experience Date: 16 Jun 1995 10:12:18 -0500 Organization: University of Minnesota Message-ID: <3rs70i$l5a@maroon.tc.umn.edu> References: \ <3rquof$he9@horus.infinet.com> In article <3rquof$he9@horus.infinet.com>, Steve Roggenkamp \ wrote: >Phil Kime (Phil.Kime@ed.ac.uk) wrote: >: I'm posting to get a rough idea of the sort of time scales and costs >: associated with SGMLing large publications. I have a colleage who is >: interested in the possibility of SGMLing a large document (not dissimilar >: from, say, a University Calendar) from scratch. He's really appreciate >: an estimate of the sort of time it would take to get someone to do >: this. We are talking probably around 500,000 words. I was thinking >: along the lines of 2 or 3 weeks to write a DTD, around 2 months for >: markup and 2 months for a decent primary LPD. Plus a month for general >: hacks and filter writing. Is this the right sort of time scale for an >: SGML consultant? Also, what about cost? What does your average >: consultant charge nowadays? Is SGML even the right thing to do? The >: requirement is roughly that the document should be available on the Web >: and in printed form without unecessary duplication of time marking up >: into HTML from scratch or from the terrible output from the HTML export >: of most WPs and DTP systems. There is possibly even interest in >: interfacing with a database ... it sounds on the face of it as thought >: SGML is ideal for this but they really have no idea about the kind of >: time or cost involved and this seems like an ideal place to get rough >: idea ... > >: p >: -- >: = Phil Kime (Phil.Kime@ed.ac.uk) = >: = Centre for Cognitive Science/Dept of Philosophy = >: = Edinburgh University = > >If HTML meets your needs, you could markup your document using >HTML. This would eliminate the effort to write yet another DTD and you >could interface to the database using CGI. I have used a filter named [stuff deleted] ...or better yet, use an existing DTD for more complicated documents such as the DOCBOOK DTD. Many of these DTD are support by freeware/shareware applications such as Panorama (Softquad's SGML view for the WEB). I would suggest looking at an SGML ftp site to see if any of the "standard" DTD might fit what you have to do. (try ftp.ifi.uio.no) HTML may not be sufficient for your needs. Depending on *how* complex your documents are, document analysis could take from a couple days to several weeks. If it is one-person/one-document, you are probably over-estimating the time for an experienced SGML person. On the other hand, if you are an novice to SGML, you have a learning curve to get over. One of the places an SGML consultant can help you is making sure that your starting out the right way for your needs. In some sense, SGML provides many ways to approach the same problem--some better than others. You may need a consultant/expert type of person to help you setup the kinds of structures and conventions that will allow reuse of documents. I do not suggest approaching SGML piece-meal. Think of the big picture and then try to apply what is practical. ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu Copernican Solutions Incorporated (612) 825 - 4132 -- ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu Copernican Solutions Incorporated (612) 825 - 4132 From: Treva Bolton \ Newsgroups: comp.text.sgml Subject: Sr. SGML Analyst Date: 16 Jun 1995 15:17:46 GMT Organization: Jeppesen Sanderson, Inc Message-ID: <3rs7aq$e6d@windom.jeppden.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) To: MKSMSSD@Jeppden.com For more than 60 years, airline, corporate, and general aviation pilots have relied on Jeppesen quality and accurancy. As the leading publisher of lfight information and pilot training, Jeppesen aviation products and services have become the standard by which others are judged. With our experience in data acquisition, interpretation, integration, and distribution, Jeppesen has launched Maintenance Information Services. We are using leading edge technology to create simultaneous, synchronous multiple document displays. If you are challenged by advanced technical work involving development, testing, implementation, and maintenance of applications programs... If you work best under pressure while breaking new ground... If you enjoy sharing your expertise and providing guidance to other professionals... Then Jeppesen may have the perfect position for you. ******************** SR. SGML ANALYST ******************** This is an advanced technical and professional position involving the analysis of maintenance technical publications systems and processes, the development of Document Type Definitions and providing detailed specifications to Filter Writers and Programmers for the application of SGML. Analysis - Thorough knowledge of SGML concepts, terminology, and production processes. Thorough knowledge of a variety of modern programming languages and techniques. Leadership - Ability to perform the most complex tasks as required by this position. Exhibit ability to lead and mentor colleagues within this functional group. Bachelor's degree in Computer Science or equivalent and a minimum of five years experience as a systems analyst with emphasis on technical publishing and SGML applications. Knowledge of ATA-100 specifications for aeronautical documentation a plus. Must have the ability to identify problems and implement solutions with the ability to plan, organize, and carryout the assignments effectively. Must be able to establish and maintain effective working relationships and communicate effectively both orally and in writing. Jeppesen offers excellent compensation and benfits. For immediate consideration, send your resume in complete confidence to Linda Nikkel, Jeppesen, 55 Inverness Drive East, Englewood, CO 80112 or fax to (303)784-4121 or E-mail to MKSMSSD@Jeppden.com. ************************************************************************************ For more than 60 years, airline, corporate, and general aviation pilots have relied on Jeppesen quality and accuracy. As the leading publisher of flight information and pilot training, Jeppesen aviation products and services have become the standard by which others are judged. With our experience in data acquisition, interpretation, integration, and distribution, Jeppesen has launched Maintenance Information Services. We are using leading edge technology to create simultaneous, synchronous multiple document displays. If you are challenged by advanced technical work involving development, testing, implementation, and maintenance of applications programs... If you work best under pressure while breaking new ground... If you enjoy sharing your expertise and providing guidance to other professionals... Then Jeppesen may have the perfect position for you. ***************** FILTER WRITER ***************** This is an advanced technical and professional position. This position consists of professional tasks and assignments requiring implementation of data conversion solutions. The employee participates in the writing of "filter" programs to convert data from a variety of input sources into SGML files. Work is reviewed for quality of professional judgment and appraisal of results. Bachelor's Degree in Computer Science and 3-5 years as a programmer, or the equivalent combination of training and experience. Experience with SGML, SGML tools and/or C programming is a plus. In depth understanding of data processing practices and procedures. Practical knowledge of SGML concepts, terminology, and production processes. Detailed working knowledge of applicable computer programming languages and systems: Window, UNIX, Microsoft Word. Must be able to establish and maintain effective working relationships and communicate effectively both orally and writing. Jeppesen offers excellent compensation and benefits. For immediate consideration, send you resume in complete confidence to Linda Nikkel, Jeppesen 55 Inverness Drive East, Englewood, CO 80112 or fax to (303)784-4121 or E-mail to MKSMSSD@Jeppden.com. ************************************************************************************ For more than 60 year, airline, corporate, and general aviation pilots have relied on Jeppesen quality and accuracy. As the leading publisher of flight information and pilot training, Jeppesen aviation products and services have become the standard by which others are judged. With our experience in data acquisition, interpretation, integration, and distribution, Jeppesen has launched synchronous multiple document displays. If you are challenged by advanced technical work involving development, testing, implementation, and maintenance of applications programs... If you work best under pressure while breaking new ground... If you enjoy sharing your expertise and providing guidance to other professionals... Then Jeppesen may have the perfect position for you. ********************************* MANAGER, SOFTWARE DEVELOPMENT ********************************* This is a management position. This position consists of professional work involving the development, testing implementation, and maintenance of application programs and systems in support of the comany's Maintenance Information Services. An important part of the work is to assure that all systems developed conform to established coding, quality control, and documentation standards. Employee assigns work, supervises, and provides guidance to subordinates who are assigned to the program. Employee reports to the Director of Maintenance Information Services Publications. 10 years experience in application software development, rapid prototype development, and system design/implementation required, including: 7 years system analysis and design, 5 years C/C++, Visual Basic, Windows, and X-Windows programming, 5 years database design and implementation, and 3 years UNIX system development. Five years of technical management experience highly desired. A Bachelor's degree in Computer Science or equivalent required. An equivalent combination of education and experience will be considered. Must be able to establish and maintain effective working relationships and communicate effectively both orally and in writing. Jeppesen offers excellent compensation and benefits. For immediate consideration, send your resume in complete confidence to Linda Nikkel, Jeppesen, 55 Inverness Drive East, Englewood, CO 80112 or fax to (303)784-4121 or E-mail to MKSMSSD@Jeppden.com. Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 16 Jun 1995 11:52:58 -0400 From: "Claude L. Bullard" \ Message-ID: <9506161552.AA07859@source.asset.com> Subject: SGML Notations Thanks to everyone for the answers they are providing to this question. I'm wrestling with my ignorance on the subject. I also realize I am "off the VRML track" here, but I think this has implications for VRML behaviors and interoperability with other applications. I shall cross-post to comp-text-sgml, as perhaps this discussion belongs there. [Greg Seidmann] | The way one is *supposed* to know what a client can process is from the | HTTP_ACCEPT field that the client should be sending to the server. Ok. That works to signal the type and notation as long as these are the same and the syntax is not alterable. Because of the SGML Declaration and Document Type Declaration concepts, one can't count on that for SGML. With the release of Panorama, one can expect non-HTML SGML applications on the Web. I don't know how these issues are worked out for Panorama. What about a document instance that must indicate to the client what will be needed to process it? That is, "I have five notations including some xenoforms (code in foreign notation not recognized by my parser"). Are all of these indicated in a MIME header for the document to be sent? In SGML, a Notation declaration can indicate: 1. the notation (e.g., CGM, CCITT Group IV Raster, SGML itself, Microsoft Bitmap, .gif, VRML), 2. perhaps the application (in SGML, HTML is an application as SGML is the generalized markup language from which an HTML application is created for structuring a WWW hypertext type), 3. the system software for processing any data of that type 4. attributes or arguments to pass to the system software when any instance of this notation type is processed. All of these are defined in the Document Type Definition for that application (a kind of schema for all instances of the type). Entity declarations indicate what components make up the document instance. The declarations can use the notation declarations to indicate what notations the entities are written in. The entity declarations can themselves have attributes whose values are passed when a particular instance is processed. Entity references within the document instance can indicate where such components are used, or specific element types can be declared similar but not identical to the WWW anchor. By these mechanisms, SGML specifies information that can be used in the MIME headers. Public declarations in entities can be used to indicate system-wide support much as a .mailcap file is used. A document instance can use an entity reference to add these to its declaration subset. Now, an SGML application can use these features to indicate that interapplication communication much as helper applications pass a filename and maybe a zoom area (get this file and display it at this named area or set of origin-based offsets). One can also declare scripting behaviors within the application (rather than pointing out as one might with Java). Given that applications such as HTML and some CGM applications define widget types within their own notations, if interoperation is expected, it seems likely that they need a way to indicate to a system requesting them what capacity will be needed to process them and any applications they require. It seems insufficient to register what type a client can process, but also what the maximum resources an instance of that type requires. I understand that one browser (A Behlendorf Swiss Army knife) might be processing all of these document types, but that seems unlikely in the long run as more and more notation applications clamor for presence on the Internet. What is more likely, as is the current case for integrated work environments, is that separate applications with more than browsing functionality will appear soon. HTML forms already indicate the direction. In these applications, there is not the fixed widget style of a Mosaic-like application, but rather a description of transactions to logical widgets whose actual rendering or class is determined by the local system using stylesheets or internal mechanisms. Using the SGML constructs, applications indicate to the receiving system what notation support must be available and what the capacities of the these are through support declarations and the SGML Declaration. Note that it is the information designer, not the IETF WG, that makes this decision. The designer can extend their own use of notations, but cannot unilaterally extend the MIME types if I understand what you are saying. In this scenario, the WWW becomes more than an online library and provides online collaborative tasking. More "beyond_the_browser" features are declared in the document. It may be using resources at the receiving site, resources at the sending site as well as other sites it knows about. The application will be cooperating with other applications to send command packets (made up term for the occasion) using protocols defined specifically for the applications and notations the instance knows about. What I must determine (and I confess I don't know enough about MIME), is how the current protocols must be augmented to do this. I do understand the mechanism by which a URL identifies FTP, WAIS, etc. However, it doesn't seem possible to do all of the above without extending the URL to an absurd condition to achieve it. It appears to me that some working groups and vendors (Netscape META?) go out of their way to avoid using standard mechanisms for this, and instead, create hacks to do it that are guaranteed to either become convoluted or to violate the base standards (e.g., SGML, CGM, etc). I don't want a flame war on the subject; I simply want to understand how far the current IETF/W3O standards can be used before some serious rework becomes necessary. Len Bullard From: christian.w@pi.se (Christian Wallgren) Newsgroups: comp.text.sgml Subject: Re: GUI for DTD editing? Date: 16 Jun 1995 17:09:38 GMT Organization: Publishing Development AB Message-ID: <3rsdsj$j8e@news.pi.se> References: <3rpght$ru8@usenet.srv.cis.pitt.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Newsreader: WinVN 0.92.2 In article <3rpght$ru8@usenet.srv.cis.pitt.edu>, suthers+@pitt.edu (Daniel D Suthers) says: > >GUIs for editing a document under a given DTD are coming out, but is anyone >offering a GUI for writing the DTD themselves? We have a product called SGML Companion(TM) ver 1.0.3 which lets you: - Build the DTD graphically / interactively, taking care of all syntactical issues - Concurrently validate the declarations, stopping any creation of ambiguity, even complicated ones like exclusions ruling out required elements - Read any DTD made "by hand" - Combine fragments into one dtd by using external parameter entities. Frequently used fragments like CALS-tables may be shared between the DTDs - Easy-to-use entity manager, keeping track on which DTD is calling what entity - Languange customized descriptions to be added on any element or attribute, one for each language. Swapped in a snap, also in the diagram. More features to come in the next release!! Req: Windows 3.1, pref. 8MB Christian Wallgren ******************************************************* christian@pubdev.se Publishing Development AB Torpvagen 10 S-175 43 JARFALLA SWEDEN +46-8-580 375 79 From: tad@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 16 Jun 1995 18:34:15 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <3rsir7$639@cliffy.lfwc.lockheed.com> References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> <3ro04f$ql5@crl.crl.com> <3rruuc$1kh@argo.hks.com> X-Newsreader: TIN [version 1.2 PL1] Glenda Jeffrey (jeffrey@hks.com) wrote: : I know this is a newbie question, but can you tell me why : such content models are a "bad idea"? : Is mixed content bad even in situations where minimization : is not being used? Yes, mixed content is a "bad idea" even in situations where minimization is not being used. An example of one of the problems follows: \ \ ... \ \text in 'text' element\ text in 'mixed' element \ ... When the SGML parser gets to the newline following the \ tag, it interprets it as PCDATA (so it is at the end of 'mixed's content model). Then when it gets to \ it is a syntax (parse) error because text is not allowed after PCDATA in 'mixed' element. This however parses OK: \\some text\ text in 'mixed' element \ It is easier to avoid the situation altogether than try to figure out each time if it will cause a problem. Mixed content is easy to avoid if you can modify the DTD: \ \ \ Now the parser will not be expecting PCDATA until after \ tag, so it does not get confused. -- --------------------------------------------------------------------- | Tad McClellan, Logistics Specialist (IETMs and SGML guy) | Lockheed Martin Tactical Aircraft Systems | voice: (817) 763-6314 | email: mcclellantj@lfwc.lockheed.com --------------------------------------------------------------------- | ... Because I command the sheriff of the county to lead you away to | some remote spot, swing you by the neck from a knotting bough of | some sturdy oak, and let you hang until dead. | | And then, Jose Manuel Miguel Xaviar Gonzales, I further command such | officer retire quickly from your dangling corpse, that vultures may | descend from the heavens upon your filthy body until nothing shall | remain but bare, bleached bones of a coldblooded, bloodthirsty, | throatcutting, murdering Son of a Bitch. | | Judge Roy Bean (the Hanging Judge) --------------------------------------------------------------------- From: ganesh@red.seas.upenn.edu (Ganesh Ramaswamy) Newsgroups: comp.text.sgml Subject: \\boldmath with \\huge Date: 17 Jun 1995 00:45:10 GMT Organization: University of Pennsylvania Message-ID: <3rt8im$gmu@netnews.upenn.edu> X-Newsreader: TIN [version 1.2 PL2-upenn1.3] Is there a way to use \\boldmath with \\huge on latex? Presently, I have the command: \\newcommand{\\mb}[1]{\\mbox{\\boldmath $#1$}} which, when used with \\huge, does not work. Latex gives me a warning message and does a \\unboldmath automatically giving me the raw font for the math characters. Could someone suggest a way by which can circumvent this problem. Thanks in advance, Ganesh From: tim@maths.tcd.ie (Timothy Murphy) Newsgroups: comp.text.sgml Subject: Re: \\boldmath with \\huge Date: 17 Jun 1995 02:28:35 +0100 Organization: Dept. of Maths, Trinity College, Dublin, Ireland. Message-ID: <3rtb43$at@bell.maths.tcd.ie> References: <3rt8im$gmu@netnews.upenn.edu> ganesh@red.seas.upenn.edu (Ganesh Ramaswamy) writes: >Is there a way to use \\boldmath with \\huge on latex? Isn't this in the wrong newsgroup ... However, I find no difficulty in combining \\boldmath with \\huge, eg ==================================================== \\documentclass{article} \\begin{document} \\huge \\boldmath $x = y$ \\end{document} ==================================================== -- Timothy Murphy e-mail: tim@maths.tcd.ie tel: +353-1-2842366 s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml,comp.text.tex Subject: Re: \\boldmath with \\huge Followup-To: comp.text.tex Date: 16 Jun 1995 18:31:49 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3rtba5$198@crl2.crl.com> References: <3rt8im$gmu@netnews.upenn.edu> In comp.text.sgml, in article <3rt8im$gmu@netnews.upenn.edu>, Ganesh Ramaswamy \ wrote: > >Is there a way to use \\boldmath with \\huge on latex? There might be, but this is the wrong place to ask... try comp.text.tex, down the hallway and on the left... [ As far as I know, the LaTeX size-changing commands -- and most text font commands for that matter -- don't affect math mode, since text font selection works quite differently from math font selection. Why do you need to do this? Perhaps \\displaystyle is what you're looking for. ] --jenglish@crl.com From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 17 Jun 1995 03:30:19 GMT Organization: \ Message-ID: <3rti8b$536@news1.delphi.com> References: <19950614T133921Z@naggum.no> X-Newsreader: Galahad 1.1b Erik Naggum wrote: >insofar as we still want capacities, this is a good proposal. however, >capacities should be reworked to be descriptive, instead of normative. >i.e., an SGML document should include the actual capacity points needed, >and the capacity point calculations should be broken up into more useful >units. the present design is almost completely useless, and ignored. I was only proposing something modest. You are proposing a radical restructuring of the entire SGML declaration. I agree with your aims. Could you give me a hint as to what you would like to see the SGML declaration look like. >this has the unfortunate side effect of changing or invalidating existing >SGML documents. as long as we tolerate that, how about ignoring _all_ >whitespace after a start-tag and before an end-tag, regardless of this >(annoying) inclusion business? or is this too radical a change? Not too radical for me. I have played with things even more radical, like ignoring _all_ whitespace after both start and end tags. This works but has some nasty side affects. In particular if you have something like a \ tag, you cannot easilly use in the middle of a sentace. You need to do something like\␣\bold text\\␣ in the middle of a paragraph. Another option to consider is to ignore white space following both start and end tags. This has the advantage of being symetrical. You treat all tags the same. This would mean you would type: You need to do something like \bold text \in the middle of a paragraph. Your proposal would require: You need to do something like \bold text\ in the middle of a paragraph. I prefer the symetrical rule because it simplifies the treatment of tags. ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: seanb@nando.net (Sean Birdsell) Newsgroups: comp.text.sgml Subject: Re: Web Notations Date: 16 Jun 1995 23:05:37 -0500 Organization: /usr/lib/news/organization Message-ID: \ References: <9506141830.AA45783@source.asset.com> <3rp5ql$jj4@news.cais.com> In-reply-to: harvey's message of 15 Jun 1995 11:33:41 GMT X-Newsreader: (ding) Gnus v0.81 Betty, As to your question regarding HotJava (I'm new to emacs, and I haven't figured out how to quote yet), I believe you are exactly right: Code (called applets) is shipped over the wire to your HotJava browser where it is executed. I think I can describe what I saw. Imagine a web page with a gif of Mickey Mouse (I hope Disney doesn't sue 8-), now picture that gif waving at you. That is a small, and probably poor, description of the type of thing you can do with HotJava. One of Sun's examples actually includes a 20 cell working spreadsheet. Hope this helps your visualization. Sean seanb@nando.net -- Sean Birdsell seanb@nando.net From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 17 Jun 1995 10:28:18 GMT Organization: Naggum Software; +47 2295 0313 Expires: 1 Jul 1995 10:28:17 GMT Message-ID: <19950617T102818Z@naggum.no> References: \ <19950615T000059Z@naggum.no> <3ro04f$ql5@crl.crl.com> <19950615T104745Z@naggum.no> <3rq1n8$o6j@crl2.crl.com> [Erik Naggum] | well, what if #PCDATA was treated as an element in its own right, | except that it had no start- or end-tag, such that the treatment that | is now given to the entire content model would only be given to those | situations where the content models explicitly allows #PCDATA? that | is, if #PCDATA is not a possible "element" at an given point, the rules | for element content (i.e., ignore white space) applies. [Joe English] | That might interact weirdly with delimiter recognition modes. The | current rules work like USEMAP [*], where the parser uses the same mode | for the whole element. The suggested rules would work more like | DATATAG, where the mode changes at different points in the content | model. since it is not possible now, and I do not suggest to make it possible, there is no way to make short reference maps change in the #PCDATA pseudo element (from the DTD that is). recall that short reference delimiters are recognized prior to ignoring whitespace (ps) even in element content. I'd appreciate an explanation why you think this problem arises. btw, DATATAG is a non-issue. it should simply never be used. just because it's in the standard is no reason to use it, especially not when it's there only because it was not known at the time that features could be removed from the standard on the route to the final text. don't use RANK, either. RANK is in the standard only because one U.S. company threatened to destroy the fragile consensus if it was not included. such is reason enough not to implement it, and to completely ignore it, even remove it from future standards. it's pathetic that SGML parser developers include it only to get a higher number in the _informative_ Annex G "FSV" classification scheme. ignore that annex, too. let "because it's there" apply to mountains, not to standards. | so I guess it would still be *possible* to create content models that | cause inexplicable errors, just a little bit harder. I agree that a combination of OR groups with #PCDATA and other connects that expose the #PCDATA outside of its group will be confusing. I'm more interested in supporting SEQ groups with #PCDATA as an individual element. | That's just a design principle that I try to stick to. For any | elements with a structure more rigid than "text with some marked-up | subphrases" (i.e. a repeatable OR group with #PCDATA), I figure that if | character data can only appear at certain points in the content model, | then the PCDATA elements at those points must have some contextual | significance and deserve to have their own GI. this is a very good point, although the dreaded DATATAG argument applies: sometimes you want (or, worse, already have) "noise tokens" that should ideally be removed. this is actually where I have heard that this has been requested the most. when the data is actually meaningful, I agree with you -- yours is a good observation. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 17 Jun 1995 10:35:52 GMT Organization: Naggum Software; +47 2295 0313 Expires: 1 Jul 1995 10:35:52 GMT Message-ID: <19950617T103552Z@naggum.no> References: \ <3rl9ra$18v@news1.delphi.com> <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> <3rp8du$7in@cliffy.lfwc.lockheed.com> [Erik Naggum] | do you mean mixed as in _both_ elements and PCDATA ("mixed content" | is an unfortunate term in the standard that means "has PCDATA in it | somewhere"), [Tad McClellan] | My understanding of paraphrasing "mixed content" is "has PCDATA in it | somewhere along with at least one other element". | | So, \ would not be "mixed content" in my | interpretation. | | Do you agree? the standard has defined "mixed content" thus: If "#PCDATA" or a data tag group is present in a model group, the element's content is mixed content; if not, it is element content. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 17 Jun 1995 10:46:20 GMT Organization: Naggum Software; +47 2295 0313 Expires: 1 Jul 1995 10:46:20 GMT Message-ID: <19950617T104620Z@naggum.no> References: <19950614T133921Z@naggum.no> <3rti8b$536@news1.delphi.com> [Erik Naggum] | insofar as we still want capacities, this is a good proposal. however, | capacities should be reworked to be descriptive, instead of normative. | i.e., an SGML document should include the actual capacity points needed, | and the capacity point calculations should be broken up into more useful | units. the present design is almost completely useless, and ignored. [Jeffrey McArthur] | I was only proposing something modest. You are proposing a radical | restructuring of the entire SGML declaration. I agree with your aims. | Could you give me a hint as to what you would like to see the SGML | declaration look like. "a radical restructuring"? well, I might, but not in this case. the syntax of the SGML declaration may remain constant while the semantics of the capacity declaration changes from normative to descriptive. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 17 Jun 1995 13:49:59 GMT From: Paul Grosso \ Message-ID: <9506171349.AA29962@texcel.no> References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no> <3ro04f$ql5@crl.crl.com> <3rruuc$1kh@argo.hks.com> Subject: Re: SGML spec [Joe English] | What I meant was elements whose content model contains both #PCDATA and | subelements, and is not a repeatable OR group... but now that I think | about it, that wouldn't be all that useful after all since there are | plenty of other reasons why such content models are a bad idea. [Glenda Jeffrey] | I know this is a newbie question, but can you tell me why such content | models are a "bad idea"? I have read Van Herwijnen's book, and | understand the problems when tag minimization is used. But if you are | using any sort of front end to generate the SGML, minimization is | generally not an issue. | | Is mixed content bad even in situations where minimization is not being | used? Just to be clear, "mixed content" just means a content model containing elements and data characters (and "other content"). There is nothing wrong with such a model in general. The de-recommended case [8879 seems careful not to call it "deprecated" as far as I can see] is a mixed content model where data characters (i.e., objects that would match a #PCDATA part of the model) are not permitted everywhere in the content. The reason this kind of model causes problems unfolds in three steps: (1) "Separator characters, which are recognized as separators in element content, are treated as data in mixed content" (8879, first Note in clause 11.2.2). In any mixed content, the recognition of data characters at a certain point means [speaking informally] that the parser state moves to the next available #PCDATA in the content model. This is true for all mixed content models and is usually no problem if data characters are allowed anywhere in the content (i.e., there is always a #PCDATA in the content model that can be matched). (2) However, in the de-recommended case, when the recognition of some data characters causes the need for the parser to match a #PCDATA in the model, there won't always be a #PCDATA to match. Furthermore, even if the data characters are allowed at that point, the fact that the recognition of them "moved" the parser to some other place in the model means that you may now be in a state where certain elements that were in context are now no longer in context and/or subsequent data characters will no longer be in context. This means that the unintended introduction of data characters in one part of the content of an element with de-recommended mixed content may affect the parsing of the rest of that element's content so as to make subsequent elements and/or data characters out of context where they would have been properly in context without the unintended introduction of data characters. Note that the immediate cause of the problem is "unintended introduction of data characters." If the unintended never happened, there would be no problem. (3) Separator characters (e.g., spaces and non-ignored records ends) are not treated seriously by many people and programs. However, SGML does treat them seriously in mixed content. This leads to separator characters being highly implicated in the "unintended introduction of data characters" which in turn leads to all the problems with de-recommended mixed content models. So now to answer Glenda's question. First, nothing is wrong with mixed content in general. The SGML standard de-recommends certain types of mixed content models. The problem with the de-recommended mixed content model goes beyond minimization. Issues with minimization can make the problems even messier, but as per my earlier discussion, there are problems that have nothing to do with minimization. What makes the de-recommended mixed content models problematic is the tendency for unintended introduction of data characters, most especially, separator characters such as spaces and record ends. If, as Glenda suggests, you are using a fully SGML-aware front end to generate the SGML, it should present an editing interface that avoids any issue of creating invalid content, and it should produce SGML that knows the 8879 rules about character (including separator) recognition so that it doesn't unintentionally introduce data characters. In this case, the de-recommended mixed content model should not be a problem. If, however, there is any chance that your SGML file will be modified by any person or program that may not be as careful about the unintended introduction of separators, and if it is just as satisfactory for your DTD to avoid a de-recommended mixed content model, then it is probably safest to do so. paul -- Paul Grosso VP Research, ArborText, Inc. and Chief Technical Officer, SGML Open Email: paul@arbortext.com Newsgroups: comp.text.sgml From: Phil.Kime@ed.ac.uk (Phil Kime) Subject: Why is LINK so bad? Message-ID: \ X-Attribution: Phil X-Www: http://www.cogsci.ed.ac.uk/~philkime/home.html Sender: usenet@cogsci.ed.ac.uk (C News Software) Reply-To: Phil.Kime@ed.ac.uk (Phil Kime) Organization: Centre for Cognitive Science X-Comments: MIME/PGP welcome X-Newsreader: Emacs 19.27.2 (GNUS 4.1) Date: Sat, 17 Jun 1995 14:02:21 GMT I notice a general aversion to the SGML LINK feature. Why is this? Correct me if I'm wrong but it is basically there to provide part of the functionality the application filter would have to do in order to get the parsed SGML uptput into an application-specific format. The advantage it has is that it is not a purely syntactic process as it has access to SGML data structures etc. so it should make things easier. Is this right? p -- = Phil Kime (Phil.Kime@ed.ac.uk) = = Centre for Cognitive Science/Dept of Philosophy = = Edinburgh University = From: jonah1@ibm.net Newsgroups: comp.text.sgml Subject: DWScript (IBM's PC GML/Script)Driver Request Date: 17 Jun 1995 18:07:17 GMT Message-ID: <3rv5kl$30cb@news-s01.ny.us.ibm.net> Reply-To: jonah1@ibm.net X-Newsreader: IBM NewsReader/2 v1.09 Does anyone know where I might begin to locate an HP LaserJet III/IV or PostScript printer driver for this 1989 software package. Of course IBM doesn't support this DOS package any longer. I am open for source suggestions. From: harvey \ Newsgroups: comp.text.sgml Subject: Re: Web Notations Date: 17 Jun 1995 18:25:28 GMT Organization: Electronic Commerce Connection Message-ID: <3rv6mo$n4a@news.cais.com> References: <9506141830.AA45783@source.asset.com> <3rp5ql$jj4@news.cais.com> \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) seanb@nando.net (Sean Birdsell) wrote: >Betty, > >As to your question regarding HotJava (I'm new to emacs, and I haven't figured out >how to quote yet), I believe you are exactly right: Code (called applets) is shipped >over the wire to your HotJava browser where it is executed. I think I can describe >what I saw. Imagine a web page with a gif of Mickey Mouse (I hope Disney doesn't >sue 8-), now picture that gif waving at you. That is a small, and probably poor, >description of the type of thing you can do with HotJava. One of Sun's examples >actually includes a 20 cell working spreadsheet. Hope this helps your visualization. > Thanks Sean: I thought this was what it did. I am really anxious to get a demonstration (I talked with our Sun Salesperson yesterday). I am wondering just how robust and interactive the programming language is. I am thinking that HotJava can really be useful for Electronic Technical Manuals (ETM) and provide more functionality and interactivity that is necessary. We know that Mosaic Communications (Netscape) has stated they are going to support HotJava. My next question, could HotJava be useful in a full-SGML environment DBMS->Electronic SGML Viewer to provide full-function IETM capabilities? It should be an interesting and exciting time during the next couple of years! <><><><><><><><><><><><><><><><><><><><><><><><><><><><><> Betty Harvey | PERSONAL David Taylor Model Basin | harvey@eccnet.com harvey@navysgml.dt.navy.mil | ---------------------------------------------------------- Navy CALS WWW Server: http://navysgml.dt.navy.mil <><><><><><><><><><><><><><><><><><><><><><><><><><><><><> From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 17 Jun 1995 11:50:40 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3rv860$5h4@crl3.crl.com> References: \ <19950615T104745Z@naggum.no> <3rq1n8$o6j@crl2.crl.com> <19950617T102818Z@naggum.no> Erik Naggum \ wrote: >since it is not possible now, and I do not suggest to make it possible, >there is no way to make short reference maps change in the #PCDATA pseudo >element (from the DTD that is). recall that short reference delimiters are >recognized prior to ignoring whitespace (ps) even in element content. > >I'd appreciate an explanation why you think this problem arises. Currently, the parser only needs to examine the stack of open elements to determine whether SEPCHARs are treated as data or not; with the proposed change, they would have to examine the current point in the content model to make this decision. (By analogy, USEMAP affects delimiter recognition for the whole element, while with DATATAG the delimiters can change from point to point in the content model. That's what I was getting at; I agree completely that DATATAG shouldn't be used.) Not that I think this is necessarily a problem -- changing the lexical mode at different points in the content model is exactly what's desired -- just that it might add another complication to parser imlementations. --Joe English jenglish@crl.com From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: Why is LINK so bad? Date: 17 Jun 1995 12:17:26 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3rv9o6$5pl@crl3.crl.com> References: \ Phil Kime \ wrote: >I notice a general aversion to the SGML LINK feature. Why is this? I think this is partly because it's not widely understood, and partly because it's not widely implemented. (It may not be widely implemented precisely *because* it's not widely understood, so users and application developers tend not to request or require the feature.) LINK would not be that difficult to implement, except for one thing: parameter entity declarations in LPDs override declarations in the base DTD, which makes it necessary to parse the prolog in two passes. Without this feature, only one pass is needed. This is probably the main reason it's not generally available. >Correct me if I'm wrong but it is basically there to provide part of the >functionality the application filter would have to do in order to get >the parsed SGML uptput into an application-specific format. The >advantage it has is that it is not a purely syntactic process as it has >access to SGML data structures etc. so it should make things easier. Is >this right? Essentially, yes. LINK is a useful way to map a document type onto an architectural form without modifying the base DTD. --Joe English jenglish@crl.com From: seanb@nando.net (Sean Birdsell) Newsgroups: comp.text.sgml Subject: Re: Web Notations Date: 17 Jun 1995 17:04:41 -0500 Organization: /usr/lib/news/organization Message-ID: \ References: <9506141830.AA45783@source.asset.com> <3rp5ql$jj4@news.cais.com> \ <3rv6mo$n4a@news.cais.com> In-reply-to: harvey's message of 17 Jun 1995 18:25:28 GMT X-Newsreader: (ding) Gnus v0.81 Betty, I wish I could answer your more detailed questions, but I can't. I am not a programmer (I only play one on TV :-) and I am extremely new to SGML. (My interest came about from doing an on-line manual to be delivered through web browsers.) I happened to see HotJava at a client site. Good luck in getting the answers to your questions. Sean seanb@nando.net -- Sean Birdsell seanb@nando.net From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 17 Jun 1995 23:05:29 GMT Organization: ACM Network Services Message-ID: <3rvn3p$dle@hopper.acm.org> References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no>,<3ro04f$ql5@crl.crl.com> Reply-To: davep@ACM.ORG In article <3ro04f$ql5@crl.crl.com>, jenglish@crl.com (Joe English) writes: >What I meant was elements whose content model contains >both #PCDATA and subelements, and is not a repeatable OR >group... but now that I think about it, that wouldn't >be all that useful after all since there are plenty of >other reasons why such content models are a bad idea. And there are plenty of reasons why such content models are a good idea. The only reason they're bad that's generic to SGML is that they trick untrained users into making errors *when those users must enter data with SGML-unaware editors*. An application designed for use with SGML-smart editors shouldn't have to worry about the ignored-RE-that's-still-data problem; the editor won't let the user put that bad RE in in the first place. This "mixed content problem" is a chimera that is overrated. Dave Peterson SGMLWorks! davep@acm.org From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 18 Jun 1995 18:11:17 GMT Organization: ACM Network Services Message-ID: <3s1q85$gf6@hopper.acm.org> References: \ <19950615T104745Z@naggum.no> <3rq1n8$o6j@crl2.crl.com> <19950617T102818Z@naggum.no>,<3rv860$5h4@crl3.crl.com> Reply-To: davep@ACM.ORG In article <3rv860$5h4@crl3.crl.com>, jenglish@crl.com (Joe English) writes: >(By analogy, USEMAP affects delimiter recognition for >the whole element, while with DATATAG the delimiters can >change from point to point in the content model. That's >what I was getting at; I agree completely that DATATAG >shouldn't be used.) > >Not that I think this is necessarily a problem -- >changing the lexical mode at different points in the content model >is exactly what's desired -- just that it might add another >complication to parser imlementations. One of the things we who are working seriously on the SGML revision have been looking at recently is a description of SGML syntax that separates out the lexical analysis from the grammatical analysis at a level higher than individual characters. Let me assure you, as one who has looked closely at the subject, that the lexical mode must change at many, many points in the content model even with the current syntax. Dave Peterson SGMLWorks! davep@acm.org (SGMLWorks! is neither a part of nor employed by the ACM. ACM just provides my network access--for a fee--since I am a member thereof who choses to avail myself of the service. SGMLWorks! is a consulting firm; I'm the Principal Consultant.) From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: Blessings and Curses Date: 18 Jun 1995 18:09:05 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3s2inh$7h8@crl10.crl.com> References: <9506092043.AA37374@source.asset.com> \ Sebastian Rahtz \ wrote: >In article <9506092043.AA37374@source.asset.com> > "Claude L. Bullard" \ writes: > >>Instead, a lone college student and some friends dictated our HTML future. > >I think this is historically a rather strange view. HTML came from CERN, from >professional computer scientists employed by a very large international >research organization. Marc Andreessen and NCSA may have made it popular and >useable with Mosaic, but they didnt invent HTML, or substantially alter its >character. Andreessen came up with the \ tag for inline images; this was a departure from Tim BL's original idea that the \ (anchor) element should be used for this (and for any embedded resource, not just bitmapped images). NCSA Mosaic introduced several formatting-based tags like \
(horizontal rules) and (I believe) \ and \ for selecting boldface and italic fonts. (Before that, all the phrase-level elements were semantic ones derived from TeXinfo.) NCSA also invented fill-out forms, which, though the idea is a very good one, the syntax used to specify them is by far the most awkward and badly-specified part of the DTD. So yes, I'd say that Andreessen and NCSA did substantially alter the character of HTML. --Joe English jenglish@crl.com From: vincent@halcyon.com (Richard D. Vincent) Newsgroups: comp.text.sgml Subject: unstuffing an HTML Editor Followup-To: comp.text.sgml Date: 19 Jun 1995 01:30:37 GMT Organization: Northwest Nexus Inc. Message-ID: \ I have a PowerMac6100/60 and have just downloaded a couple of HTML Editors for the Mac from the net.Both of them were described as teachtext files but when I tried to open them I got a message stating that they are too large to be opened by teachtext.Next I tried to go into Microsoft word and do it from there.I got a page of what looks like stuffed material and a message " Must decode with binhex4".Now I am confused and it has been awhile since I have decompressed any files. I do have the application stuffit deluxe.Can someone walk me through the steps topen up my HTML editor or tell me of a URL where there is a self extracting archived HTML Editor?It might be a couple of months before I figure it out myself and of course it will have been about five simple steps.Any infois appreciated in advance. Richard Vincent vincent@halcyon.com From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: CALS table processing Date: 19 Jun 1995 01:54:24 GMT Organization: \ Message-ID: <3s2lcg$1o1@news1.delphi.com> References: <3rqiui$h1i@newsgate.dircon.co.uk> X-Newsreader: Galahad 1.1b Bruce Hunter wrote: >>What I did was write a simple little Lex program that scans through the >>table and checks for situations that can cause problems in the typesetting >>engine. I incorporate that into our editing application. > >And I'll bet it wasn't that simple. But where did you get the semantic >knowledge in order >to write the lex program? Is there some "official" description of this, or >did you just roll >your own? Just a few notes about the program. I tend to prefer a version of Lex that outputs Pascal instead of C. The program extracts one other tag \. This is a tag that occurs in many of the documents we are currently working with. Extracting \ allows more meaningful error messages; but you could easilly remove that from the program. The other check that the program does is see if entities follow the rule that they only contain letters, numbers, and period. This will flag any "&"; but allow \& to pass through. The program source also contains the version control headers I use. So here is the program: ------------------- cut here ------------------ %{ program CntCols; uses LexLib; const { ***keyword-flag*** '%v'} VersionNumber='2'; { ***keyword-flag*** '%d %t'} VersionDate='26-Jul-94 21:48:06'; { **> revision history <** } (* 1 CNTCOLS.L 26-Jul-94,21:46:54,`SJSM' First Version. Added entity check *) (* 2 CNTCOLS.L 26-Jul-94,21:48:06,`SJSM' Bug in comment. *) { **> revision history <** } var brandName: string; badEnts, problems, defs, cols: integer; procedure ProcessColSpan(s: string); var i, c: integer; begin { 12345678901 colspan="1" } Val(Copy(s, 10, Length(s) - 10), i, c); if (c <> 0) or (i <= 0) then begin WriteLn(yyoutput, 'Illegal colspan: ', s); end else begin inc(cols, i - 1); end; end; procedure IllegalEntity; begin WriteLn(yyoutput); WriteLn(yyoutput, 'Illegal Entity on line ', yylineno, ' near column ', yycolno); WriteLn(yyoutput, yyline); WriteLn(yyoutput, ' ':yycolno-2, '^'); inc(badEnts); start(0); end; %} TBLCDEFS [Tt][Bb][Ll][Cc][Dd][Ee][Ff][Ss] TBLCDEF [Tt][Bb][Ll][Cc][Dd][Ee][Ff] TBLROW [Tt][Bb][Ll][Rr][Oo][Ww] TBLCELL [Tt][Bb][Ll][Cc][Ee][Ll][Ll] COLSPAN [Cc][Oo][Ll][Ss][Pp][Aa][Nn] BRNDNAME [Bb][Rr][Nn][Dd][Nn][Aa][Mm][Ee] ENTCHARS [a-zA-Z0-9] ENTEND ; %start tcel bnam ent celent %% \\<{BRNDNAME}\\> start(bnam); \\\< start(0); \[^\\<]* brandName := yytext; \& start(celent); & start(ent); \{ENTCHARS} ; \{ENTCHARS} ; \; start(tcel); \; start(0); \. IllegalEntity; \. IllegalEntity; \\<{TBLCDEFS} defs := 0; \\<{TBLCDEF} inc(defs); \\<{TBLROW} cols := 0; \\<{TBLCELL} begin inc(cols); start(tcel); end; \\<\\/{TBLROW} begin if defs <> cols then begin WriteLn(yyoutput); WriteLn(yyoutput, 'Table problem in ', brandName); WriteLn(yyoutput, 'near line ', yylineno); WriteLn(yyoutput, ' defs = ', defs); WriteLn(yyoutput, ' cols = ', cols); inc(problems); end; end; \\\> start(0); \{COLSPAN}=\\"[0-9]+\\" ProcessColSpan(yytext); \.|\\n ; |\\n ; %% begin WriteLn('CntCols version ', VersionNumber, ' ', VersionDate); brandName := '????'; problems := 0; badEnts := 0; if ParamCount >= 1 then begin Close(yyinput); Assign(yyinput, ParamStr(1)); Reset(yyinput); end; if ParamCount >= 2 then begin Close(yyoutput); Assign(yyoutput, ParamStr(2)); ReWrite(yyoutput); end; yylex; WriteLn; WriteLn; if problems = 0 then WriteLn('No table problems') else WriteLn(problems, ' table problems'); if badEnts = 0 then WriteLn('No entity problems') else WriteLn(badEnts, ' entity problems'); end. ------------------- cut here ------------------ ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: tactics@magna.com.au (David Churchland) Newsgroups: comp.text.sgml Subject: SGML - HTML newbie looking for input Date: 19 Jun 1995 02:09:10 GMT Organization: TACTICS Consulting Pty Ltd Message-ID: <3s2m86$8iq@kettle.magna.com.au> Mime-Version: 1.0 X-Newsreader: WinVN 0.93.14 I have been doing a considerable amount of research into SGML and have started doing some work with HTML. I have been thinking that HTML might be a more accessable and, initially, less painful way of putting some of our documentation on-line prior to moving toward a full SGML solution. My problem is that I am having to battle the Windows help proponents who love the low initial co$t of using the help engine. If anyone can give me some input on the following topics I would really appreciate it: - are there any good (freeware/shareware) indexing/text retrieval tools for searching multiple HTML documents, - What downsides/traps are there in doing internal documentation in HTML and viewing with Netscape/Mosaic, and - Is there anyone who is doing it, why/ why not. Any responses would be much appreciated, especially anything that I can use to combat the MS WinHelp camp. Thanks in advance, David From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 18 Jun 1995 19:11:37 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3s2mcp$8fk@crl10.crl.com> References: \ <19950617T102818Z@naggum.no> <3rv860$5h4@crl3.crl.com> <3s1q85$gf6@hopper.acm.org> \ wrote: >In article <3rv860$5h4@crl3.crl.com>, jenglish@crl.com (Joe English) writes: > >>(By analogy, USEMAP affects delimiter recognition for >>the whole element, while with DATATAG the delimiters can >>change from point to point in the content model. That's >>what I was getting at; I agree completely that DATATAG >>shouldn't be used.) > [ ... ] >One of the things we who are working seriously on the SGML revision have been >looking at recently is a description of SGML syntax that separates out the >lexical analysis from the grammatical analysis at a level higher than >individual characters. Let me assure you, as one who has looked closely >at the subject, that the lexical mode must change at many, many points in >the content model even with the current syntax. Are there any cases where the lexical mode to be chosen depends on the current point in the content model? I.e., after parsing or implying an end-tag for a subelement, does the lexical mode ever change to something other than what it was before the subelement was opened? Other than DATATAG, I can't think of any cases where this happens. --jenglish@crl.com From: fdrake@csgrad.cs.vt.edu (Fred L. Drake Jr.) Newsgroups: comp.text.sgml Subject: Re: SGML - HTML newbie looking for input Date: 19 Jun 1995 00:06:59 -0400 Organization: Virginia Tech, Blacksburg, Virginia Message-ID: <3s2t53$fve@csgrad.cs.vt.edu> References: <3s2m86$8iq@kettle.magna.com.au> In article <3s2m86$8iq@kettle.magna.com.au>, David Churchland \ wrote: >- What downsides/traps are there in doing internal documentation in HTML >and viewing with Netscape/Mosaic, and >- Is there anyone who is doing it, why/ why not. David, I don't know about the search engines, but I do know that at least one company here in town is using Mosaic as the display engine with an in-house server for all documentation and bug-tracking activity. There's even a public interface for their clients to make bug reports when needed. They're main reasons for using Web technology instead of WinHelp included portability (they typically have 12 - 13 different platforms to support) and extensibility (interfacing to a bug-tracking database). I know there are things that can be done with WinHelp, but I'm not sure how difficult it is to add things. (C/C++/VB programming?) Using a Web server allows CGI scripts to be used, and those can be written in just about anything that even acts remotely like a program. I hope this helps in your decision making! -Fred -- http://csgrad.cs.vt.edu/~fdrake/ From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML spec Date: 19 Jun 1995 12:54:19 GMT Organization: Naggum Software; +47 2295 0313 Expires: 3 Jul 1995 12:54:19 GMT Message-ID: <19950619T125419Z@naggum.no> References: \ <19950614T133921Z@naggum.no> <3rncmn$cn8@crl9.crl.com> <19950615T000059Z@naggum.no>,<3ro04f$ql5@crl.crl.com> <3rvn3p$dle@hopper.acm.org> [Dave Peterson] | The only reason they're bad that's generic to SGML is that they trick | untrained users into making errors *when those users must enter data | with SGML-unaware editors*. An application designed for use with | SGML-smart editors shouldn't have to worry about the ignored-RE-that's- | still-data problem; the editor won't let the user put that bad RE in in | the first place. | | This "mixed content problem" is a chimera that is overrated. I disagree. you have only moved the problem from the parser to the generator of the SGML document. this is moving it in the wrong direction. already, the amount of knowledge required to generate a valid SGML document is too large, and is not for naive software. I want the amount of knowledge required to be _sharply_ decreased in the revised standard SGML. we should have learned much from the first edition of the standard, but I would prefer if this knowledge did not take the form of "we know there's a problem -- here's how you get around it" which I have so far seen as the only answer, but rather "we know there's a problem -- here's how we plan to remove it". as time has went by, I have been moving in the direction of making the SGML declaration mandatory in the revised SGML, and of changing the edition indicator, both of which seem obvious for a revised standard. that is, to move from \ to \. at this point, we are free to change the form of the SGML declaration, as well as some other troublesome aspects of the language that do not change its fundamental nature. as always, existing tools will know they can't deal with a new-edition SGML document merely by inspecting the SGML declaration. (it was a grave mistake to allow the SGML declaration to be omitted, but that may have been because the SGML declaration itself was so unwieldy that not allowing it would have been an impediment to the acceptance of the standard.) to make this task simpler, we need to define _two_ explicit formats for SGML: one for local system files and one for interchange, possibly by making certain things optional and others required in each of them. this is to obviate the need for an SDIF that nobody uses because they don't see the difference between local files and interchange. I think the "mixed content problem" needs to be addressed beyond finding clever ways to say "somebody else's problem". above all, it should be a lot simpler to generate valid SGML, and a lot harder to generate invalid SGML than it is today. the "mixed content problem" is only part of this greater problem. during the Lisp standardization, it became a standing joke that "a sufficiently smart compiler" could figure out various complicated things. the SGML world has "the sufficiently smart editor" for the same reason. I don't think this is a good idea, or we will wind up with only "sufficiently smart users", just like Lisp does. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: pm@piau () Newsgroups: comp.text.sgml Subject: Rq: ISO 9573 DTD (Equations) Date: 19 Jun 1995 14:07:30 GMT Message-ID: <3s40b2$kvt@news.cict.fr> Reply-To: patrick.mathe@avions.aerospatiale.fr X-Newsreader: TIN [version 1.2 PL2] Where can I retreive ISO 9573 standard (equations DTD)? Thanking you in anticipation. -- Patrick MATHE patrick.mathe@avions.aerospatiale.fr H21A, Blagnac Orientation - Toulouse (3102MP) FRANCE From: ietmteam@mail.worldgate.edmonton.ab.ca (ietmteam) Newsgroups: comp.text.sgml Subject: ASCII Editor for WindowsNT Date: 19 Jun 1995 14:09:25 GMT Organization: CAE Aviation Message-ID: <3s40el$meu@scanner.worldgate.edmonton.ab.ca> X-Newsreader: WinVN 0.92.6+ I am looking for a text editor for windowsNT. The editor must accept long (unix type) file names and save the text in a unix format (LF only not CR and LF). If anyone knows of such a beast, please Email me with the name and location of the app. Thanks Vincent Savard From: milor001@maroon.tc.umn.edu (R A Milowski) Newsgroups: comp.text.sgml Subject: Comments on ideas... Date: 19 Jun 1995 09:27:01 -0500 Organization: University of Minnesota Message-ID: <3s41fl$r0k@maroon.tc.umn.edu> I wanted to get comments from all of you in net-land about some ideas I have that are going into a paper. Some of them are SGML related and some of them are SGML programming related. (Two separate audiences?) 1. Entity management, as a whole, has not been treated as of equal importance as parsing--case in point, there are *no* entity management products. Although there are a few SGML database products on the horizon, if not now, they may or may not address the issue of consistent entity management. 2. SGML processing models have traditionally been linear in the sense that an application asks a parser to parse an SGML document and the parser asks the entity manager (if any) for the document. This is a unfortunate standard for SGML processing since a true SGML system should be a system where Entity Management, Parsing/Validating, and Application work together in a "triangle of interdependency". That is, the functions and services that they provide are very much interrelated and interdependent. 3. The parser is the only way in which raw SGML information can be correctly "understood" by an SGML application. 4. There is not *standard* interface between an SGML parser and the application. 5. All SGML parsers supply a subset of the information provided by raw SGML as denoted in the ISO 8879 standard. 6. The only way to provide a "standard interface" that serves everyone's needs to the an SGML parser is to serve the productions of the SGML standard. This, of course, would be a low-level interface. 7. A standard object-framework, along the lines of Fresco, would be of great benefit to *application* writers--not tool vendors. 8. A standard set of services (object-framework?) for entity management would be of great benefit to application writers and tool vendors. Although SGML open has a definition for specification of this, there needs to be some kind of API/object-framework specification. 9. (The most important one) There is a great need in the SGML industry for programming-level standards so that "components" can be developed that allow for component-level inter-operation. That is, so that a variety of tools and applications can be integrate with ease to create open SGML applications. 10. Such programming-level standardization is necessary for the success of SGML since the practicality of the matter is that people won't use what is difficult or costly to integrate and develop for. Ideas? Comments? Suggestions? Additions? Please post them here... I want everyone who reads this group to see them. -- ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu Copernican Solutions Incorporated (612) 825 - 4132 From: pm@piau () Newsgroups: comp.text.sgml Subject: Rq: ISO 12083 standard Date: 19 Jun 1995 14:37:45 GMT Message-ID: <3s423p$l92@news.cict.fr> Reply-To: patrick.mathe@avions.aerospatiale.fr X-Newsreader: TIN [version 1.2 PL2] Where can I retreive ISO 12083 standard? Thanking you in anticipation. -- Patrick MATHE patrick.mathe@avions.aerospatiale.fr Newsgroups: comp.text.sgml From: ak@oasis.icl.co.uk (Alfie Kirkpatrick) Subject: Re: SGML - HTML newbie looking for input Message-ID: \ Sender: news@oasis.icl.co.uk Reply-To: ak@oasis.icl.co.uk Organization: ICL, Bracknell, UK X-Newsreader: TIN [version 1.2 PL2] References: <3s2m86$8iq@kettle.magna.com.au> <3s2t53$fve@csgrad.cs.vt.edu> Date: Mon, 19 Jun 1995 16:20:33 GMT Fred L. Drake Jr. (fdrake@csgrad.cs.vt.edu) wrote: : David, : I don't know about the search engines, but I do know that at least one : company here in town is using Mosaic as the display engine with an in-house : server for all documentation and bug-tracking activity. There's even a public : interface for their clients to make bug reports when needed. This reminded me that SCO used HTML and Mosaic for their on-line documentation for a recent SCO UNIX release. They have a search mechanism, but it's nothing fancy. I don't believe they author in HTML, but convert from troff. The whole look is quite nice. I wouldn't advise authoring in HTML as maintaining links and/or contents pages will be a nightware and moving to "proper" SGML may be harder than you think. If you author in SGML, you could even convert into Winhelp and satisfy those people wanting a cheap browser while future proofing the data and publishing in other ways. If the product is Windows and will only ever be Windows, SGML will be hard to justify. There are too many tools around that help you create "whizzo" help systems and SGML will always _look_like_ a step backwards. My 2p, Alfie. -- +-Professional Publishing Services---------------+ | Alfie Kirkpatrick ICL | | external: +1344 472500 Lovelace Road | | internal: 7263 2500 Bracknell, Berks | | mail: ak@oasis.icl.co.uk RG12 8SN | +------------------------------------------------+ From: jeffrey@hks.com (Glenda Jeffrey) Newsgroups: comp.text.sgml Subject: Help with content model? Date: 19 Jun 1995 19:27:58 GMT Organization: Hibbitt, Karlsson & Sorensen, Inc. Message-ID: <3s4j3u$uo@argo.hks.com> X-Newsreader: TIN [version 1.2 PL2] I'm trying to put together a conten model like this: \ Because of the "no look-ahead" rule, this is an ambiguous content model. One option is this: \ \ but this is very non-intuitive for the authors. I've resorted to playing a name game, like this: \ This is only slightly more intuitive than the Wrapper option. Does anybody have a better idea? Can anybody explain to me why this is such a Bad Thing to want to do that the standard does not permit it? I mean, this is really the structure I have -- the order of A and C must be preserved, and all the items really are optional... TIA... -- Glenda Jeffrey Email: jeffrey@hks.com Hibbitt, Karlsson & Sorensen, Inc Phone: 401-727-4200 1080 Main St. Fax: 401-727-4208 Pawtucket, RI 02860 From: Peter Murray-Rust \ Newsgroups: comp.text.sgml Subject: Chemical Markup Language (CML) Date: 19 Jun 1995 21:58:55 +0100 Organization: Myorganisation Sender: news@news.demon.co.uk Message-ID: <803595163snz@ursus.demon.co.uk> Reply-To: Peter@ursus.demon.co.uk X-Newsreader: Demon Internet Simple News v1.29 X-Posting-Host: ursus.demon.co.uk I have been very grateful to many of you for your help in guiding me in the SGML tools and techniques needed to create a Chemical Markup Language using SGML. I have created a prototype application which I hope to announce shortly but would be grateful for a little help before then. 1. Are comments (-- --) allowed within model groups in ATTLIST ( I hope I have the terminology right). This would help greatly in documenting my DTDs, e.g: \ 2. I have realised (it's obvious in retrospect!) that SGML can provide little in the way of generic tools apart from a parser and things like dtd2html (for both of which many thanks). Has anyone produced a generic editor/authoring tool which can take a DTD and put up windows or fill-in forms for the content and attributes? Ideally I would like this in tcl/tk since this is my preferred GUI environment - but that's probably asking a lot. I'll probably have to write it ... 3. I require an SGML declaration file (because some of the names exceed 8 characters , etc) and at present, using SGMLS, I include this on the command line. Is there a way of referring to it via the catalog or otherwise tying it closer to the program? 4. I have to include a lot of tabular data, including multidimensional matrices, etc and I've seen on this newsgroup that this is an area that SGML doesn't provide readymade support for. One possible approach could be to include this data using another (defined) standard, such as HDF by defining \ ... \ as part of the DTD. Is this a useful way forward? (I have already decided that certain legacy filetypes will be encapsulated in the first version of CML - see the DTD fragment above...) 5. This is probably difficult to express precisely but I'd welcome comments. I have built the present DTD to represent identifiable 'chunks' of information. These are things like BIBLIOGRAPHY, GLOSSARY, MOLECULE, DATA, FOREIGNfile, which I've allowed to appear in a fairly flexible order in the document instance. I shall create a viewer and editor to support this approach. In particular it means that if someone is sent a *.cml file, a single set of DTDs will allow it to be parsed , rendered, edited, etc. This in itself will be a major benefit to our community. But it means that the overall document has a flexible and semantically undefined structure. It's rather as if the highest level of organisation in a book was VERSE, PARA, etc and that this worked equally whether the book was an anthology or a newspaper. From my reading of TEI this is catered for by having many different DTDs which can be selected for different types of 'book'. Presumably the recipient (I'm thinking in terms of something coming over the WWW) decides from the DOCTYPE which set of DTDs to use on that particular instance. The rendering software then has to be able to cater for all types - e.g. It can render it as a book of verse or a newspaper accordingly. This is far too ambitious for our community at present (I wouldn't know what types of document we might have, but they would range through: thesis patent, computer output, picture, notebook, abstract, publication, etc). Is there any alternative way that at least some of the implied semantics can be captured whilst not producing inflexible DTDs . (I expect I'm asking for the impossible!) TIA Peter -- Peter Murray-Rust, Biomolecular Structure, Glaxo Research and Development, Stevenage, UK, pmr1716@ggr.co.uk (posted from domestic net connexion) From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: Help with content model? Date: 19 Jun 1995 21:58:19 GMT Organization: Naggum Software; +47 2295 0313 Expires: 3 Jul 1995 21:58:19 GMT Message-ID: <19950619T215819Z@naggum.no> References: <3s4j3u$uo@argo.hks.com> [Glenda Jeffrey] | I'm trying to put together a conten model like this: | | \ \ | Can anybody explain to me why this is such a Bad Thing to want to do | that the standard does not permit it? I mean, this is really the | structure I have -- the order of A and C must be preserved, and all the | items really are optional... the only thing the standard does is to force a lot of theory on the DTD designers that DTD designers should not have had to know quite that intimately. graph rewriting is not necessarily a trivial operation. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. Newsgroups: comp.text.sgml From: papresco@calum.csclub.uwaterloo.ca (Paul Prescod) Subject: Re: GUI for DTD editing? Message-ID: <3s59mo$918@calum.csclub.uwaterloo.ca> Sender: news@undergrad.math.uwaterloo.ca (news spool owner) Organization: University of Waterloo Computer Science Club References: <3rpght$ru8@usenet.srv.cis.pitt.edu> <9506152027.AA47346@source.asset.com> Date: Tue, 20 Jun 1995 01:53:28 GMT In article <9506152027.AA47346@source.asset.com>, Claude L. Bullard \ wrote: >You correctly interpret my original question. Others have answered that >the multipart option is available but not well supported. Multipart file files are not often requested on the Web because references give the illusion of multiple parts. Obviously there are performance benefits to being able to package things in a single stream, but it is not as high on the user-demand list as are other features. To be clear, individual objects requested in a "multi-media" document are each negotiated. The only benefits that I can see to multipart mime encoding are a little performance and file maintenance ease. > Another issue >that I don't understand is how or if MIME supports cooperative peers: i.e., >applications passing commands to one another, a.k.a, push and pull. It may >be that a mail protocol simply isn't up to this. I've much to learn about >all of this. What you are discussing would seem more at home in HTTP. I am not sure what you mean by "commands" though. >Combining HTML with Java may be one approach to this, but it leaves only >the fixed widget types and the HTTP addressing and location techniques. As I undersatnd it, Java should allow you to discard _both_ the HTML "widget" set and HTTP. You should be able to plug in new applets and new protocols. Paul Prescod Newsgroups: comp.text.sgml From: papresco@calum.csclub.uwaterloo.ca (Paul Prescod) Subject: Re: Web Notations Message-ID: <3s5bku$b6r@calum.csclub.uwaterloo.ca> Sender: news@undergrad.math.uwaterloo.ca (news spool owner) Organization: University of Waterloo Computer Science Club References: <9506141830.AA45783@source.asset.com> <3rp5ql$jj4@news.cais.com> <9506152026.AA17483@source.asset.com> Date: Tue, 20 Jun 1995 02:26:38 GMT In article <9506152026.AA17483@source.asset.com>, Claude L. Bullard \ wrote: >The original question has one extra facet that I'm not completely clear on. >An SGML notation also indicates the flavor or version of the notation as >well as the helper application which services it. Can the MIME types and >the .mailcap conventions specify the flavor? You can specify parameters. I am not sure how one draws the line between types, flavors and parameters. Is HTML a type, or a flavor of SGML? "Type parameters Parameters on the content type are extremely useful for describing resolutions, colour depths, etc. They will allow a client to specify in the Accept: field the resolution of its device. This may allow the server to economise greatly on transmission time by reducing the resultion of an image, for example, and enable a more appropriate custom-designed black and white image to be selected rathther than giving the client a color image to convert into monochrome. These parameters are to be specified when types are registered." You might want to look at the document that came from: http://www.w3.org/hypertext/WWW/Protocols/HTTP >Another issue that becomes important is the ability to discern in advance >if the local system has the capacity to process the incoming file. For >HTML types, which are conventionally broken into reasonably sized chunks, >this may not be such a concern unless one has a very small cache. On the >other hand, once one adds scripts or MID applications which define complete >processing environments as well, having a few of these around in local >memory could be problematic. So we must still define ways to indicate the >processing environment required for the instance. Again, from the HTTP spec: mx The maximum number of bytes in the Entity-Body accepted by the client. The default value is mxb=undefined (i.e. infinity) This is one of the parameters that can be specified when you are accepting data. Paul Prescod Newsgroups: comp.text.sgml From: papresco@calum.csclub.uwaterloo.ca (Paul Prescod) Subject: Re: Web Notations Message-ID: <3s5ceh$cf6@calum.csclub.uwaterloo.ca> Sender: news@undergrad.math.uwaterloo.ca (news spool owner) Organization: University of Waterloo Computer Science Club References: <9506141830.AA45783@source.asset.com> <3rp5ql$jj4@news.cais.com> \ <3rv6mo$n4a@news.cais.com> Date: Tue, 20 Jun 1995 02:40:17 GMT > I thought this was what it did. I am really anxious to >get a demonstration (I talked with our Sun Salesperson yesterday). >I am wondering just how robust and interactive the programming >language is. I am thinking that HotJava can really be useful >for Electronic Technical Manuals (ETM) and provide more functionality >and interactivity that is necessary. Java is basically a simplified C++ that has been made "safe" from a computational and security point of view (or so they say...I've never been through the code). So it should be fairly robust and interactive. It is interpreted and portable, so it is not as fast as C++. > We know that Mosaic Communications (Netscape) has stated they are >going to support HotJava. My next question, could HotJava be useful >in a full-SGML environment DBMS->Electronic SGML Viewer to provide >full-function IETM capabilities? Java is a full programming langauge. It should be able to do whatever you feel like. Large values of "whatever you feel like" may require throwing away HTML, HTTP and doing a bunch of coding from scratch. Small values will probably be satisfied by freeware. > It should be an interesting and exciting time during the next >couple of years! I agree. Java is the kind of thing everybody talks about, and finally someone has done it. Java promises to be a free, portable, graphical, internet-aware development environment. It could change more than just the web. It has particularly interesting implictations for the SGML community. People will invariably want to put up complex, structured data (in addition to mindless tripe). Java allows them to create and deliver new file formats and viewers on demand. That puts us in a good position to offer SGML as an encoding format. Paul Prescod From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: Help with content model? Date: 20 Jun 1995 04:24:48 GMT Organization: ACM Network Services Message-ID: <3s5iih$o3c@hopper.acm.org> References: <3s4j3u$uo@argo.hks.com> Reply-To: davep@ACM.ORG In article <3s4j3u$uo@argo.hks.com>, jeffrey@hks.com (Glenda Jeffrey) writes: >I'm trying to put together a conten model like this: > > \ > >Because of the "no look-ahead" rule, this is an ambiguous content model. Use \ Hope this helps Dave Peterson SGMLWorks! davep@acm.org From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: Chemical Markup Language (CML) Date: 20 Jun 1995 04:30:56 GMT Organization: ACM Network Services Message-ID: <3s5iu0$o3c@hopper.acm.org> References: <803595163snz@ursus.demon.co.uk> Reply-To: davep@ACM.ORG In article <803595163snz@ursus.demon.co.uk>, Peter Murray-Rust \ writes: >1. Are comments (-- --) allowed within model groups in ATTLIST ( I hope I have >the terminology right). This would help greatly in documenting my DTDs, e.g: >\ filetype (smiles, -- a SMILES file from Daylight, Inc. -- > swissprot, -- a SWISSPROT protein sequence file -- > pdb -- A protein data bank structure -- > ) #REQUIRED > Sorry--no comments inside groups. Probably will be considered for the revision. BTW, they're "name token groups", not "model groups", at this point in the attribute definition list declaration. Dave Peterson SGMLWorks! davep@acm.org From: martyn davison \ Newsgroups: comp.text.sgml Subject: Re: GUI for DTD editing? Date: 20 Jun 1995 09:52:33 +0100 Organization: OMI Logistics Ltd Sender: news@news.demon.co.uk Message-ID: <66560144wnr@omimjd.demon.co.uk> References: <3rpght$ru8@usenet.srv.cis.pitt.edu> Reply-To: martyn@omimjd.demon.co.uk X-Newsreader: Newswin Alpha 0.7 X-Posting-Host: omimjd.demon.co.uk In article: <3rpght$ru8@usenet.srv.cis.pitt.edu> suthers+@pitt.edu (Daniel D Suthers) writes: > > GUIs for editing a document under a given DTD are coming out, but is anyone > offering a GUI for writing the DTD themselves? I agree with others on this thread. Try Near\&Far. It does not get your around the need for the analysis but it does help a lot creating the DTD -- Martyn Davison From: dld@chem.psu.edu (Drake Diedrich) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 20 Jun 1995 13:42:20 GMT Organization: Maui High Performance Computing Center Message-ID: <3s6j7s$1013@hearst.cac.psu.edu> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ In article \, Gordon Kent \ wrote: >Home page size (in k) >index.html: 9k >GIF Header: 108k >Misc. GIFS: 45k,36k,50k > >Total:248k It would be terribly easy to make disparaging comments about the author of this "typical" home page. Suggestion: look at ways of quantizing the colors in each gif. (pnmdepth, ppmquant) Shrink them a bit (pnmscale). Center the header. (\

...) Most gifs of this nature can be reduced to only a few K. Use jpeg instead of gif if these are photo-like. From: Kai Grossjohann \ Newsgroups: comp.text.sgml Subject: Re: SGML conversion utilities Date: 20 Jun 1995 16:58:43 +0200 Organization: University of Dortmund, Germany Message-ID: \ References: \ Reply-To: Kai Grossjohann \ In-reply-to: jjr@panini.att.com's message of Wed, 14 Jun 1995 22:47:15 GMT To: jjr@panini.att.com (Jeff Rocca), jjr@pez.att.com X-Newsreader: (ding) Gnus v0.86 >>>>> "Jeff" == Jeff Rocca \ writes: Jeff> I'm looking for UNIX-based utilities available in the public Jeff> domain to convert various formats to and from SGML (e.g., Jeff> SGML-to-LaTeX and LaTeX-to-SGML). Any pointers? Thanks. There is gf (General Formatter) which can convert from SGML to LaTeX, RTF, HTML and a number of other formats. Get it from ftp.th-darmstadt.de. I don't know about the other way 'round. \\kai{} -- Life is hard and then you die. Newsgroups: comp.text.sgml From: Steve Pepper \ Subject: SGML Tools and Vendors Message-ID: <1995Jun20.152402.17643@falch.no> Organization: Falch Infotek, Oslo, Norway Date: Tue, 20 Jun 1995 15:24:02 GMT I have now posted a new version of The Whirlwind Guide to SGML Tools and Vendors at the following locations: ftp://ftp.falch.no/pub/sgmltool/sgmltool.txt http://www.falch.no/~pepper/sgmltool/ The PostScript version of the listing is being discontinued. The web version, on the other hand, will now be updated as regularly as the text version (i.e. about once a month, or whenever I have gathered enough new information to justify a new edition). As always, I am very grateful for suggestions and updated information. I would particularly urge vendors to take the time to check the accuracy of the information about their companies and products (but please read the guidelines carefully first!) Best regards, Steve -- pepper@falch.no -------------------------------------------------------------- falch infotek a.s, postboks 130 kalbakken, n-0902 oslo, norway tel +47 2290 2733 fax +47 2290 2599 http://www.falch.no/ ^^^^^^^^^^^^ NEW ^^^^^^^^^^^^ Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 20 Jun 1995 13:11:57 -0400 From: "Claude L. Bullard" \ Message-ID: <9506201711.AA13777@source.asset.com> References: <9506092043.AA37374@source.asset.com> \ Subject: Re: Blessings and Curses [Sebatian Rahtz] | I think this is historically a rather strange view. HTML came from | CERN, from professional computer scientists employed by a very large | international research organization. Marc Andreessen and NCSA may | have made it popular and useable with Mosaic, but they didnt invent | HTML, or substantially alter its character. Just to be clear, the quote which you attribute to me is a quote from a magazine article that I did not author. No problem but since they archive these posts... |-) Apologizing in advance for the mixed metaphors herein... Yes, what you state are the essential facts. What is odd to me is how many people buy into the myth that the internet utilities such at the WWW are being created by "lone hackers". In many cases, they are being created by powerful organizations who contribute freeware to the community which then "hacks" at the core. Unfortunately, as the twig is bent so grows the tree. This "freeware" starts a snowball whose course is difficult to alter. One has to pray the snowball is well-formed and well-aimed. So far, the WWW is proving to be a nice warm-up act for a main act yet to be introduced. Maybe Panorama; maybe Sculptor; maybe not. The key issue is the middleware protocol handling. Yes, Marc Andreesen worked on Mosaic. His corporation under the tutelage of a former Silicon Graphics exec, James Clark, (not to be confused with the writer of SGMLS), was also the first to add private extensions to HTML thus invalidating the freeware they created and the myth of "the anarchic good will" of freeware. SGI contributed formats to the VRML movement, and thus, they are able to get a leg up on other 3D-Web browsers. This movement by industry to "colonize" the Internet is both a polite and cunning way to convince the public that something done for private interests is in the public interest. Sometimes it is. It certainly isn't wrong. It is clever capitalism. But, let's be honest here, it is still Geeks Bearing Gifts. The next version of the freeware is always commercial. As with any acquisition, shop around. There were three points in the original article: 1. Myths are bad history but sometimes useful. The media only want good stories. Facts are for historians and engineers, not journalists and marketeers. Caveat emptor. But we need facts. 2. There is a groundswell of forces with various origins and intentions who are realizing that the current WWW is a weak design whose momentum threatens the global network industries just as it helps to jumpstart them. In chaos is opportunity or disaster. HTML is a blessing and a curse particularly to SGML vendors who in their race to reap long awaited profits, may only strangle the goose by not acknowledging the goose lays golden eggs but also bites. Caveat vendor. But we need sales. 3. If the WWW and the W3O are to realize any of their original intentions, they must compete with good designs. They will be well-served by research that considers more than a design for a mail system. They will be well-served by considering other standards. They will be well-served if they know more about sound document design. Their knowledge of network communications and protocol design is sound. They are fine programmers. They are terrible document designers. But they got the party into high dudgeon, and I thank them for that. They deserve respect. They should accord it as well. We all deserve that. Len Bullard Break the Chain of Mediocrity. Support Your Local DTD Designer. From: Andrew Haigh \ Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 20 Jun 1995 19:10:25 GMT Organization: Imagen Communications Inc. Message-ID: <3s76f1$5td@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 32bit) If you are looking for a file format that is so compact that it blows away PDF for bandwidth conservation, transmission speed, and graphic retention, check out RTC, Rich-Text-Compressed, a page format developed by Imagen Communications Inc. Check out our home page listed below for examples of the files, and the free-viewer (which also blows Acrobat out of the water - 170K zipped as opposed to 1.4megs) -- Andrew Haigh Imagen Communications mailto:ahaigh@imagen.net http://www.imagen.net Newsgroups: comp.text.sgml Approved: erik@naggum.no Message-ID: <9506202221.AA0720@notes.microstar.com> From: Hugh Chatfield \ Date: 20 Jun 1995 15:19:28 -0400 References: <3s4j3u$uo@argo.hks.com> Subject: Re: Help with content model? Glenda Jeffrey said: > I'm trying to put together a conten model like this: > > \ > > Because of the "no look-ahead" rule, this is an ambiguous content model. Erik Naggum has suggested an equivalent content model: > \ I took an approach which tried to deal with the underlying information model which is being expressed in SGML. First Glenda Jeffrey asked: > Can anybody explain to me why this is such a Bad Thing to want to do that > the standard does not permit it? It is not bad - just ambiguous. Since all the elements are optional, then you are saying the following are all valid instance strings (with spaces added to the explanation to try to clarify for us humans which B we are talking about): 1) BBB (B BB) 2) ABCB (A B C B) 3) ABBBCBBBB (A BBB C BBB) 4) BBBCBBB (BBB C BBB) 5) AB (A B) 6) B ( B) i.e the second B In all examples the parser cannot know whether the first B belongs to the 'B's following the A or the 'B's following the C (not even when the B follows an A). To fix it for the parser, you need a mechanism which will allows the two B's to be distinguished (like the space tries to do for us humans in my examples). If I presented you with option 1 or 6 without the space, I presume you couldn't tell which B it was either. I suggest your model is perhaps incomplete - it is missing that semantic which would allow the authors (and the parser) to distinquish between the two B's. >I mean, this is really the structure I have -- the order of A and C must >be preserved, and all the items really are optional... Which gives a clue, that what you really have is two higher level abstractions, an optional A thing followed by a optional C thing, whose order must be preserved. I assume these things have meaning in your world and the authors could recognize them. I further assume (like Erik) that the distinguishing feature of the A thing and the C thing is that if they exist at all they must contain an A and C respectively. In which case the following model should be something your authors would recognize and relate to. \test-- > \C thing-- > \A thing-- > or if the tag names are not really relevant or useful, then we get back to a model similar to what Erik proposed. \ Hope this helps. Cheers...Hugh -- W. Hugh Chatfield I.S.P. Phone: +1 613 727-5696 x317 /\\ /\\ Fax: +1 613 727-9491 \\/ \\/ Computer Microstar Software Ltd. WATS: 1 800 267-9975 /\\ /\\ Aided 34 Colonnade Road North E-mail: hchatfie@microstar.com \\/ \\/ Document Nepean Ontario Info: cade@microstar.com /\\ /\\ Engineering CANADA K2E-7J6 List: cadelist-request@ \\/ \\/ Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 20 Jun 1995 17:42:44 -0400 From: "Claude L. Bullard" \ Message-ID: <9506202142.AA51434@source.asset.com> Subject: The Protocol Game Thanks to everyone for the public and private email I've received on the subject of notations and protocols. It is all very helpful. This much is certain, without strong agreements for hypermedia middleware standards, we are going nowhere quickly. We can look at HyTime and discover if the structures for client-server interaction are there for predefined application protocols. IMO, it will work for SGML applications. The question is whether it will work for other notations. There are really good minds on the CTS. Someone said it was inspiring to get up and do something rather than wait for the powers-that-be to do it. I said that was free research for the powers. It's true, but SFW. I'm a dodo. Instead of debating or kicking PDF and HTML in the teeth every day, anyone game for designing a HyTime application protocol to explore the issues of using HyTime in a model of a client-server architecture? For example, what is needed to pass an independent link (ilink) in the message body of a server request? Careful. The ilink is an archform or metatype. An instance uses a generic identifier to infer the processsing semantic. Therefore, middleware for processing a HyTime linktype must select among the HyTime types AND it or the server must select among the application types as well. Here is an instance of a hytime ilink called \. When a client passes a get to the server, the server collects and concatenates the data. The client processes the data at the point of the get in the client script. The space attribute is used to tell whether members of an aggregate should be delimited by a single space. Client processing is determined by the return type of the target data. \ \ Notice the entity %locs. In HyTime, a link probably points to a location type. There are several and the application designer is expected to choose the ones they want to implement. One might use an entity to group them such as: \ For this example, the location type is a named location. It looks like this: \ \ The name list will include an IDREF(s) for an entity which may also have a notation declaration associated with it. NOTE: If I have any of this out of whack, please correct me. (Eliot??) That's it for the HyTime for the moment. HyTime cognoscenti know that I removed the HyQ query type from the nameloc. It isn't needed for the example, but document querying is very important as the queries can use property definitions and that is how one might construct catalogs for document databases. But, walk first. The first order of business is to figure out how to pack this information into a message packet and sort it out at the server. Let's look at the code to begin defining the message packets. The following code is from the Novell book (ISBN 0-7821-1248-X) written by Jeffrey D. Schank and modified by me. As Schank uses C, so does this. Coding this into a communications class would be another exercise. As archforms are metaclasses for SGML, this could be important. |-) Correct this as you see fit. For Request/Response, the message protocol is predefined and requires a structured packet as shown: typedef struct HyTMsg MSG; typedef struct _hylink { WORD clientID; // comm layer fills. WORD linkType; // comm layer fills. // denotes server function LONG returnCode; // server fills. // contains return code of server operation. MSG linkData; // message passed/returned } HYLINK; Note that the linkData is the message filled for server transmission. Now we need a handler for the application protocol. MSG HyTMsgHandler(QUEUE serverQueue, BYTE executionMode, WORD linkType, MSG *dataMsg) { HYLINK *hylink; // process protocol structures hylink = (HYLINK *) alloc(sizeof(HYLINK)); hylink.linkType = linkType; // copy dataMsg into hylink memcpy(\&hylink.linkData, dataMsg, sizeof(MSG)); // send to link server CallLinkServer(serverQueue, hylink) if (executionMode & SYNCHRONOUS) return(GetLinkTarget(serverQueue, linkType)); } Anyone know if this works? Of course, a switch to dequeue link types and dispatch them to linktype handlers would be nice and freeing up the memory is also nice. What else? Anyone care to recode this into a C++ class capable of handling multiple hytime and application link types and the location types? LISP or Scheme examples are welcome. But, if you do that, keep a copy in both languages. Language bindings prove the abstract nature of the spec. Len Bullard From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: Help with content model? Date: 20 Jun 1995 16:12:06 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3s7kk6$jfl@crl6.crl.com> References: <3s4j3u$uo@argo.hks.com> <9506202221.AA0720@notes.microstar.com> >Glenda Jeffrey said: >> I'm trying to put together a conten model like this: >> >> \ >> Hugh Chatfield \ wrote: > I took an approach which tried to deal with the underlying information model > which is being expressed in SGML. [ followed by an excellent analysis, suggesting either ] > \test-- > > \C thing-- > > \A thing-- > [ or ] > \ [Back to Glenda Jeffrey] >>I mean, this is really the structure I have -- the order of A and C must >>be preserved, and all the items really are optional... Hugh's suggestion is a good one if the order of the Bs relative to the A and the C is important, since it accurately reflects the logical structure. If on the other hand the order of the B elements *doesn't* matter, and you only want to allow them anywhere in the content model for user convenience, I would suggest constraining them to appear in a fixed place: \ or \ If you later want to add some new elements, say a D (which has to come after A and C) and an E (which can appear multiple times, but the order doesn't matter), then this: \ will turn into this: \ and it will only get worse from there... If you fix the order up front, future DTD maintenance will be more tractable: \ (Again, this is only if the order of the B elements relative to the others is not significant; otherwise Hugh's or Erik's suggestions are better ones.) There was recently a very similar situation in the HTML working group regarding the content model of HEAD. The HEAD element in HTML contains * a TITLE element (required), * a BASE element (optional) * a few other optional elements, * zero or more LINK elements, and * zero or more META elements, all in any order. The order really *doesn't* matter in this case, but to support legacy documents (which tend to have LINK and META sprinkled all over the place), the DTD ended up using a rather awkward content model, with inclusion and exclusion exceptions that don't reflect the logical structure at all. --Joe English jenglish@crl.com From: Larry Mull \ Newsgroups: comp.text.sgml Subject: Need Book on Structured Authoring Date: 20 Jun 1995 23:54:17 GMT Organization: Mentor Graphics Corp. Message-ID: <3s7n39$spn@hpmos.wv.mentorg.com> X-UserAgent: Version 1.1.3 X-XXMessage-ID: \ X-XXDate: Tue, 20 Jun 95 00:02:44 GMT I'm looking for a good book and/or articles on structured authoring. Specifically not on the SGML syntax, but on understanding how to effectively author for SGML. Any help and direction would be appreciated! Thanks, Larry larry_mull@mentorg.com From: AJTowle@wco.com (AJT) Newsgroups: comp.text.sgml Subject: Re: SGML - HTML newbie looking for input Date: Wed, 21 Jun 1995 01:51:12 GMT Organization: WEST COAST ONLINE Message-ID: <3s7tqi$20k@news.wco.com> References: <3s2m86$8iq@kettle.magna.com.au> Reply-To: AJTowle@aol.com X-Newsreader: Forte Free Agent v0.55 tactics@magna.com.au (David Churchland) wrote: >My problem is that I am having to battle the Windows help proponents who >love the low initial co$t of using the help engine. The primary cost either way will be authoring. Though the help engine is free in Windows, documents must be crafted to get through the help compiler just as documents must be crafted to by viewed with a low-priced (or free) browser. Crafting links/jumps takes a similar amount of planning with both. If you must also have a published document, you should consider using MS-Word. MS offers the (free) MS-Word "Internet Assistant" which will use MS-Word formatting to make VERY BASIC HTML docs. There are at least 2 MS-Word-to-help applications which operate in a similar manner, Doc-2-Help and RoboHelp. --Bert Towle From: tja@phobos.kbs.citri.edu.au (Tim Arnold-Moore) Newsgroups: comp.text.sgml Subject: HMSO DTDs for legislation Date: 21 Jun 1995 12:19:43 +1000 Organization: CITRI Message-ID: \ Reply-To: tja@citri.edu.au I'm doing some research on encoding legislation in SGML and was wondering if there is an easy way of getting hold of the DTDs which HMSO (the UK publisher of legislation) has constructed for statutes and statutory rules and regulations. FTP, or even an email address, contact name, Fax or phone number would be handy. -- Tim Arnold-Moore (LL.B - Melb) | Multimedia Database Systems, CITRI | tja@citri.edu.au | 723 Swanston St ---------------- Phone: +61 3 9282 2487 | Carlton 3053 | simul iustus Fax: +61 3 9282 2490 | Victoria, Australia | et peccator From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: Help with content model? Date: 21 Jun 1995 04:16:16 GMT Organization: \ Message-ID: <3s86eg$41u@news1.delphi.com> References: <3s4j3u$uo@argo.hks.com> X-Newsreader: Galahad 1.1b > I'm trying to put together a conten model like this: > > \ > > Because of the "no look-ahead" rule, this is an ambiguous content model. Since you have already considered a form of "wrapper" let me make a suggestion that works on this and other similar situations. Instead of one wrapper, use a wrapper around each section with optional elements: \ \ \ \ \ ]> Now you need to make two consistant changes in your tagging. You never use just a \, you always use the combination \\. Also you never use just \, you use the combination \\. You no longer have ambiguous content. Also the changes to tagging can be done programatically. Using the definitions above the following parses: \ \\\A\\ \\\B\\B\\ \\\A\\B\\B\\ \\\\C\\ \\\A\\\C\\ \\\B\\B\\\C\\ \\\A\\B\\B\\\C\\ \\\B\\B\\ \\\A\\B\\B\\ \\\B\\B\\B\\B\\ \\\A\\B\\B\\B\\B\\ \\\\C\\B\\B\\ \\\A\\\C\\B\\B\\ \\\B\\B\\\C\\B\\B\\ \\\A\\B\\B\\\C\\B\\B\\ \ This method also works if you need to something even more complex like: \ \ \ \ \ \ ]> ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. From: votetaker@f10.facts.uky.edu (Kenneth Herron) Newsgroups: news.announce.newgroups,news.groups,comp.text,comp.text.desktop,comp.text.sgml,comp.text.tex,comp.text.interleaf,comp.text.frame,comp.publish.cdrom.software,comp.publish.cdrom.multimedia,comp.multimedia Subject: RESULT: comp.publish.electronic.{misc,developer,end-user}: All pass Supersedes: \ Followup-To: news.groups Date: 21 Jun 1995 01:22:05 -0400 Organization: UUNET Technologies Inc, Fairfax VA USA Sender: tale@uunet.uu.net Approved: tale@uunet.uu.net Message-ID: \ References: <3qaumu$r91@rodan.UU.NET> \ \ RESULT comp.publish.electronic.{misc,developer,end-user} results - 207 valid votes Yes No : 2/3? >100? : Pass? : Group ---- ---- : ---- ----- : ----- : ------------------------------------------- 171 27 : Yes Yes : Yes : comp.publish.electronic.misc 173 25 : Yes Yes : Yes : comp.publish.electronic.developer 168 28 : Yes Yes : Yes : comp.publish.electronic.end-user 6 invalid votes There is a five day discussion period after these results are posted. If no serious and significant allegations of voting irregularities are raised, the moderator of news.announce.newgroups will create the newsgroups shortly thereafter. This vote was conducted by a neutral third party. For voting questions only contact votetaker@f10.facts.uky.edu. For questions about the proposed groups contact Mike Gilpin \. Newsgroups lines: comp.publish.electronic.misc General electronic publishing issues. comp.publish.electronic.developer Electronic publishing developer tools. comp.publish.electronic.end-user Electronic publishing end-user tools. CHARTERS These three groups are intended for those interested in the creation of "serious" publications in electronic form, the tools available, and format issues. comp.publish.electronic.misc comp.publish.electronic.misc is an unmoderated newsgroup created with the purpose of posting general information and discussions items regarding Electronic Publishing issues. Electronic Publishing is taken to mean the publication of documents, images (still and moving), and sound in a digital format for use with computer based viewing applications. comp.publish.electronic.developer comp.publish.electronic.developer is an unmoderated newsgroup created with the purpose of posting information and discussion items relating to tools used to create and develop electronic publishing data collections. This can include discussion on tools used to create the media, as well as tools used to insert and manage hyperlinked data. Current examples are Director, Adobe Distiller, CD-ROM formatters, Authorware, Toolbook, etc. comp.publish.electronic.end-user comp.publish.electronic.end-user is an unmoderated newsgroup created with the purpose of posting information and discussion items relating to applications used by end-users to view collections of electronic publishing data. This typically means viewing software. Current examples are Common Ground, Acrobat, WorldView, etc. Also, this group should be a forum for discussion of end-user products, particularly where they are technologically advanced. For example Microsoft Encarta. However, games discussions should be kept to comp.publish.cdrom.software and discouraged here. comp.publish.electronic.end-user -----+ comp.publish.electronic.developer ----+| comp.publish.electronic.misc ---+|| ||| 75052.207@CompuServe.COM Lela Whisnant YYY aa699@cleveland.Freenet.Edu Tony Thomas YYY aaron@chrysopylae.com Aaron Priven YYY admmss@icbl.heriot-watt.ac.uk Martin S Smith YYY ahdavidson@aol.com Andrew Davidson YYY Alden.Klovdahl@anu.edu.au al klovdahl YYY AltesJ@USA.RED-CROSS.ORG James Altes Y-Y altmann@slspyr Monika Altmann YYY amyk@ns.moneng.mei.com Amy Karls Y-Y andrew@haslam.win-uk.net Andrew Haslam YYY andrew_mcrae@harvard.edu Andrew McRae YYY arromdee@blaze.cs.jhu.edu Ken Arromdee N-- asandler@encore.com Arlene Sandler YYY ault@cs.albany.edu Jim Ault YYY b.wage@elsevier.nl Bart Wage -YY bailey@hagar.ph.utexas.edu Ed Bailey YNN bajema@alc.com Bruce D. Bajema YYY bducharme@riasmtp.riatax.com Bob DuCharme YYY bessler@highland.swb.de Wolfgang Bessler YYY billo@grinch.HQ.ileaf.com Bill O'Donnell YYY bjv@herbison.com B.J. Herbison Y-- books@inforamp.net Erich Volk YYY bouillon@cs.uni-sb.de Peter G. Bouillon YYN brent.bailey@mosby.com Brent Bailey YYY brucec@ix.netcom.com Bruce Covell YYY by177@cleveland.Freenet.Edu D. L. Mosher NYY campbelg@limestone.kosone.com Gordon Campbell YYY carey.smith@natinst.com carey smith YYY caseyt@teleport.com Casey Tichenor YYY cdupree@qcktrn.com Chuck Dupree YYY Chris.P.MILLER@chetco.wrd.state.or.us Chris miller YYY christy.davis@srs.gov Christy S. Davis YYY cifsab@inet.uni-c.dk Anders Bjerre ( CIFS/Copenhagen ) YYY ckkang@iti.gov.sg Kang Chwee Kiat YYY claird@Starbase.NeoSoft.COM Cameron Laird YYY clstampe@mailbox.syr.edu Chruis Stamper -YY crm@ebt.com Christopher R. Maden -Y- crouchkp@flidh101.delcoelect.com K. Crouch --- csaamw@urc.tue.nl Michiel Wijers NNN cverver@u.washington.edu Craig Verver YYY cward@Think.COM Christopher Ward NNN dannya@xs4all.nl Danny A. Nijburg YYY dapst26+@pitt.edu Daniel Ponech YYY dcitron@pts.mot.com David Citron YYY deborah.board@sap-ag.de Deborah Board YYY delancey@herbie.unl.edu Andrew DeLancey YYY dhartung@mcs.com Dan Hartung YYY diermann@mad.pb.owl.de Markus Diermann YYY DKnowles@po.asm-intl.org Doug Knowles YYY donday@bga.com Don R. Day YYY dorai@ses.com Dorai Sitaram YYY egarbos@sanders.com Eric Garbos YYN eggert@twinsun.com Paul Eggert NNN ELLISORJ@mtc2.mid.tec.sc.us James D. Ellisor YYY export4@aol.com Michael Hammer YYY fatman@cpcsat.sfos.ro Catalin Gras -YY faubert@CAM.ORG Luc Faubert YYY fh@dune.swb.de Frank Haefemeier NYY fishe@casbah.acns.nwu.edu Carwil James YYY fredw@fred1.demon.co.uk Fred Weil YYY frees@castle.net Frank Rees YYY fuy1@gl.umbc.edu Francis Uy Y-- geert.bryon@infoboard.be Geert Bryon YYY geo@knex.mind.org Geo Singlemann YYY gess@knex.mind.org Gess Shankar YYY giniels@aol.com Niels Nielsen YYY gordon@yuki Gordon Engel NNN grath@math.duke.edu Steve Grathwohl YYY greaber@reed.edu Grant Reaber YYY gt@ee.latrobe.edu.au Geoff Tobin YYY Hans-Walter.Gabler@engliph.uni-muenchen.d400.de Hans Walter Gabler YYY harvey.forman@bardo.com HARVEY FORMAN YYY hh@hhdo.do.home.venture.net Henning Holtschneider YYY hsauer@marnie.teuto.de Henning Sauer YYY hug@faw.uni-ulm.de Klaus Hug YYY hyphen@ibmPCUG.CO.UK Dave Fawthrop YYY info@love.campbell.ca.us Lance Chun YYY irina@rempt.xs4all.nl Irina Rempt YYY isg@bball.HQ.ileaf.com Ira Gerstein YYY J.Goldberg@cranfield.ac.uk Jeffrey Goldberg YYY jabbott@picasso.dehavilland.ca John Abbott YYY james.hart@chrysalis.org James Hart YY- jcohen@DRI.cornell.edu Jeffrey David Cohen NYY jeb@cris.com K. Larson YYY jfortt@dorsai.dorsai.org Joseph Fortt YYY jhammes@PrimeNet.Com jack hammes YYY jmcgarry@uoguelph.ca James McGarry YYY jme@c-art.com Yukon Kid NNN jocasta@ra.isisnet.com Betty Dobson YYY john@prism.nmt.edu John W. Shipman YYY jon@sep.Stanford.EDU Jon Claerbout YYY jonathan@book.demon.co.uk Jonathan Peachey YYY jonboy@io.org Trevor Tymchuk YYY jrm@globalvillag.com John R. MacWilliamson NNN jsuttor@interramp.com Jeff Suttor NNN juergen.bachteler@itap.physik.uni-stuttgart.de Juergen Bachteler YYY justin@pendred.demon.co.uk Justin Cormack Y-- karl@ppe.ppd.ti.com Karl Rasmussen YYY katnagel@EZNET.NET Kat Nagel YYY khan@JTP.CS.NYU.EDU Hasnain Khan NNN kieran@interport.net Aaron Dickey YYY kim@shell.portal.com Kim DeVaughn YYY Klaus.Rehm@extern.uni-regensburg.de Klaus Rehm YYY kmc@specialform.com Keith M. Corbett YYY krishnan@engr.LaTech.edu Ashok Krishnan YYY lauraj@nsmdserv.cnd.hp.com Laura Johnson NNN len@contec.COM Leonard Mills NYN lenoil@catalogic.com Robert Lenoil NNN lf7z@midway.uchicago.edu Glenn Carnagey YYY lisa@gordian.com Lisa Leone NYY list-votes@dream.hb.north.de Martin Schr\\"oder YYY Lutz.Donnerhacke@Jena.Thur.De Lutz Donnerhacke YYY M.J.Elphick@ncl.ac.uk Michael Elphick YYY m1mth01@FRB.GOV Mary Hoffman YYY Maggioni@aol.com Suzanne Maggioni YYY mal@infinet.com Mark A Lynch YNY marc@ALGORITHMICS.COM Marc Moorcroft NNN marpirfv@individual.puug.pt Mario Filipe Pires YYY martin@hchworth.demon.co.uk Martin Dransfield YYY martyn@omimjd.demon.co.uk Martyn Davison YYY matsd@physchem.kth.se mats dahlgren YYY mau@beatles.cselt.stet.it Maurizio Codogno YYY mb@arachnid.com Mary Beth Schulte YYY mcevilly@netcom.com Carlos McEvilly YYY mcjones@pa.dec.com Paul McJones YYY ME@addix.swb.de Michael Ellis YYY mg@cosmos.swb.de Michael Grundel YYY Mike@expedite.demon.co.uk Michael Cox YYY mjg@io.com Mitchell J. Gross YYY mmt@RedBrick.COM M Mike Taksar --- mooring@grimoire.Tymnet.COM Ed Mooring YYY mourik@python.konbib.nl Hans van Mourik (Koninklijke Bibliotheek, NL) -YY mox@vpnet.chi.il.us William Moxley YYY mthul@netaxs.com Matt Thul YYY mtr@addix.swb.de Mark Rudolph YYY MWOOD@INDYVAX.IUPUI.EDU Mark H. Wood YYY neubert@lusi.uni-sb.de Boris Neubert YYY nick@inferno.fc.hp.com Nick Ingegneri YYY norman@astro.gla.ac.uk Norman Gray YYY Ollivier.Robert@hsc.fr.net Ollivier Robert YYY orc@pell.com david parsons NYY ornberg@mcs.com Rick Ornberg YYY otto@vaxb.acs.unt.edu M. Otto --- papresco@calum.csclub.uwaterloo.ca Paul Prescod YYY pepper@falch.no Steve Pepper YYY pgb@arbortext.com P.G. Bartlett YYY phase3@gate.net Don Fasig YYY phoffman@proper.com Paul Hoffman YYY PHOTOPRES@aol.com Jeff Schmidt (photopres@aol.com) YYY phughes@infohwy.com Peter Hughes YY- pierre@shell.portal.com Pierre Uszynski NNN pivo@cipserv.ikm.uni-mannheim.de Andreas Ruf YYY pjotr@pop.jaring.my Pjotr van Schothorst YYY poynton@inforamp.net Charles A. Poynton NNN ptbast@owl.WPI.EDU Pete Bastien NNN qramark@lmera.ericsson.se Markus Berg YYY quinlan@best.com Daniel Quinlan NNN randrews@cas.org Brad Andrews YYY ranfry@python.magnet.ca Marc Lombart YYY rbrown@ctsnet.cts.com J. Reese Brown YNY rcimek@globalx.net Ryszard Cimek YYY rew@CrystalData.COM Ryan Waldron NNN ricday@mbfw451.mlnet.com Richard S. Day YYY richa050@gold.tc.umn.edu Mark E. Richards YYY rick@bcm.tmc.edu Richard Miller YNN rieger@bse.de Wolfgang Rieger YYY rms@arachnid.com Richard Stucky YYY rtfm@cyberspace.com Dan Thies YYY rueger@netaccess.on.ca Barry Rueger YYY russw@powergrid.electriciti.com Russell Wagner YYY rutie@asp.co.il Rutie Adar YY- s639823@aix2.uottawa.ca Michel LeSann YYY schorsch@pz-oekosys.uni-kiel.d400.de Georg Hoermann YYY shawn@iglou.com Shawn Curry YYY silveste@falch.no Silvester Pozarnik YYY Singlis@bcde.demon.co.uk Steven Inglis YYN SKIRS@dbsoftware.com Susan Kirshbaum YYY skn@hut.fi s. k. niiranen NYN slb@fs.com.au Stuart Burnfield YYY slcramer@cuok.cameron.edu Sheryl L. Cramer YYY snow@crl.com Aaron J. Snow YYY SPB@tntv1.ntrs.com Steve Bonine YYY ssacks@InterLink.NET Steven Sacks YYY stainles@bga.com Dwight Brown NNN stevenn@kluwer.be Steven Noels YYY stewaeh@alpha.unisa.ac.za Elizabeth Stewart YYY stuart@cosc.canterbury.ac.nz Stuart Yeates YYY studio.mb@magnet.at studio.mb@magnet.at YNY tbas@sybase.com Tony Basoglu NNN tccn@tccn.com Coralie Allen YYY tgromak@kilroy.id.net Tom Gromak Jr. YYY thilo@asciel.deceiver.org Thilo Pfennig YYN tim@canon.co.uk Tim O'Donoghue YYY tmodica@connix.com Ted Modica YYY tobiasr@stud.cs.uit.no Tobias Rischer YYY tprodin@cadcam.pms.ford.com Tim Prodin YYY tra@tpone.telepac.pt Tiago Rodrigues Antao YYY twells@netcom.com Tabor J. Wells YYY twpierce@midway.uchicago.edu Tim Pierce YNN urlichs@smurf.noris.de Matthias Urlichs -YY us006471@interramp.com Valerie Cline YYY uthke@ite.iabg.de Ekkehard Uthke YYY weyer@netaxs.com Steve Weyer YYY widenius@cc.helsinki.fi Risto Widenius NNN WiN@addix.swb.de Robert Winter YYY Winfried.Mueller@Jena.Thur.De Winfried Mueller YYY young@rizer.computize.com Bill Young YYY Votes in error ------------------------------------------------------------------------------ facteur@paulmont.demon.co.uk Paul Montgomery ! No vote statement in message hal@enterprise.advance.com Hal Hammond ! No vote statement in message hmiller@rosie.Read.TASC.COM Hal Miller ! No vote statement in message keie@cs.vu.nl EG Keizer ! No ballot munro@ncsu.edu Molly Munro ! No vote statement in message tb@HTWM.DE Thomas Beierlein ! No vote statement in message Newsgroups: comp.text.sgml From: jta_swets@hgl.signaal.nl (Swets J.T.A.) Subject: Metaphase2.0 implementations of SGML doc.man systems? Message-ID: <1995Jun21.062058.25075@hgl.signaal.nl> Keywords: compound document database systems Sender: usenet@hgl.signaal.nl (USENET News System) Organization: Hollandse Signaal apparaten BV X-Newsreader: TIN [version 1.2 PL2] Date: Wed, 21 Jun 1995 06:20:58 GMT I would like to know if there are people in the world who are aware of, or even have experiences with implementations of SGML document management systems based on a tool called "Metaphase 2.0", What SGML document management support functionality is standard available and what customizations in order to support SGML document management have been implemented? Metaphase is a Product Data Management package running on several platforms. Vendor of this tool is SDRC / ControlData. =============================== Unclassified ================================== J.T.A. Swets Dept. DSA-PIM email: jta_swets@hgl.signaal.nl Research, Development & Technology tel: (+31)-(0)74-482539 (work) Document Technology Consultant Fax: (+31)-(0)74-484009 -------------------------------- Hollandse Signaalapparaten BV PO Box 42, 7550 GD Hengelo The Netherlands. =============================== Unclassified ================================== Newsgroups: comp.text.sgml From: robh@cwi.nl (Rob Hoksbergen) Subject: AAP DTD: article & maths (+NOTATION, etc used by these) needed Message-ID: \ Keywords: AAP DTD Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam Date: Wed, 21 Jun 1995 06:23:04 GMT HI, I have been unable to locate the AAP DTD's for article and maths on the network (they are not in the SGML repository at ftp.ifi.uio.no, as far as i know). Is there anyone who could mail these to me? Also, i would need, of course, the NOTATION (and possibly others) that these DTD's call for. A pointer to a location where i could ftp them from myself would also be convenient. Thanks in advance. Rob From: mandel@calefon (Luis Mandel) Newsgroups: comp.text.sgml,comp.sys.next.software Subject: SGML Browser Date: 21 Jun 1995 09:34:09 GMT Organization: Institut fuer Informatik der Universitaet Muenchen Message-ID: <3s8p2h$3nv@arcadia.informatik.uni-muenchen.de> X-Newsreader: Alexandra.app (Version 0.8) I would like to know if it is available any SGML browser for NeXT. Thanks, Luis --- +-------------------------------------------------------------------------+ | /// Luis Mandel mandel@informatik.uni-muenchen.de | | (. .) http://www.pst.informatik.uni-muenchen.de/~mandel | +oOO-(_)-OOo--------------------------------------------------------------+ -- +-------------------------------------------------------------------------+ | /// Luis Mandel mandel@informatik.uni-muenchen.de | | (. .) http://www.pst.informatik.uni-muenchen.de/~mandel | +oOO-(_)-OOo--------------------------------------------------------------+ From: jang@xs4all.nl (Jan Grootenhuis) Newsgroups: comp.text.sgml Subject: Re: GUI for DTD editing? Date: 21 Jun 1995 10:59:42 GMT Organization: XS4ALL, networking for the masses Message-ID: <3s8u2u$i3j@news.xs4all.nl> References: <3rpght$ru8@usenet.srv.cis.pitt.edu> <66560144wnr@omimjd.demon.co.uk> X-Newsreader: NN version 6.5.0 #666 (NOV) X-Posted-By: jang@xs1.xs4all.nl martyn davison \ writes: [...] >I agree with others on this thread. Try Near\&Far. >It does not get your around the need for the analysis but it >does help a lot creating the DTD If that's the sort of help you need, you should not be creating DTDs in the first place. BTW Near\&Far does not even support the reference Concrete Syntax in NAMECASE GENERAL YES ENTITY NO, let alone any changes in the other NAMING aspects. Cheers, From: bruce@sgml.dircon.co.uk (Bruce Hunter) Newsgroups: comp.text.sgml Subject: Re: Browser for HTM or SGML files on a LOCAL PC Date: 21 Jun 1995 11:51:48 GMT Organization: SGML Systems Engineering Message-ID: <3s914k$2cc@newsgate.dircon.co.uk> References: <3ropjq$hmv@imp.demon.co.uk> X-Newsreader: WinVN version 0.81 In article <3ropjq$hmv@imp.demon.co.uk>, dmatthewson@cix.compulink.co.uk (David K Matthewson) says: > >Does anyone know of a browser for use in a non net environment - ie: >for viewing local files of a disc? The SGML viewer we are developing is (in its simplest mode) a local file viewer. The runtime viewer will be free, and it will come set up to view html files. Unfortunately, as it uses a true SGML parser, the html files will have to conform to a DTD :-( Contact me by email or phone (+44 934 822911) for further details. regards, Bruce Hunter From: Kai Grossjohann \ Newsgroups: comp.text.sgml Subject: Re: SGML - HTML newbie looking for input Date: 21 Jun 1995 14:26:33 +0200 Organization: University of Dortmund, Germany Message-ID: \ References: <3s2m86$8iq@kettle.magna.com.au> Reply-To: Kai Grossjohann \ In-reply-to: tactics@magna.com.au's message of 19 Jun 1995 02:09:10 GMT To: tactics@magna.com.au (David Churchland) X-Newsreader: (ding) Gnus v0.87 >>>>> "David" == David Churchland \ writes: David> - are there any good (freeware/shareware) indexing/text David> retrieval tools for searching multiple HTML documents, There is freeWAIS-sf (to be had from the mirros of ls6-ftp.informatik.uni-dortmund.de): ,----- | Please get freeWAIS-sf from a mirror site: | | ftp.germany.eu.net | /pub/infosystems/wais/Unido-LS6/freeWAIS-sf-1.1/ | ftp.switch.ch | /mirror/wais/servers/sf-alpha/freeWAIS-sf-1.1/ `----- Using freeWAIS-sf, you can do free text search using a ranking facility, and you can search structured fields (`sf'). There is a program SFproxy, to be had from ls6-ftp.informatik.uni-dortmund.de, too, which can operate as a WAIS proxy in addition to being able to retrieve a list of documents via HTTP and indexing them. There is a program SFgate (from ls6-ftp, as before), which allows you to query the WAIS database via a WWW browser (SFgate operates as a CGI script). \\kai{} -- Life is hard and then you die. From: Kai Grossjohann \ Newsgroups: comp.text.sgml Subject: Re: SGML - HTML newbie looking for input Date: 21 Jun 1995 14:30:26 +0200 Organization: University of Dortmund, Germany Message-ID: \ References: <3s2m86$8iq@kettle.magna.com.au> Reply-To: Kai Grossjohann \ In-reply-to: tactics@magna.com.au's message of 19 Jun 1995 02:09:10 GMT To: tactics@magna.com.au (David Churchland) X-Newsreader: (ding) Gnus v0.87 >>>>> "David" == David Churchland \ writes: David> I have been doing a considerable amount of research into SGML David> and have started doing some work with HTML. I have been David> thinking that HTML might be a more accessable and, initially, David> less painful way of putting some of our documentation on-line David> prior to moving toward a full SGML solution. My problem is David> that I am having to battle the Windows help proponents who David> love the low initial co$t of using the help engine. Please note that you can, in principle, author your documents in SGML, according to some DTD, and then transform them into HTML documents, to be viewed via WWW. This way, you have a `future compatible' underlying document format and are not tied to HTML. Also, it is possible to use semantic tagging with SGML which aids you when you want to index your documents (as, with semantic tagging, it is easy to search for, say, a document's author or the title or the publication date or something). freeWAIS-sf supports structured fields and is thus a good indexer that can be used to index the SGML files. (cf my previous posting) \\kai{} -- Life is hard and then you die. From: boo@primenet.com (Walter Ian Kaye) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: Wed, 21 Jun 1995 05:35:12 -0700 Organization: Getting organized... a little more... a little more... Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> X-Web-Site: http://www.primenet.com/~boo/ X-FTP-Site: ftp://ftp.primenet.com/users/b/boo/ X-Quote: More insomniacs get their news from World News Now than from any other source! X-Newsreader: Value-Added NewsWatcher 2.0b24.3+ In article <3s76f1$5td@googol.bctel.net>, Andrew Haigh \ wrote: > If you are looking for a file format that is so compact that > it blows away PDF for bandwidth conservation, transmission > speed, and graphic retention, check out RTC, > Rich-Text-Compressed, a page format developed by Imagen > Communications Inc. Check out our home page listed below for > examples of the files, and the free-viewer (which also blows > Acrobat out of the water - 170K zipped as opposed to 1.4megs) > > -- > Andrew Haigh Imagen Communications > mailto:ahaigh@imagen.net http://www.imagen.net Uh, I tried www.imagen.net, imagen.net, imagen.com, and www.imagen.com -- no DNS (pronounced "no dice!"). Is it new and not fully propagated yet? # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # Walter Ian Kaye: (602) 942-6390 FoxPro/Excel Programmer; Guitarist # # Correspond to: boo@primenet.com, boodlums@genie.com # # BinHex files: boo@primenet.com WWW: http://www.primenet.com/~boo/ # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # From: Arjan Loeffen \ Newsgroups: comp.text.sgml Subject: DTD to meta-DTD conversion? Date: 21 Jun 1995 12:50:39 GMT Organization: Faculty of Arts, Utrecht University, NL Message-ID: <3s94iv$h9r@ruulch.let.ruu.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1b3 (Windows; I; 16bit) Dear reader, Does anyone have a DTD-to-Meta-DTD converter lying around? I mean, a program that reads a DTD and creates a document instance describing the declarations of the document type definition? I need such a program to generate the tree structure of some large DTD's as an image file, for visual support. I'd be most interested in public domain software. Money is tight. Thanks in advance, Arjan (similar message also posted on TEI-LIST) From: s.rahtz@srahtz.elsevier.co.uk (Sebastian Rahtz) Newsgroups: comp.text.sgml Subject: Re: SGML conversion utilities Date: 21 Jun 1995 13:09:31 GMT Organization: Elsevier Science Ltd, Oxford, UK Message-ID: \ References: \ In-reply-to: jjr@panini.att.com's message of Wed, 14 Jun 1995 22:47:15 GMT I'm looking for UNIX-based utilities available in the public domain to convert various formats to and from SGML (e.g., SGML-to-LaTeX and LaTeX-to-SGML). If SGML and LaTeX were markup languages, this would be more reasonable; since SGML is a language for describing markup, and LaTeX is a programming language, its almost impossible to describe, let alone implement, the tools you want. At some point, you have to do some work mapping some LaTeX markup to some DTD; given that, any of the SGML converters can do the job. I find *arbitrary* LaTeX to SGML an interesting challenge; you can look at latex2html (which is after all latex to SGML) to see just how problematic it can get. If you come to the 1995 TeX Users Group meeting, i'm giving a paper on a working LaTeX to SGML - this is called a \ :-} sebastian Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla From: jbeverly@adobe.com (Jamie D. Beverly) Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Message-ID: \ Sender: usenet@adobe.com (USENET NEWS) Organization: Adobe Systems Incorporated References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> Date: Wed, 21 Jun 1995 13:53:52 GMT In article <3s76f1$5td@googol.bctel.net>, Andrew Haigh \ wrote: > If you are looking for a file format that is so compact that >it blows away PDF for bandwidth conservation, transmission >speed, and graphic retention, check out RTC, >Rich-Text-Compressed, a page format developed by Imagen >Communications Inc. Check out our home page listed below for >examples of the files, and the free-viewer (which also blows >Acrobat out of the water - 170K zipped as opposed to 1.4megs) > I tried to get the free viewer, but found us Mac users left in the cold. Seems UNIX and DOS users are out of luck too. I guess "blown away" can be taken literally... >-- >Andrew Haigh Imagen Communications >mailto:ahaigh@imagen.net http://www.imagen.net "We're here to protect democracy, not practice it" Gene Hackman in "Crimson Tide" From: Kai Grossjohann \ Newsgroups: comp.text.sgml Subject: Novice question -- Reusing a DTD? Date: 21 Jun 1995 16:11:27 +0200 Organization: University of Dortmund, Germany Message-ID: \ Reply-To: Kai Grossjohann \ X-Newsreader: (ding) Gnus v0.87 Hi there, being a rather novice user of SGML, I've got a question which doesn't quite measure up to the high level prevalent in this group, I'm afraid. Please bear with me... I would like to define a DTD (call it `kai') where documents look like this: \ \ ..arbitrary HTML text can go here.. ..like, say: .. ..\

cool heading\

.. \ \ ..arbitrary HTML text can go here.. \ \ How do I best make use of the existing html.dtd? Thanks a lot for your help in advance, yours, \\kai{} -- Life is hard and then you die. Newsgroups: comp.text.sgml From: robh@cwi.nl (Rob Hoksbergen) Subject: Examples of use of ISO12083:1993 math dtd Message-ID: \ Keywords: mathematics Sender: news@cwi.nl (The Daily Dross) Organization: CWI, Amsterdam Date: Wed, 21 Jun 1995 14:19:58 GMT Hi, is there anyone who has used the above dtd (in combination with, say, article) and is willing to send me some examples of code? I am having trouble dealing with: -equation numbers (tags): how to id and rid them -accents in math (x dot, x arrow, x circle) -matrices and pile (case) constructs: Also, what is the "labelled arrows" section in this DTD for? A few clear examples will probably help me (mathematician with lots of TeX/EQN knowledge) on the way. Thanks in advance. From: Martin Bryan \ Newsgroups: comp.text.sgml Subject: Re: Comments on ideas... Date: Wed, 21 Jun 95 15:30:09 GMT Organization: The SGML Centre Message-ID: <803748609snz@sgml-cen.demon.co.uk> References: <3s41fl$r0k@maroon.tc.umn.edu> Reply-To: mtbryan@sgml-cen.demon.co.uk X-NNTP-Posting-Host: sgml-cen.demon.co.uk X-Newsreader: Demon Internet Simple News v1.29 Comments on contribution from: milor001@maroon.tc.umn.edu (R A Milowski) in: comp.text.sgml Subject: Comments on ideas... Date: 19 Jun 1995 09:27:01 -0500 Organization: University of Minnesota Message-ID: <3s41fl$r0k@maroon.tc.umn.edu> Mr Milowski comments: >1. Entity management, as a whole, has not been treated as of equal importance > as parsing--case in point, there are *no* entity management products. > Although there are a few SGML database products on the horizon, if not now, > they may or may not address the issue of consistent entity management. The WG8 team published as standard for SGML entity management (ISO 9069) way back in 1988. Users, however, did not demand that suppliers adopt this optional feature of the SGML standard (as with other optional features whose benefits are only now becoming obvious to everyone!) The Open-SGML consortium has tried to tackle this problem by introducing the concept of a catalog. In a forthcoming revision to HyTime the whole concept will be taken even further with the idea of formal system identifiers and notations to identify entity managers being introduced to provide a comprehenzive, portable, methodology for interchanging the information needed to manage entity managers. >2. SGML processing models have traditionally been linear in the sense that > an application asks a parser to parse an SGML document and the parser > asks the entity manager (if any) for the document. This is a > unfortunate standard for SGML processing since a true SGML system should be > a system where Entity Management, Parsing/Validating, and Application > work together in a "triangle of interdependency". That is, the functions > and services that they provide are very much interrelated and interdependent. The concept of managing data fragments as sub-trees has been around in most advanced SGML systems for some time. The DSSSL standard will give you ways of transforming subtrees so that they can be reused more easily. >3. The parser is the only way in which raw SGML information can be correctly > "understood" by an SGML application. > >4. There is not *standard* interface between an SGML parser and the application. The work currently going on in looking at how SGML can be extended will help in this area. DSSSL has already introduced the idea of an originating object structure environment (OOSE) that can be used as a standardized representation of the components of an SGML document. HyTime has been extended to operate with this OOSE structure, which will, I suspect, become the standard way of interfacing between parsers and applications. >5. All SGML parsers supply a subset of the information provided by raw SGML > as denoted in the ISO 8879 standard. The subset of information to date has often been that defined in the Element Structure Information Set, which has always been recognized as being too limited. The OOSE structure will extend this information set, though there will still be a few gaps (related to rarely used functions such as RANK and SHORTREF) >6. The only way to provide a "standard interface" that serves everyone's needs > to the an SGML parser is to serve the productions of the SGML standard. > This, of course, would be a low-level interface. > >7. A standard object-framework, along the lines of Fresco, would be of great > benefit to *application* writers--not tool vendors. The WG8 committee has been trying to define a standard object framework for SGML for the last three years. OOSE is the result. >8. A standard set of services (object-framework?) for entity management would > be of great benefit to application writers and tool vendors. Although > SGML open has a definition for specification of this, there needs to be > some kind of API/object-framework specification. It would be nice if we could get a standardized API approved as part of the SGML review, but I suspect that this would upset too many of those with existing tools to get enough votes to be accepted as a standard. >9. (The most important one) There is a great need in the SGML industry for > programming-level standards so that "components" can be developed that > allow for component-level inter-operation. That is, so that a variety of > tools and applications can be integrate with ease to create open SGML > applications. >10. Such programming-level standardization is necessary for the success of > SGML since the practicality of the matter is that people won't use what > is difficult or costly to integrate and develop for. Why not join ANSI X3V1 and start contributing to our regular discussions on this very topic. You would be most welcome. -- Martin Bryan @ The SGML Centre, Churchdown, Glos. GL3 2PU, UK (+44 1452 714029) From: Andrew Haigh \ Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 21 Jun 1995 16:09:04 GMT Organization: Imagen Communications Inc. Message-ID: <3s9g70$ei7@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 32bit) >Uh, I tried www.imagen.net, imagen.net, imagen.com, and www.imagen.com -- >no DNS (pronounced "no dice!"). Is it new and not fully propagated yet? We have been up and online for about two months now, so there is a chance that our eddress hasn't fully propagated yet, Yes. In fact we have had previous reports from primenet.com users that they have ben unable to access our web site. Try the numeric eddress 204.239.245.1 instead of the name and you should get through. -- Andrew Haigh Imagen Communications mailto:ahaigh@imagen.net http://www.imagen.net From: milor001@maroon.tc.umn.edu (R A Milowski) Newsgroups: comp.text.sgml Subject: Re: Comments on ideas... Date: 21 Jun 1995 12:03:55 -0500 Organization: University of Minnesota Message-ID: <3s9jdr$537@maroon.tc.umn.edu> References: <3s41fl$r0k@maroon.tc.umn.edu> <803748609snz@sgml-cen.demon.co.uk> I certainly hope I wasn't sounding like I was complaining. In fact, these are problems for which I think I have some or part of the solutions to. In article <803748609snz@sgml-cen.demon.co.uk>, Martin Bryan \ wrote: > >Comments on contribution from: milor001@maroon.tc.umn.edu (R A Milowski) >in: comp.text.sgml >Subject: Comments on ideas... >Date: 19 Jun 1995 09:27:01 -0500 >Organization: University of Minnesota >Message-ID: <3s41fl$r0k@maroon.tc.umn.edu> > >Mr Milowski comments: > >>1. Entity management, as a whole, has not been treated as of equal importance >> as parsing--case in point, there are *no* entity management products. >> Although there are a few SGML database products on the horizon, if not now, >> they may or may not address the issue of consistent entity management. > >The WG8 team published as standard for SGML entity management (ISO 9069) >way back in 1988. Users, however, did not demand that suppliers adopt >this optional feature of the SGML standard (as with other optional features >whose benefits are only now becoming obvious to everyone!) >The Open-SGML consortium has tried to tackle this problem by introducing >the concept of a catalog. In a forthcoming revision to HyTime the whole >concept will be taken even further with the idea of formal system >identifiers and notations to identify entity managers being introduced >to provide a comprehenzive, portable, methodology for interchanging >the information needed to manage entity managers. Pardon my ignorance: 9069? Is that SDIF? If so, that's not the kind of entity management I was thinking of. I like the catalog format from SGML-Open. The problem is that I *can* extend the catalog to provide, say, SQL query access, but there is no way of attaching the semantics in a standard way. If you look at systems like Java/HotJava that can "request" the semantics from a web server for things it does know about, you see how a standard scheme and API/object-interface for entity management can be really useful. It is not that such schemes will provide *complete* openness (i.e. if I don't have the database or data in it for the SQL query I am at a loss). It will provide a means for setting up an application to be able to integrate into such an environment with coding or much effort. > >>2. SGML processing models have traditionally been linear in the sense that >> an application asks a parser to parse an SGML document and the parser >> asks the entity manager (if any) for the document. This is a >> unfortunate standard for SGML processing since a true SGML system should be >> a system where Entity Management, Parsing/Validating, and Application >> work together in a "triangle of interdependency". That is, the functions >> and services that they provide are very much interrelated and interdependent. > >The concept of managing data fragments as sub-trees has been around in >most advanced SGML systems for some time. The DSSSL standard will >give you ways of transforming subtrees so that they can be reused more >easily. I guess I would have to disagree. I see most systems as linear--even some of the most advanced ones--which is unfortunate. I like to think of systems in terms of "documents are databases, how well do you implement this?" and see if I can perform such operations. Now, when true DSSSL systems come about, my attitude will surely change. > >>3. The parser is the only way in which raw SGML information can be correctly >> "understood" by an SGML application. >> >>4. There is not *standard* interface between an SGML parser and the application. > >The work currently going on in looking at how SGML can be extended will >help in this area. DSSSL has already introduced the idea of an originating >object structure environment (OOSE) that can be used as a standardized >representation of the components of an SGML document. HyTime has been >extended to operate with this OOSE structure, which will, I suspect, >become the standard way of interfacing between parsers and applications. Now, this is new!! OOSE? What exactly is this and where can I get information on it. [stuff deleted] >>6. The only way to provide a "standard interface" that serves everyone's needs >> to the an SGML parser is to serve the productions of the SGML standard. >> This, of course, would be a low-level interface. >> >>7. A standard object-framework, along the lines of Fresco, would be of great >> benefit to *application* writers--not tool vendors. > >The WG8 committee has been trying to define a standard object framework >for SGML for the last three years. OOSE is the result. > >>8. A standard set of services (object-framework?) for entity management would >> be of great benefit to application writers and tool vendors. Although >> SGML open has a definition for specification of this, there needs to be >> some kind of API/object-framework specification. > >It would be nice if we could get a standardized API approved as part >of the SGML review, but I suspect that this would upset too many of those >with existing tools to get enough votes to be accepted as a standard. I heard the same argument at SGML '94 from someone at Microstar who seemed to agree with me that such program-level integration was good but would never be allowed to pass by the vendors. Are the SGML vendors really that self interested? I'm tempted to say no--if it was done in a fair way. Shall we allow the SGML community to be stricken by this plague? Why not do what is being done by X/Open with Fresco or Sun with Java? If you develop it and give it away, they will come.... ;) Seriously, what is needed here is free implementations without any legal strings attached. >>9. (The most important one) There is a great need in the SGML industry for >> programming-level standards so that "components" can be developed that >> allow for component-level inter-operation. That is, so that a variety of >> tools and applications can be integrate with ease to create open SGML >> applications. > >>10. Such programming-level standardization is necessary for the success of >> SGML since the practicality of the matter is that people won't use what >> is difficult or costly to integrate and develop for. > >Why not join ANSI X3V1 and start contributing to our regular discussions >on this very topic. You would be most welcome. Well, I'm on the mailing list.... now, if I only had the time. ;) Actually, I'm really trying to make it to a meeting in the next 6 months--but, time is short for me. Thanks for the comments--all are appreciated! -- ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu Copernican Solutions Incorporated (612) 825 - 4132 From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: DTD to meta-DTD conversion? Date: 21 Jun 1995 10:54:21 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3s9mcd$oe@crl2.crl.com> References: <3s94iv$h9r@ruulch.let.ruu.nl> Arjan Loeffen \ wrote: >Does anyone have a DTD-to-Meta-DTD converter lying around? I mean, a >program that reads a DTD and creates a document instance describing the >declarations of the document type definition? >I need such a program to generate the tree structure of some large DTD's >as an image file, for visual support. Earl Hood's DTD2HTML package may do something close to what you want, and if not, it's well-written enough that hacking it to suit your needs should be fairly easy. It's all in Perl. See \ for details... --jenglish@crl.com From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: Novice question -- Reusing a DTD? Date: 21 Jun 1995 11:13:57 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3s9nh5$11r@crl2.crl.com> References: \ Kai Grossjohann \ wrote: > >I would like to define a DTD (call it `kai') where documents look like >this: > \ > \ > ..arbitrary HTML text can go here.. > ..\

cool heading\

.. > \ > [...] >How do I best make use of the existing html.dtd? That's a neat idea... you could create a "wrapper" DTD around HTML to keep track of a larger collection of nodes, something that HTML doesn't do very well on its own. Anyway, the following technique works: -- Cut Here -- \ \ \ \ %HTML-DTD; \ \ \ \ \ \ \ ]> \ \ \

Part A can contain any block-level elements. \ \ \ \\Subdocument part B\\ \ \

Part B must contain a complete HTML document. \ \ \ \ -- Cut Here -- --Joe English jenglish@crl.com From: thiel@acsu.buffalo.edu (Kathleen M Thiel) Newsgroups: comp.text.sgml Subject: Characters in sgml tags Date: 21 Jun 1995 19:24:19 GMT Organization: UB Message-ID: <3s9rl3$qgo@azure.acsu.buffalo.edu> I wanted to know which are the characters that are allowed in an sgml tag, BUT have to be in "" or '' . I know that a string like ABCD does not have to be in quotes but a string like "AB#CD" must be in quotes. What are the other characters besides # and /? Thanks Kathy From: stevmorgan@aol.com (StevMorgan) Newsgroups: comp.text.sgml Subject: Re: SGML based databases Date: 21 Jun 1995 16:24:20 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3s9v5k$pbe@newsbf02.news.aol.com> References: <1995Jun14.114905.55821@ucl.ac.uk> Reply-To: stevmorgan@aol.com (StevMorgan) The REAL answer to your question is in a new product called LivePage by Inforium. They have an excellent Web page, and have great literature to send. Good luck. Steve Morgan Can best be E-Mailed at StevenL@erinet.com, rather than AOL, I check net mail daily, AOL mail weekly. From: ahaigh@imagen.net (Andrew Haigh) Newsgroups: comp.os.ms-windows.apps.misc,comp.os.ms-windows.apps.utilities,comp.os.ms-windows.apps.word-proc,comp.os.ms-windows.programmer.multimedia,comp.text,comp.text.desktop,comp.text.sgml,misc.creativity,rec.arts.books,3b.misc Subject: Compressed RTF (RTC) to challenge PDF on the WWW Date: 21 Jun 1995 20:50:45 GMT Organization: Imagen Communications Inc. Message-ID: <3sa0n5$mre@googol.bctel.net> Mime-Version: 1.0 X-Newsreader: WinVN 0.99.3 (Vancouver, Canada) Imagen Communications Inc. has announced enhancements to its RTC (Rich Text Compressed) file format which will offer substantial advantages over its chief rival, Adobe Acrobat's PDF (Portable Document Format). Recent WEB functionality built into the RTC viewer, which can be used as a "helper application" with any browser, allows the reader to complete hyperlinks both internally within documents and externally to view references anywhere on the WEB without the assistance of the browser. Imagen Product Manager Johnathan Steere comments that " RTC documents now request files directly through Winsockets, so there is no need for complicated communications with individual browsers. You cannot at present insert HTML hyperlinks within PDF files." RTC Based On The RTF Standard The new RTC format is based on compressed RTF (Rich Text Format) files, which are the industry standard for document exchange. Competing formats such as Novell's Envoy, Farallon's Replica, and No Hands' Common Ground are completely proprietary, and do not offer conversion to and from the RTF standard. Imagen CEO Dwight Jones observes that "RTC files are built for speed, we do not concern ourselves with Postscript and fancy font emulations, which bloat file size. Although RTF can support any font, we feature the system fonts everyone already has." Imagen also uses the sound and MPEG video decompression to be supported in Microsoft operating systems, so that the RTC viewer remains small and nimble. "When the screen resolution is 72 DPI", says Jones, "why would we haul around Postscript? Our mandate is to stay inside the 28.8 K envelope supported by ordinary phone lines and modems. That's where the mass market is, and shall remain, until we are all connected to fiber." RTC Offers Advantages Over Basic HTML The main advantage of compressed RTF is that such documents can offer a true magazine interface, with multiple columns and images, large fonts- what you would expect from a slick magazine. Pages are also much easier to make up or convert(in any major word processor and saved as RTF) and special HTML editors are not required. Support for HTML, Sound and Video Hyperlinks are supported with standard HTML syntax, and these are executed directly by the RTC "helper application". Compressed sound compression/decompression is currently included as will be MPEG video on Microsoft platforms. Compression Included Free of Charge Imagen is aggressively distributing compression free of charge, along with the free reader. The registered version ($190) includes support and Help files, and allows unlimited publishing from a single server. A Return to Standard Word Processors "The whole problem has been that standard word processors have not accepted compressed images" said Jones, " and now that RTC enables that we anticipate a migration back to major word processors, which all support RTF, as the standard authoring tools." Because RTC optionally converts back to RTF after transmission, documents remain searchable, editable and readable. Pages are not imaged (rasterized) as with some competitors, and so remain "live" documents. This is of particular value for indexing on WAIS servers. WEB, Wireless and Fax May Benefit Tiny page sizes (typically 2Kb-20Kb) mean fast WEB transfers, wireless documents that can be actual pages rather than strings of text, and editable faxes that complete in a few seconds. "Online publishing has been needlessly complicated, " says Jones, "and we're overdue to return to simple documents supporting established standards, utilizing display code and fonts built into existing software." Imagen Communications Inc., is a private company founded in 1990 Imagen has pioneered online imaging and specializes in document compression for publishers. Contact: Press: djones@imagen.net Dwight Jones (604) 537-1950 Product Information: ahaigh@imagen.net Andrew Haigh (604) 687-7511 URL: http://www.imagen.net From: Andrew Haigh \ Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 21 Jun 1995 22:47:22 GMT Organization: Imagen Communications Inc. Message-ID: <3sa7hq$qqc@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 32bit) jbeverly@adobe.com (Jamie D. Beverly) wrote: >I tried to get the free viewer, but found us Mac users left in the cold. >Seems UNIX and DOS users are out of luck too. I guess "blown away" can be >taken literally... There are plans for a Mac and a Unix version to be implemented down the line. As for Dos, once Windows `95 hits the streets, Dos is as dead as my old Sinclair ZX-80. -- Andrew Haigh Imagen Communications mailto:ahaigh@imagen.net http://www.imagen.net From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Another reason I like SGML Date: 22 Jun 1995 00:43:40 -0700 Organization: Call Really Late Dialup Internet Access Message-ID: <3sb6vc$mtv@crl7.crl.com> I've been working on a paper that contains dozens of short Ada-like pseudocode fragments. The document is in SGML with a homebrew DTD, and I'm translating to LaTeX to produce hardcopy. I didn't much like the look of the code fragments -- they were all in a monospaced typewriter font -- but I certainly wasn't about to boldface every keyword and italicize every comment by *hand* when the computer can figure out how to do that on its own. So I downloaded 'lgrind' (a code beautifier for LaTeX) from CTAN, added "\" to the DTD, changed the relevant \

tags to \, and hacked my translation script to pipe the content of these elements through lgrind. The whole process took about half an hour. Presto! Nice looking code fragments, and I didn't have to mangle my source document to get it. Try *that* with FrameMaker! --Joe English jenglish@crl.com From: rcstage9@urc.tue.nl (Maarten Deen) Newsgroups: comp.text.sgml Subject: WordPerfect Internet Assistant Date: Thu, 22 Jun 95 11:42:11 GMT Organization: TU Eindhoven Message-ID: <3sbadb$rm5@tuegate.tue.nl> X-Newsreader: News Xpress Version 1.0 Beta #3 Hi, Has anyone have experience with the WP IA? I can't seem to open a HTML document and have it recognised as an HTML document. WP wants to read it as and ASCII file and does not give an option to read it as HTML. Is there a way to do this right? (Read as HTML I mean) Maarten -- Maarten Deen (mccidd@ns.via.nl) http://ns.via.nl/users/mccidd From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: Characters in sgml tags Date: 22 Jun 1995 12:07:21 GMT Organization: ACM Network Services Message-ID: <3sbmdp$6s9@hopper.acm.org> References: <3s9rl3$qgo@azure.acsu.buffalo.edu> Reply-To: davep@ACM.ORG In article <3s9rl3$qgo@azure.acsu.buffalo.edu>, thiel@acsu.buffalo.edu (Kathleen M Thiel) writes: >I wanted to know which are the characters that are allowed in an >sgml tag, BUT have to be in "" or '' . I know that a string >like ABCD does not have to be in quotes but a string like >"AB#CD" must be in quotes. What are the other characters >besides # and /? The simple answer is: practically any character. Tags come in two flavors, start- and end-. Presumably you're talking about start-tags, wherein you may have attribute specifications. What characters are allowed depends on where you are in the start-tag; the only place where you quote anything is the attribute value, a part of the attribute specification. In that context, you are permitted to have "replaceable character data" ([34] in ISO 8879). That means any legal SGML character (as defined in the SGML declaration) except the "lit" or "lita" (usually double or single quote) that started the string. Even non-SGML characters and that "lit" or "lita" character can be included indirectly by using an entity or character reference. The net result is that, if you're using delimiter strings from the reference concrete syntax, which almost everyone does almost all of the time, you can include directly as data any SGML character except the ampersand (which usually, although not under every possible circumstance, will be recognized as markup-- the start of an entity or character reference--rather than data) and either the single or double quote, whichever you used to start the string (which will always be recognized as markup--the end of the string). Inside the replacement text of a referenced text entity, even the single and double quotes will be data; inside the replacement text of a referenced data entity, every character will be data. ISO 8879 is IMHO not clear what should happen if external data entities are referenced in an attribute value since they always come with a notation attached. Hope this helps Dave Peterson SGMLWorks! davep@acm.org Newsgroups: comp.text.sgml Subject: Re: SGML based databases Message-ID: <1995Jun22.131710.1318@ittpub> From: miked@ittpub.nl (Mike Davis) Date: 22 Jun 95 13:17:09 WET References: <1995Jun14.114905.55821@ucl.ac.uk> ubacr63@pluto.dcs.bbk.ac.uk (Dan Smith) wrote: >I am currently researching textual databases and am wondering if there >are any extant SGML based (experimental / non-experimental) database >systems in existence. >As there seems to be a lot of data being produced in SGML format it >seems sensible to explore this approach. >If anyone has any information I would be grateful so that I know what >to and what not to follow up! >TIA >Dan I bumped into one OCLC were using. They're over in the US. I can't give you any more help. From: jeffrey@hks.com (Glenda Jeffrey) Newsgroups: comp.text.sgml Subject: Re: Help with content model? Date: 22 Jun 1995 13:41:37 GMT Organization: Hibbitt, Karlsson & Sorensen, Inc. Message-ID: <3sbruh$4tn@argo.hks.com> References: <3s4j3u$uo@argo.hks.com> <9506202221.AA0720@notes.microstar.com> X-Newsreader: TIN [version 1.2 PL2] So far, I've gotten lots of nice suggestions for alternatives to my ambiguous content model, which was: : > \ Dave Peterson and Erik Naggum suggested: : > \ So far, it appears to me that this solution is the most easily understood for my authors. However, Hugh Chatfield and Jeffrey McArthur suggested some interesting alternatives (and Hugh provided a nice explanation of the original ambiguity). Hugh, in particular, points out that the "right" solution depends on the true meaning of the elements A, B, and C. So, before I take Dave's and Erik's simple solution, let me tell you what these elements do, and see whether you agree that the simple solution is the best one. All three elements are wrapper elements for lists of "parameters" (in my company's jargon). The lists are actually DocBook variablelists, with a slight twist. Element A is for "required" parameters (required in our software, not required in the DTD) and element C is for "optional" parameters. Element B is for parameters which are either required or optional, and which have other peculiar conditions attached to them. Elements A and C will have automatically generated headings, since their purpose is known from the get-go. Element B allows the author to input a "custom" heading. So, the content models go like this: \ \ \ (Obviously, I could have used attributes to distinguish A and C, rather than having separate elements -- except that I need to enforce the order (A before C), and I can't do that with attributes.) An example of a composed instance corresponding to this DTD would be: ================================ Required Parameters: \ Parameter Alpha Definition Parameber Beta Definition Required parameters if the zeta capability is used: \ Parameter Gamme Definition Parameter Delta Definition Optional parameters: \ Parameter Epsilon Definition Parameter Chi Definition Optional, mutually exclusive parameters, but only when the moon is full: \ Parameter Eta Definition Parameter Theta Definition Parameter Iota Definition =============================== Obviously, I could make my authors input the "Required Parameters:" and "Optional Parameters:" headings, but then I can't enforce the order. (and I miss the opportunity to toy with automatic text generation ;-) There is some possibility that I could call the first B something like \ (where A is \, C is \, and the second B is \), but I wouldn't want to bank on the first B always being some form of required parameter. So, now that you know the full story, any further comments? Thanks so much for your help so far -- you folks are lifesavers! -- Glenda Jeffrey Email: jeffrey@hks.com Hibbitt, Karlsson & Sorensen, Inc Phone: 401-727-4200 1080 Main St. Fax: 401-727-4208 Pawtucket, RI 02860 From: milor001@maroon.tc.umn.edu (R A Milowski) Newsgroups: comp.text.sgml Subject: Re: Compressed RTF (RTC) to challenge PDF on the WWW Date: 22 Jun 1995 09:03:00 -0500 Organization: University of Minnesota Message-ID: <3sbt6k$ss8@maroon.tc.umn.edu> References: <3sa0n5$mre@googol.bctel.net> In article <3sa0n5$mre@googol.bctel.net>, Andrew Haigh \ wrote: >(Vancouver, Canada) Imagen Communications Inc. has announced >enhancements to its RTC (Rich Text Compressed) file format which >will offer substantial advantages over its chief rival, Adobe >Acrobat's PDF (Portable Document Format). What does this have to do with SGML? Cross-posts are just plain evil! -- ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu Copernican Solutions Incorporated (612) 825 - 4132 From: Kai Grossjohann \ Newsgroups: comp.text.sgml Subject: Re: Novice question -- Reusing a DTD? Date: 22 Jun 1995 16:51:06 +0200 Organization: Universitaet Dortmund, CS Dept, Chair 6 Message-ID: \ References: \ <3s9nh5$11r@crl2.crl.com> Reply-To: Kai Grossjohann \ In-reply-to: jenglish@crl.com's message of 21 Jun 1995 11:13:57 -0700 To: jenglish@crl.com (Joe English) X-Newsreader: (ding) Gnus v0.87 >>>>> Kai Grossjohann \ wrote: Kai> How do I best make use of the existing html.dtd? >>>>> "Joe" == Joe English \ writes: Joe> That's a neat idea... you could create a "wrapper" DTD around Joe> HTML to keep track of a larger collection of nodes, something Joe> that HTML doesn't do very well on its own. Actually, I've got a database of SGML files, intended to be displayed in HTML. Rather than making up my own syntax for things like imagemaps, links, and suchlike, I thought I'd just reuse the html stuff. The database is to contain documents of different kinds, semantically tagged for the convenience of later retrieval. (A lecture would be tagged with attributes like title, lecturer, time and place, abstract and others.) I though it would be nice if the users could embed any HTML markup they cared in, say, the abstract field for nice looks (or in the lecturer field, for that matter, to embed a link to their homepage or something). Would you think this is an unreasonable approach? I'm aware of the fact that this approach means transition to a different standard is more complicated (when HTML dies). Oh, well... Joe> Anyway, the following technique works: [...] Gee, complicated syntax. I sure hope I will learn this some time! At least, now I know how to `#include' a file (in C parlance). Thanks a bundle for your quick help! \\kai{} -- Life is hard and then you die. From: Kai Grossjohann \ Newsgroups: comp.text.sgml Subject: NAMELEN -- SGML Declaration -- How to? Date: 22 Jun 1995 17:39:12 +0200 Organization: Universitaet Dortmund, CS Dept, Chair 6 Message-ID: \ Reply-To: Kai Grossjohann \ X-Newsreader: (ding) Gnus v0.87 Hi all, a very short question: How do I allow names up to 72 characters? Some more info: I know that I need an SGML declaration for this which might look like this: \ ...or it might not. The problem is: Where do I put this string so that the parser interprets it correctly? I now have two files: foo.sgml and foo.dtd where foo.sgml starts with the following line: \ I tried putting the SGML declaration above and below this line, but it didn't work either way. (Above: error msg `sgml declaration not allowed here', below: didn't have any effect.) Is this question answered in Goldfarb's book? I now have only ``SGML -- An Author's Guide'' by Martin Bryan which doesn't seem to explain this. tia, \\kai{} -- Life is hard and then you die. From: eoaai@csv.warwick.ac.uk (Chris Tilbury) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Followup-To: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Date: 22 Jun 1995 17:31:55 +0100 Organization: University of Warwick, Coventry, UK Message-ID: <3sc5tr$je0@crocus.csv.warwick.ac.uk> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Newsreader: TIN [version 1.2 PL2] Andrew Haigh (ahaigh@imagen.net) wrote: : jbeverly@adobe.com (Jamie D. Beverly) wrote: : >I tried to get the free viewer, but found us Mac users left in the cold. : >Seems UNIX and DOS users are out of luck too. I guess "blown away" can be : >taken literally... : There are plans for a Mac and a Unix version to be : implemented down the line. As for Dos, once Windows `95 hits : the streets, Dos is as dead as my old Sinclair ZX-80. The only people who are losing out by making such a ludicrous assumption are yourselves - there is a massive installed base of DOS users out there, who may have no need to upgrade. Somehow, I can't see everyone immediately rushing out and getting Windows '90-whatever the instant it's released. You're not wearing Microsoft-tinted glasses, by any chance, are you? Cheers, Chris -- Chris Tilbury, Estates, University of Warwick, England, CV4 7AL MIME mail welcomed - text/html makes much more interesting reading mailto:Chris.Tilbury@estate.warwick.ac.uk From: ahaigh@imagen.net (Andrew Haigh) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 22 Jun 1995 16:43:31 GMT Organization: Imagen Communications Inc. Message-ID: <3sc6jj$257@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> <3sc5tr$je0@crocus.csv.warwick.ac.uk> Mime-Version: 1.0 X-Newsreader: WinVN 0.99.3 In article <3sc5tr$je0@crocus.csv.warwick.ac.uk>, eoaai@csv.warwick.ac.uk says... >The only people who are losing out by making such a ludicrous >assumption are yourselves - there is a massive installed >base of DOS users out there, who may have no need to upgrade. > >Somehow, I can't see everyone immediately rushing out and getting >Windows '90-whatever the instant it's released. > >You're not wearing Microsoft-tinted glasses, by any chance, are >you? > >Cheers, Nope, but I was a dyed-in-the-wool Dos Fanatic until I started using `95. I would never go back now. IT has become the OS that 3.1 was supposed to be, instead of the toy that it was. Frankly I can't stand Billy-boy and Megasoft, but they have a good product with this OS Andrew Haigh Imagen Communications mailto:ahaigh@imagen.net http://www.imagen.net From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: Re: NAMELEN -- SGML Declaration -- How to? Date: 22 Jun 1995 17:10:30 GMT Organization: SEMA Group Belgium Message-ID: <3sc866$na4@news.Belgium.EU.net> References: \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) Kai Grossjohann \ wrote: >Hi all, > >a very short question: > > How do I allow names up to 72 characters? > >Some more info: > >I know that I need an SGML declaration for this which might look like >this: > > \ > >...or it might not. The problem is: Where do I put this string so >that the parser interprets it correctly? > .. You need to specify the value for NAMELEN in ***the*** (not just "an") SGML declaration for your application. The tools you are using (which ones, BTW ?) assume an SGML declaration. If you don't specify one, the defaults (SGMLREF) are assumed, which limit NAMELEN to 8. Read something on this in the Annexes to ISO 8879, reproduced and explained in Charles Goldfarb's SGML Handbook, or in the other classics from Martin Bryan, Eric Van Herwijnen, and others. A lot of public DTDs come with their SGML declaration The declaration you mention will choke every parser. Here's an example of the SGML declaration changing the default value for NAMELEN. \ Hereafter the statements of your DTD, as \ \ \ .. ]> should follow. You may want to change other things, such as the BASESET in order to use 8-bit characters such as ISO 8859-1 (you live in Germany and not in the US, after all). It depends on the tools you use how you make your environment "aware" of the non-reference SGML declaration. You can place it at the start of your DTD file, although not all tools will cope with that. With sgmls, you can specify declaration, dtd and document instance as parameters and sgmls will concatenate these files. Mail me if you want some real examples of non-reference SGML declarations. Best regards and good luck, ----------------------------------------------------------- Jacques Deseyne \ SEMA Group Belgium Stallestraat 96 (we have moved !) B-1180 Brussels Belgium Tel. + {32} 2 333 53 24 Fax. + {32} 2 333 55 22 ----------------------------------------------------------- Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 22 Jun 1995 14:07:48 -0400 From: "Claude L. Bullard" \ Message-ID: <9506221807.AA24304@source.asset.com> References: <3s41fl$r0k@maroon.tc.umn.edu> <803748609snz@sgml-cen.demon.co.uk> <3s9jdr$537@maroon.tc.umn.edu> Subject: Re: Comments on ideas... This is a REAL long one. Caveat emptor. [Alex Milowski] | Are the SGML vendors really that self interested? I'm tempted to say | no-- if it was done in a fair way. | | Shall we allow the SGML community to be stricken by this plague? Why | not do what is being done by X/Open with Fresco or Sun with Java? If | you develop it and give it away, they will come.... ;) | | Seriously, what is needed here is free implementations without any | legal strings attached. OOSE is part of the HyTime corrigendum currently being voted on in ISO. As such, it's contents are in draft form and have not been released to the general public. Most on the mailing list for it accede to Dr. Goldfarb's request not to disseminate it publicly. As Charles reads this board, he will be aware of your questions. Adoption of the corrigendum (which has some other things you will like, particularly in the objectification of SGML using architectural forms) should occur by the July meeting in Dublin. That meeting could result in the alignment of the SGML standards family (e.g., SGML, DSSSL, HyTime). Keep your fingers crossed and say a prayer for the WG8 members. They are trying very very hard and succeeding for the most part. Wish I could be there. The virtue of selfishness: Vendors are exactly that self-involved. The SGML community does consist of some pretty big egos (i got mine) and self-interested companies. In that sense, it is little different from any industry. Timing and self-publicizing are the essence of market domination. From time to time, my posts are written to give the good-hearted a hint to all of this without slamming individuals. It isn't polite to crucify people or companies in public, it can be slander, and well, there is no class in that. We all live in that glass house. Understand, the industry reps are enlightened capitalists. They pay to send people to WG*, the ISG, ANSI, etc., so in effect they are trying to do their bit for the community. They pay for SGML Open. Most are really intelligent folks doing their best. But they have to make money. Some will soon offer their designs and some of their code, no doubt as freeware. The problem is bigger than fairness. The solutions may not interoperate or provide portability across platforms. Try to build a protocol for querying a server that doesn't use anything but ANSI calls for memory management and is still efficient both at the user interface client and at the server database. Then you get a Charles Goldfarb or a James Clark who creates code and gives us all something really wonderful to support the users and the industry. Why? Because they can, they will, and they do. But they support SGML. The merely obvious: Not only SGML, and HTML, but every notation standard, their respective committees and industry reps are out to dominate the Internet and get the profits. How long has the PDF vs HTML thread been going? Look at who is joining it with yet_another_purported "industry standard" notation and who actually controls the definition of the notation itself. Even the STEP vendors are joining the party. Yes, we can all do the *helper polka* together can't we? The WWW really caught the industry by surprise. No one thought things would move that fast and once in motion, would dominate so quickly. That's what happens when the solution let out for free provides the 10% of the functionality that everyone thinks they want. That's just market psychology. However, now everyone must choose to fight the Web or join it. Those who join, get quick returns but in a quickly saturating industry. Those who fight it are shouted down or vilified by those who want to make money or continue to control the direction of evolution, and so assure their increasing market share or ego boost. My Hope: The Internet supports many protocols, both network and application, and there is some freedom in that. I cheer for HyTime because despite its flaws, it is a workable design, is still very open to competition, will stay stable and public, but mostly because in the notation/interface wars, it could be the piece that defines the standard middleware protocol for hypermedia. HyTime middleware is the best hope for having a unified distributed hypermedia network with competition where things are best competed, and agreement where we can do the most good for the users and the industry. Layer the beast. So, but so what: ISO standards keep the forces in balance because of the policies that govern their creation, and because many of the editors for all their ego, have an eye toward the future and the well-being of the community. Those same policies and politics slow things down tremendously. Enter two new forces: consortia, such as SGML Open or W3O, and the creativity of the net hackers such as the VRML group. They speed up the process, but they are subject to the lack of any policies that keep the forces in balance. Hacked standards are underdeveloped when released. Once they get a toehold and go to more formal methods for development, their development slows down just as much as ISO or ANSI. Of these, the consortium standards work because they force vendors to confront each other directly on the issues and that does a lot of good. The consortia eventually deal with other standards bodies because they usually have the same people on them, and everyone understands the dangers of duplicative or conflicting agreements. Of course, timing and whose design gets chosen are always negotiable, and that is what turns the carrousel. Deals are being struck left and right among all of these parties to create powerbloc alliances. Eventually, something workable emerges. The powerblocs get to pick the "winners" and the "designated geniuses" to promote into the next round of play. Both pay dearly in time and effort and stress for that promotion. Some reward, eh? |-) What can a poor boy do: Hack on the network. For those of you who feel fed up or disenfranchised, I suggest the protocol game. Explore an existing but mostly unimplemented standard: HyTime/SGML. Sure, the other notation and protocol reps will vilify this, but, do you care? This is comp-text-SGML, not PDF, RTF, EXPRESS, etc. Did we go to the trouble to learn the language not to use it or to spend our time debating its competitors? The discussion occurs in the best public arena for SGML: The CTS Board. Anyone can contribute; no one need travel; everyone sees the results. There is no obligation to conform or implement. Just explore for the sake of learning and teaching something valuable. If you like the SGML Open catalog, detail how it contributes to the game. If you think a class is better than a typedef, riff on the code. If you think that HTTP is all and everything we need, show by example how it is independent of the notations and how it supports interprocess and interapplication communication without violating the notations. View the game as a large improvisation in code and ideas. Don't assume we all know what you know. Teach by example. Learn by practice. That is what the CTS does best and often. Do it with code instead of advertisements. Do that and I promise to stop spewing polemics. Who will use it: The game only results in examples, but these are useful to anyone who needs to compare solutions. They encourage better designs. At the very least, it is a way of expressing requirements and preferences without going down on one knee and putting on a ring. Keep asking questions, Alex. I do realize the quixotic nature of such a game, but sometimes a windmill IS a dragon, and sometimes, an old horse, a broken sword, and a disreputable companion are just enough to change one dragon back into one working windmill. Then make bread. Len Bullard From: rlockwood@novell.com (Robert Lockwood) Newsgroups: comp.text.sgml Subject: SGML and double-byte character sets Date: 22 Jun 1995 20:31:16 GMT Organization: Novell Message-ID: <3scjuk$grb@grok.provo.Novell.COM> X-Newsreader: WinVN 0.92.6+ Can anyone point me to information on native SGML handling of double-byte character sets, either theoretical or in actual practice? Are there existing systems that take raw SGML data and drive a production publishing process in Chinese or Japanese (for example)? On the theoretical side: - How would it be most appropriate to structure the files with regards to the markup vs. CDATA components (e.g.: do you use ASCII for the markup and a double-byte representation for the text? If so, do you expect the parser to make a byte by byte determination of whether it is reading a single- byte character or half of a double-byte character?) - What character set(s) would one expect to utilize? Are there unicode efforts that are sufficiently advanced to serve as the basis for production applications? - Assume that a robust SGML repository environment were put in place to hold information in various languages. Would one expect to be able to mix & match between, say, an English document component represented in ISO-latin, and a Chinese component in a unicode representation? - Are there platform, or other issues to be considered? On the practical side: - Do real tools exist that can deal with such representations, in the SGML editing, conversion, repository, publishing, and electronic delivery arenas? - Do any of the SGML product vendors monitoring the group care to comment on future directions in this area? From: jsc@tabitha.atype.com Newsgroups: comp.text.sgml Subject: Help using James Clark's SP parser Date: 22 Jun 1995 18:22:34 -0400 Organization: Massachvsetts Institvte of Technology Message-ID: \ Reply-To: jin@atype.com X-Newsreader: (ding) Gnus v0.87 Is there any documentation of any sort relevant to James Clark's SP SGML parser? I've been grovelling through the code and examples, but this could take a while and any documentation at all would be a great help. I just want something to convert HTML to a tree-like data structure. From: bebb@ferndown.ate.slb.com (Malcolm Bebb) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: Thu, 22 Jun 1995 22:24:01 +0000 Organization: Schlumberger Technologies ATE Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> In article <3sa7hq$qqc@googol.bctel.net>, Andrew Haigh \ wrote: > > There are plans for a Mac and a Unix version to be > implemented down the line. As for Dos, once Windows `95 hits > the streets, Dos is as dead as my old Sinclair ZX-80. ROTFL. No, not about the ZX-80. Well, all right, but separately. You've just wrecked whatever credibility you had left. Corporate users will be using DOS for donkeys years yet. As for Windows 95 (=~ Mac '83) "hitting the streets" - it's gonna need a major hardware upgrade just to run the OS, the RAM alone will cost more than a decent Windows machine (oxymoron I know, but just for the sake of argument). As far as I can see, assuming that RTC isn't just a spoof, it seems to combine the worst aspects of pdf and html while offering nothing new and arriving too late. Malcolm bebb@ferndown.ate.slb.com Tech Pubs bebb@embetech.demon.co.uk From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: RTMF (Was Re: NAMELEN -- SGML Declaration -- How to?) Date: 23 Jun 1995 07:28:00 GMT Organization: SEMA Group Belgium Message-ID: <3sdqe0$7is@news.Belgium.EU.net> References: \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) Kai Grossjohann \ wrote: >Hi all, > >a very short question: > > How do I allow names up to 72 characters? > >Some more info: > >I know that I need an SGML declaration for this which might look like >this: > > \ > >...or it might not. .. >I tried putting the SGML declaration above and below this line, but it >didn't work either way. (Above: error msg `sgml declaration not >allowed here', below: didn't have any effect.) > >Is this question answered in Goldfarb's book? I now have only ``SGML >-- An Author's Guide'' by Martin Bryan which doesn't seem to explain >this. > .. On second thought, I think you need to read something elementary about the SGML declaration. It's not done like you have tried it. The NAMELEN can be specified in the Concrete Syntax part of the declaration. The following pointers should help you : Charles Goldfarb, The SGML handbook, p. 450-477; Martin Bryan, An Author's guide to the Standard Generalized Markup Language, p. 72-73 and p. 225-262 (maybe you didn't get THAT far in the book); Eric Van Herwijnen, Practical SGML (1st ed.), p. 99-120. These books show numerous examples of variant SGML declarations. Of course, they don't tell you how to make your particular SGML tool work with a non-default SGML declaration. With most tools, putting it ahead of the Doctype and Element declarations in your DTD file should work. If you have a SGML declaration separate from the DTD, the following should work (with PATH the local filename of your SGML Decl.) : With Mark-It, specify the parameter -gPATH With OmniMark, I haven't found out (nothing in the doc, as far as I am aware), but you can include the declaration at the top of the DTD With tools which need a "compiled" version of the DTD, such as SoftQuad's RulesBuilder or WordPerfect's DTD2LGC, there's always possibility of using a switch telling what SGML declaration to use (e.g., with WordPerfect's DTD2LGC, use the parameter /s PATH). Read your manuals and documentation. Of course, they tend to disappear, even with legal copies of software. Judging from your other posts, did you come to SGML by the HTML path ? Best regards, ----------------------------------------------------------- Jacques Deseyne \ SEMA Group Belgium Stallestraat 96 (we have moved !) B-1180 Brussels Belgium Tel. + {32} 2 333 53 24 Fax. + {32} 2 333 55 22 ----------------------------------------------------------- From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: Re: SGML and double-byte character sets Date: 23 Jun 1995 08:21:28 GMT Organization: SEMA Group Belgium Message-ID: <3sdti8$7is@news.Belgium.EU.net> References: <3scjuk$grb@grok.provo.Novell.COM> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) rlockwood@novell.com (Robert Lockwood) wrote: >Can anyone point me to information on native SGML handling >of double-byte character sets, either theoretical or in >actual practice? Are there existing systems that take raw >SGML data and drive a production publishing process in Chinese >or Japanese (for example)? > Although this newsgroup is not intended for commercial purposes, I thought I can mention the 16-bit version of Mark-It, which is commercialized in the Eastern Asia region by Nissho Iwai Infocom Systems Section 1. Electronic Systems 4-5 Akasaka, 2-chome Minatu-ku Tokyo 107 Japan Mr. Akiro Mizobuchi Tel. + {81} 33 588 4598 Fax. + {81} 33 588 4883 Mr. Tetsuro Shinsu Fax. + {81} 3 58009149 which can provide you information on other SGML and publishing tools for 16-bit languages. >On the theoretical side: >- How would it be most appropriate to structure the files > with regards to the markup vs. CDATA components (e.g.: do > you use ASCII for the markup and a double-byte representation > for the text? If so, do you expect the parser to make a byte > by byte determination of whether it is reading a single- > byte character or half of a double-byte character?) It is possible to use a non-ASCII concrete syntax (although not in the SGML declaration) for your markup. >- What character set(s) would one expect to utilize? Are there > unicode efforts that are sufficiently advanced to serve as the > basis for production applications? The Editor from TimeLux is UNICODE-based. >- Assume that a robust SGML repository environment were put in > place to hold information in various languages. Would one expect > to be able to mix & match between, say, an English document > component represented in ISO-latin, and a Chinese component > in a unicode representation? I have not really investigated this, but it looks impossible to change the BASESETs from the SGML declaration within one document. Even with SUBDOC, the subdocument entity must conform to the SGML declaration of the SGML document entity. What do you mean by "mix & match" ? >- Are there platform, or other issues to be considered? > >On the practical side: >- Do real tools exist that can deal with such representations, > in the SGML editing, conversion, repository, publishing, > and electronic delivery arenas? (See above) >- Do any of the SGML product vendors monitoring the group care > to comment on future directions in this area? We do. Let us gather the necessary information. The SGML Asia '95 Conference organized by GCA (the week of Oct. 25 ?) could provide you a lot of information on these topics. ----------------------------------------------------------- Jacques Deseyne \ SEMA Group Belgium Stallestraat 96 (we have moved !) B-1180 Brussels Belgium Tel. + {32} 2 333 53 24 Fax. + {32} 2 333 55 22 ----------------------------------------------------------- From: Kai Grossjohann \ Newsgroups: comp.text.sgml Subject: Re: NAMELEN -- SGML Declaration -- How to? Date: 23 Jun 1995 12:53:28 +0200 Organization: Universitaet Dortmund, CS Dept, Chair 6 Message-ID: \ References: \ <3sc866$na4@news.Belgium.EU.net> Reply-To: Kai Grossjohann \ In-reply-to: Jacques Deseyne's message of 22 Jun 1995 17:10:30 GMT To: Jacques Deseyne \ X-Newsreader: (ding) Gnus v0.87 >>>>> "Jacques" == Jacques Deseyne \ writes: Jacques> [...] The tools you are using (which ones, BTW ?) I use Emacs with psgml.el 0.4b3 to edit the documents and sgmls 1.1 to parse them. Also, I might be using gf 0.44 to create HTML output. Jacques> assume an SGML declaration. If you don't specify one, the Jacques> defaults (SGMLREF) are assumed, which limit NAMELEN to 8. I see. Jacques> Read something on this in the Annexes to ISO 8879, Jacques> reproduced and explained in Charles Goldfarb's SGML Jacques> Handbook, or in the other classics from Martin Bryan, Eric Jacques> Van Herwijnen, and others. A lot of public DTDs come with Jacques> their SGML declaration. I read in the book from Martin Bryan but didn't think an SGML declaration needed to be `complete' in that it needs to contain all those entries. I thought just specifying those parameters deviating from the default would be enough. Jacques> The declaration you mention will choke every parser. The error message was misleading. It told me I couldn't use `\ Here's an example of the SGML declaration changing the Jacques> default value for NAMELEN. [...] You example, though, showed me that I was wrong: Just putting the SGML declaration before the \ statement seemed to work, sort of. (You were omitting parts of the SGML declaration which caused it not to work.) Jacques> You may want to change other things, such as the BASESET in Jacques> order to use 8-bit characters such as ISO 8859-1 (you live Jacques> in Germany and not in the US, after all). Good idea. Is there any example you can give me to see how I would do that? Jacques> It depends on the tools you use how you make your Jacques> environment "aware" of the non-reference SGML declaration. Jacques> You can place it at the start of your DTD file, although Jacques> not all tools will cope with that. I think I'll try that. Jacques> With sgmls, you can specify declaration, dtd and document Jacques> instance as parameters and sgmls will concatenate these Jacques> files. Jacques> Mail me if you want some real examples of non-reference Jacques> SGML declarations. I think using the HTML declaration I will be fine. Thanks for your offer, I will first try the HTML declaration and contact you again if it doesn't work. Thanks a lot for your kind help! \\kai{} -- Life is hard and then you die. From: kalliso1@ix.netcom.com (Keith Allison) Newsgroups: comp.text.sgml Subject: Implementation of SGML Date: 23 Jun 1995 13:53:07 GMT Organization: Netcom Message-ID: <3seh03$c4@ixnews2.ix.netcom.com> We are surveying the implementation of SGML among commercial firms to ascertain if our firm should make a strong commitment to SGML and if we should recommend SGML to our numerous clients. We would appreciate examples of firms that have made a commitment to SGML, not just by lip service, but by actual implementation in daily work. (For example, Novell and its production of NetWare manuals). In particular we are interested in: (1) Examples of successful implementations of SGML with specifics on how much it cost, how much it saved, and how it was done. (2) Examples of unsuccessful implmentations and why. (3) Evaluation of existing tools and their biggest weaknesses (not their strengths. This is readily available) (4) Opinions on the vendors and consultants in the industry with regard to the quality of their products, customer service, etc.. Our general impression to date has been that alot of commericial firms say they are behind SGML and it is important, but few are actually using it beyond the prototype phase and fewer are actually receiving a significant Return on Investment (ROI). Private e-mail messages are welcome and confidentiality is ensured if you prefer to remain anonymous. -- Keith Allison Web Designer for Ernst & Young at http://www.ey.com/us/tax/eyustax.htm Personal site at http://www.webcom.com/~kalliso1 From: Tony Jones \ Newsgroups: comp.text.sgml Subject: Re: Compressed RTF (RTC) to challenge PDF on the WWW Date: 23 Jun 1995 14:08:26 GMT Organization: Miles 33 Ltd, UK. Message-ID: <3sehsq$9gg@miles33.co.uk> References: <3sa0n5$mre@googol.bctel.net> <3sbt6k$ss8@maroon.tc.umn.edu> X-NNTP-Posting-Host: antigua.demon.co.uk Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) It has something to do with SGL in that acceptance as a de facto Web publishing standard will reduce the drive for implementing SGML publishing solutions From: glenn@rice-chex.ai.mit.edu (Glenn A. Adams) Newsgroups: comp.text.sgml Subject: Re: SGML and double-byte character sets Date: 23 Jun 1995 14:11:39 GMT Organization: MIT Artificial Intelligence Lab Message-ID: <3sei2r$mlt@life.ai.mit.edu> References: <3scjuk$grb@grok.provo.Novell.COM> <3sdti8$7is@news.Belgium.EU.net> In article <3sdti8$7is@news.Belgium.EU.net>, Jacques Deseyne \ wrote: >rlockwood@novell.com (Robert Lockwood) wrote: > >>- Assume that a robust SGML repository environment were put in >> place to hold information in various languages. Would one expect >> to be able to mix & match between, say, an English document >> component represented in ISO-latin, and a Chinese component >> in a unicode representation? > >I have not really investigated this, but it looks impossible to >change the BASESETs from the SGML declaration within one document. >Even with SUBDOC, the subdocument entity must conform to the >SGML declaration of the SGML document entity. I think it's important to distinguish between the character encoding used in the representation of an entity as a storage object and the document character set employed by the entity as such. One cannot vary the latter, but the former could be mixed. For example, one might employ BIG5 in the storage object of one entity and Shift JIS in the storage object of another entity and use both entities in a document whose DESCSET is based on ISO/IEC 10646. The entity manager is then responsible for translating the storage object encoding to a system character encoding which supports the semantics of (i.e., is consistent with) the document character set. Note that this system character encoding may or may not be identical to the document character set. It could be any encoding as long as it provided for an equivalent representation and semantics of every described character. Glenn Adams From: rlockwood@novell.com (Robert Lockwood) Newsgroups: comp.text.sgml Subject: Re: SGML and double-byte character sets Date: 23 Jun 1995 14:35:48 GMT Organization: Novell Message-ID: <3sejg4$buv@grok.provo.Novell.COM> References: <3scjuk$grb@grok.provo.Novell.COM> <3sdti8$7is@news.Belgium.EU.net> X-Newsreader: WinVN 0.92.6+ In article <3sdti8$7is@news.Belgium.EU.net>, Jacques Deseyne \ says: > rlockwood@novell.com (Robert Lockwood) wrote: >>- Assume that a robust SGML repository environment were put in >> place to hold information in various languages. Would one expect >> to be able to mix & match between, say, an English document >> component represented in ISO-latin, and a Chinese component >> in a unicode representation? > >I have not really investigated this, but it looks impossible to >change the BASESETs from the SGML declaration within one document. >Even with SUBDOC, the subdocument entity must conform to the >SGML declaration of the SGML document entity. >What do you mean by "mix & match" ? > Thank you for the useful information, Jacques. I agree that SGML Asia will provide a good opportunity to explore these issues further. On the "mix & match" question, suppose a \ element in turn contains \
elements. I want to build a \ using a \
authored as part of an English document and another \
authored as part of a Japanese document. If the English \
is in ISO-latin whereas the Japanese \
is in unicode, it seems I won't be able to utilize them within one \ without first converting the English \
to unicode. More generally, it would seem that the entire contents of an SGML document, as presented to a validating parser, must conform to the same character set. But since I would expect unicode to provide a superset of the possible characters supported in ISO-latin, it would seem that conversion of documents in that direction would always be feasible. From: Physicians' Online <71154.760@CompuServe.COM> Newsgroups: comp.text.sgml Subject: Help needed in SGML Interpretation Date: 23 Jun 1995 14:38:06 GMT Organization: via CompuServe Information Service Message-ID: <3sejke$8oe$1@mhadf.production.compuserve.com> Most of the messages I have seen (and the software I have seen) is based on the concept that the user wishes to create SGML documents. I am on the other side of the issue; my company is going to be receiving books marked up using SGML, and we need to be able to interpret the files. I am interested in what books and/or software libraries are out there (UNIX) which could help us with our task. If someone has some decent internal documentation of SGMLS, that would be great. Please send EMAIL to blidofsky@po.com, although email sent to this address will reach me, eventually. Bart Lidofsky From: lalthen@wtcd.DaytonOH.ATTGIS.COM (Lee Althen) Newsgroups: comp.text.sgml Subject: MS SGML Author Application Building Experience Message-ID: <1813@wtcd.DaytonOH.ATTGIS.COM> Date: 23 Jun 95 14:46:57 GMT Followup-To: Lee.Althen@daytonoh.attgis.com Organization: AT\&T Global Information Solutions - Dayton, OH USA I have recently received a copy of Micosoft SGML Author. After experimenting with it and reading the rather sparse System Administrator Guide (booklet), I am of the opinion that a rather serious development effort is required in order to have a useful tool for authoring with anything but the simplest of DTDs. The sample partial implementation of the CALS DTD included with the MS SGML Author, I believe, is an example of the development required. Does anyone have experience building / using this tool? Has anyone implemented the tool for use with DocBook? Where can I find some expertise to help with building an application? Thanks in advance, Lee Althen Lee E. Althen (WHQ-2) Lee.Althen@wtcd.DaytonOH.ATTGIS.COM Information Products Division ...!uunet!ncrcom!wtcp!wtcd!lalthen AT\&T GIS, So Patterson Bldv FAX: (513) 445-2814 Dayton, OH 45479 Tele: (513) 445-3487 -- Lee E. Althen (WHQ-2) Lee.Althen@wtcd.DaytonOH.ATTGIS.COM Information Products Division ...!uunet!ncrcom!wtcp!wtcd!lalthen AT\&T GIS, So Patterson Bldv FAX: (513) 445-2814 Dayton, OH 45479 Tele: (513) 445-3487 From: martyn davison \ Newsgroups: comp.text.sgml Subject: Re: GUI for DTD editing? Date: Fri, 23 Jun 1995 14:52:48 GMT Organization: OMI Logistics Ltd Message-ID: <588163537wnr@omimjd.demon.co.uk> References: <3rpght$ru8@usenet.srv.cis.pitt.edu> <66560144wnr@omimjd.demon.co.uk> <3s8u2u$i3j@news.xs4all.nl> Reply-To: martyn@omimjd.demon.co.uk X-NNTP-Posting-Host: omimjd.demon.co.uk X-Broken-Date: Friday, Jun 23, 1995 14.41.37 X-Newsreader: Newswin Alpha 0.7 In article: <3s8u2u$i3j@news.xs4all.nl> jang@xs4all.nl (Jan Grootenhuis) writes: > > martyn davison \ writes: > > [...] > >I agree with others on this thread. Try Near\&Far. > >It does not get your around the need for the analysis but it > >does help a lot creating the DTD > > If that's the sort of help you need, you should not be creating DTDs in > the first place. BTW Near\&Far does not even support the reference > Concrete Syntax in NAMECASE GENERAL YES ENTITY NO, let alone any changes > in the other NAMING aspects. > > Cheers, > Dear oh dear I do seem to have touched on a raw nerve don't I. As far as being a productivity tool goes it is still the best generally available. OK so there are some draw backs but isn't there a downside to most things? The problem that we have around in the real world is that with the higher profile of SGML in the market place there are more and more naive users out there who know about documentation back to front having been creating it to very rigid specifications for years but who are still unfamiliar with SGML. SGML editors are becoming much more user friendly (Incontext and Grif are examples) but as far as working with DTD's is concerned there is a long way to go. One of the main problems that these naive users are having to face is the question of multi file DTD's. Now I recognise that N\&F cannot export these (and they have good reasons for putting forward the view they do) but using a good graphical DTD programme they can see the structure. I used to know a man who would never use compilers for creating programme code - he did it all in machine code - he could read it as though it was a novel, thousands of lines of the stuff, and spot mistakes. He was really good at it. There are those around who decry productivity tools for creating DTD's being much happier to create them using a text editor and parsing them regularly. Typesetting was an are (still is to the purists) but W4W can give a very good output with care more than adequate for most uses. So for most uses why make things more complicated than necessary? What I am really saying is that if _all_ the Good parts of SGML are to be realised then there has to be the tools to (in some ways) de skill the work and bring it down from the dizzy heights to the ground level. Instead of decrying those who cannot handle the new methods why not come alongside them and help them along the path? Perhaps you touched a raw nerve/hobby horse of mine? -- Martyn Davison OMI Logistics Ltd UK. Email martyn@omimjd.demon.co.uk From: djones@imagen.net (Dwight Jones) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 23 Jun 1995 15:20:27 GMT Organization: Imagen Communications Inc. Message-ID: <3sem3r$de0@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ Mime-Version: 1.0 X-Newsreader: WinVN 0.99.3 > >Uh, I tried www.imagen.net, imagen.net, imagen.com, and www.imagen.com -- >no DNS (pronounced "no dice!"). Is it new and not fully propagated yet? It's at www.imagen.net Please call me collect at 604.537-1950 if this does not connect. Dwight Jones djones@imagen.net From: djones@imagen.net (Dwight Jones) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 23 Jun 1995 15:25:02 GMT Organization: Imagen Communications Inc. Message-ID: <3semce$de0@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ Mime-Version: 1.0 X-Newsreader: WinVN 0.99.3 > >I tried to get the free viewer, but found us Mac users left in the cold. >Seems UNIX and DOS users are out of luck too. I guess "blown away" can be >taken literally... > >It is intended for the Win95 front end. It makes use of compressed sound and the JPEG compression/decompression found in Win95. Unfortunately, we don't see the object oriented OS promised by Apple (Taligent) as being available yet or we could provide an object. We build RTC for speed and the 85% of the market that will run Win95. As the company grows we shall expand our product base. Apologies Dwight Jones From: djones@imagen.net (Dwight Jones) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 23 Jun 1995 15:29:34 GMT Organization: Imagen Communications Inc. Message-ID: <3semku$de0@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> <3sc5tr$je0@crocus.csv.warwick.ac.uk> Mime-Version: 1.0 X-Newsreader: WinVN 0.99.3 > >: There are plans for a Mac and a Unix version to be >: implemented down the line. As for Dos, once Windows `95 hits >: the streets, Dos is as dead as my old Sinclair ZX-80. > >The only people who are losing out by making such a ludicrous >assumption are yourselves - there is a massive installed >base of DOS users out there, who may have no need to upgrade. For DOS users, e.g. every caller would have to load a video driver to see the images above 16 colors. Would you predicate an image-based document format on that scenario? And a sound driver? And so on. We are Microsoft afficionados, and don't apologize for it. You can't ignore 85% of the market. From: masonjd@ornl.gov (James David Mason) Newsgroups: comp.text.sgml Subject: PDF and SGML Date: Fri, 23 Jun 1995 15:30:26 GMT Organization: Convenor, ISO/IEC JTC1/SC18/WG8 Message-ID: \ X-Newsreader: Trumpet for Windows [Version 1.0 Rev A] I haven't been reading this newsgroup an a while, and I gather I've missed some things. But I was responding to a message from another source and thought this might make fodder for discussion here. In the other context, someone had raised questions of retrievability/searchability of documents. The following is the results of some tests we've run at ORNL: I'll give you an example of the problems with searcing PDF. This is a known bug, acknowledged by Adobe. Take a document printed in multiple columns and distill it into PDF. Bring it up in an Acrobat viewer. You will see if your system is slow enough to see things happening) that the columns on the screen fill in the order in which the composition system built them--normally down column 1, then down column 2, etc. Now pick some phrase that you can see that wraps lines in either column and search for it. You won't get the hit. However, you can get a hit if you construct a search string consisting of characters at the end of line n in the last column and the beginning of line n+1 in the first column--even though no such string exists in the source document!! You can also get hits on jumps across the gutters between columns. What is happening is that after they build a page image, they search the displayed characters on the page as a whole, without regard to how the characters got there. The result is that you can't find things that are there but you do find things that aren't there. Any wonder I'd rather search my own SGML source than the PDF generated from it? I can't show you my best examples, but to show that it's not just a local phenomenon, try something from another DOE laboratory. Go to http://www.lanl.gov/computer-information/ComputingNews/1995Archives/May1995 and pick up the PDF form of the May issue of their computing newsletter. Open it in Acrobat and look for "red hand". You'll get "considered" in column 2 and "handling" in column 1, some hundreds of words earlier in the source text. Dr. James D. Mason (ISO/IEC JTC1/SC18/WG8 Convenor) Oak Ridge National Laboratory Information Management Services Bldg. 2506, M.S. 6302, P.O. Box 2008 Oak Ridge, TN 37831-6302 U.S.A. Telephone: +1 615 574-6973 Facsimile: + 1 615 574-6983 Network: masonjd@ornl.gov From: Horst Szillat \ Newsgroups: comp.text.sgml Subject: Re: Novice question -- Reusing a DTD? Date: Fri, 23 Jun 1995 17:48:23 +0200 Organization: Unlimited Surprise Systems, Berlin Message-ID: \ References: \ <3s9nh5$11r@crl2.crl.com> \ Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Sender: szillat@berlin In-Reply-To: \ On 22 Jun 1995, Kai Grossjohann wrote: > Actually, I've got a database of SGML files, intended to be displayed > in HTML. Rather than making up my own syntax for things like > imagemaps, links, and suchlike, I thought I'd just reuse the html > stuff. > > The database is to contain documents of different kinds, semantically > tagged for the convenience of later retrieval. (A lecture would be > tagged with attributes like title, lecturer, time and place, abstract > and others.) > > I though it would be nice if the users could embed any HTML markup > they cared in, say, the abstract field for nice looks (or in the > lecturer field, for that matter, to embed a link to their homepage or > something). > > Would you think this is an unreasonable approach? I am not shure if it is really a solution of your problems but perhaps SUBDOC is an idea. Horst Szillat email: szillat@berlin.snafu.de url : http://www.snafu.de/~szillat/ H. Szillat: SGML - Eine praktische Einf\ührung International Thomson Publishing, ISBM 3-929821-75-3 From: patt@ais.berger-levrault.fr (P. ATTAR) Newsgroups: comp.text.sgml Subject: Chemical entities Date: 23 Jun 1995 21:03:07 GMT Organization: Pierre Attar Message-ID: <3sfa6b$s45@vishnu.jussieu.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Newsreader: WinVN 0.92.6 Hi, I wonder where I can find the standard entity set to use for chemical symbols such as triple long line used to link molecules ? Also, it's sometimes difficult to figure out what is the graphical representation of an entity ... is there some cook file reference I can use ? What is the ISO standard describing both entities and their representation ? Thanks a lot for answering, Pierre Attar (France) patt@ais.berger-Levrault.fr From: quixote@Primenet.Com (Donald A. Hosek) Newsgroups: comp.text.sgml Subject: DTD analysis tool Date: 24 Jun 1995 00:11:35 GMT Organization: PrimeNet Message-ID: <3sfl7n$4ci@nnrp1.primenet.com> DOes anyone have any tools for analyzing DTDs? A couple things that would be handy would be - a tool to list the minimum required document frame given a DTD - a tool that takes the above one step further by indicating optional tags, stopping when it detects a recursion possibility (this would be a bit tougher to represent the output of, although even the first will require some nastiness). Any suggestions? -dh -- Don Hosek dhosek@quixote.com 909-621-1291 fax: 909-625-1342 Quixote Digital Typography http://www.quixote.com/ Publishers of _Serif: The Magazine of Type and Typography_ From: quixote@Primenet.Com (Donald A. Hosek) Newsgroups: comp.text.sgml Subject: DTD analysis tool Date: 24 Jun 1995 03:15:03 GMT Organization: PrimeNet Message-ID: <3sfvvn$v2@nnrp1.primenet.com> Cc: DOes anyone have any tools for analyzing DTDs? A couple things that would be handy would be - a tool to list the minimum required document frame given a DTD - a tool that takes the above one step further by indicating optional tags, stopping when it detects a recursion possibility (this would be a bit tougher to represent the output of, although even the first will require some nastiness). Any suggestions? -dh -- Don Hosek dhosek@quixote.com 909-621-1291 fax: 909-625-1342 Quixote Digital Typography http://www.quixote.com/ Publishers of _Serif: The Magazine of Type and Typography_ From: Paul_Hermans@protext.be (Paul Hermans) Newsgroups: comp.text.sgml Subject: conloc Date: Sat, 24 Jun 1995 11:45:03 +0100 Organization: Pro Text Message-ID: \ Are there any tools in the marketplace yet that support the content location attribute of the all-id attribute list form of the base module of HyTime ? Thanks, Paul From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: GUI for DTD editing? Date: 24 Jun 1995 13:29:41 GMT Organization: Naggum Software; +47 2295 0313 Expires: 8 Jul 1995 13:29:41 GMT Message-ID: <19950624T132941Z@naggum.no> References: <3rpght$ru8@usenet.srv.cis.pitt.edu> <66560144wnr@omimjd.demon.co.uk> <3s8u2u$i3j@news.xs4all.nl> <588163537wnr@omimjd.demon.co.uk> [Martyn Davison] | There are those around who decry productivity tools for creating DTD's | being much happier to create them using a text editor and parsing them | regularly. I don't think such people exist -- it only looks like they do because the existing tools to help create DTD's fall short of the requirements that can be fulfilled with a good text editor (we're not talking EDLIN). now that I'm allergic to (gratuitous) graphics, I probably won't be satisfied soon, but that doesn't mean I'm "decrying productivity". I _am_ more productive with text based tools and languages than I am with graphics stuff. that's because I use an editor that I can program and teach to do lots of the tedious, error-prone stuff (GNU Emacs from the Free Software Foundation). | Typesetting was an are (still is to the purists) but W4W can give a | very good output with care more than adequate for most uses. So for | most uses why make things more complicated than necessary? because it is currently easier than possible, so your view of what is "necessary" is a little skewed. also, automation and computers is about removing the need for "care" where it should have been unnecessary for the operator to do things manually. it is _always_ harder to automate things than to do it manually, because we need to provide the machine with the information that the user had while we were doing it manually. the only complication lies in communicating this intention to the computer. | What I am really saying is that if _all_ the Good parts of SGML are to | be realised then there has to be the tools to (in some ways) de skill | the work and bring it down from the dizzy heights to the ground level. SGML's good parts come from the explicit expression of the subconsciously held knowledge about the information being created, such as structure, relations between elements, meta-information, etc. to communicate this requires a certain willingness to introspect creatively that will remain above the "ground level". that's not to say we shouldn't build elevators. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: Implementation of SGML Date: 24 Jun 1995 13:52:46 GMT Organization: Naggum Software; +47 2295 0313 Expires: 8 Jul 1995 13:52:45 GMT Message-ID: <19950624T135246Z@naggum.no> References: <3seh03$c4@ixnews2.ix.netcom.com> [Keith Allison] | We would appreciate examples of firms that have made a commitment to | SGML, not just by lip service, but by actual implementation in daily | work. (For example, Novell and its production of NetWare manuals). I'm not sure I can name it, but a book is soon coming out on International Thompson Publishing that I think should satisfy your needs for information on Real Life SGML. I have reviewed this book, and I was deeply impressed. I said earlier, and am happy to to repeat, that this is the first book about SGML that I can whole-heartedly recommend. (Goldfarb's SGML Handbook isn't _about_ SGML, it _is_ SGML, and is still unsurpassed.) | In particular we are interested in: | | (1) Examples of successful implementations of SGML with specifics on | how much it cost, how much it saved, and how it was done. | (2) Examples of unsuccessful implmentations and why. | (3) Evaluation of existing tools and their biggest weaknesses (not | their strengths. This is readily available) | (4) Opinions on the vendors and consultants in the industry with regard | to the quality of their products, customer service, etc.. I can't promose that it will answer all these questions, but it should give you the important leads to the answers. | Our general impression to date has been that alot of commericial firms | say they are behind SGML and it is important, but few are actually | using it beyond the prototype phase and fewer are actually receiving a | significant Return on Investment (ROI). my impression is that adoption of SGML is moving from the publishing end closer to the source of the information. I have gone so far as to say that SGML is nearly worthless if used only for publishing. I spent some time reaching this realization myself, so I'm pleased by the speed that the market is also reaching it. there's a long way left to go, however. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML and double-byte character sets Date: 24 Jun 1995 14:05:35 GMT Organization: Naggum Software; +47 2295 0313 Expires: 8 Jul 1995 14:05:35 GMT Message-ID: <19950624T140535Z@naggum.no> References: <3scjuk$grb@grok.provo.Novell.COM> <3sdti8$7is@news.Belgium.EU.net> [Jacques Deseyne] | It is possible to use a non-ASCII concrete syntax (although not in the | SGML declaration) for your markup. this implies that the SGML declaration must be in ASCII, which is _wrong_. the SGML declaration must only be in some character set that the computer on which the parser runs understands, known as the "system character set". Martin Bryan is the source of the _extremely_ unfortunate and unsupported belief that the SGML declaration must be in ASCII. this belief has been thoroughly repudiated here in comp.text.sgml previously, and I'm sorry to see that the SEMA Group still caters to this belief. it is possible to use _any_ character set whatsoever for the entire SGML document. the SGML declaration will specify the character set used. when an SGML document is translated from one character set to another, the SGML declaration will certainly aid the process, but it must also be translated at a more semantic level to reflect the new character set. that is, the SGML declaration does _not_ serve the purpose of document interchange that some like to impute to it, however much we would like it to do so. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: Help with content model? Date: 24 Jun 1995 14:33:15 GMT Organization: Naggum Software; +47 2295 0313 Expires: 8 Jul 1995 14:33:15 GMT Message-ID: <19950624T143315Z@naggum.no> References: <3s4j3u$uo@argo.hks.com> <9506202221.AA0720@notes.microstar.com> [Hugh Chatfield] | In all examples the parser cannot know whether the first B belongs to | the "B"s following the A or the "B"s following the C (not even when the | B follows an A). To fix it for the parser, you need a mechanism which | will allows the two B's to be distinguished (like the space tries to do | for us humans in my examples). | | If I presented you with option 1 or 6 without the space, I presume you | couldn't tell which B it was either. I suggest your model is perhaps | incomplete - it is missing that semantic which would allow the authors | (and the parser) to distinquish between the two B's. I disagree with the premises of this analysis. there are no means in SGML to say "which B" in the content model, just "a B". it is irrelevant which content token a given element matches, as there is no way to find out, nor should there be a need to know this -- if the "which" is important, it should be labeled so, either with attributes or another element name. unfortunately, the "no ambiguous content models" rule in SGML may imply that the "which" question is legitimate. this flies in the face of the rest of the relationship between the DTD and the document, and it behooves a good analysis to realize that the ambiguity rule stems from an implementation issue, not from a design issue. in particular, the notion of "which" is thwarted by the _equivalence_ between such content models as (B, (B, B*)?) , (B, B*) and (B+). thus, if one is defending SGML's ambiguity restriction, it should be on implementation grounds, such that a parser should not at any one point have the option of choosing among more than one element with the same name. now, _this_ design issue may be implemented in a variety of ways; today's ambiguity rule is just one of them and not a particularly good one at that. that said, I hope that we will do away with the ambiguity rule in the revised standard, in particular such that the example under discussion would be legal and have the intuitive meaning. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: SGNX32A@prodigy.com (Calvin Wylie) Newsgroups: comp.text.sgml Subject: Re: BRF ? do you know What it is? Date: 24 Jun 1995 18:41:14 GMT Organization: Prodigy Services Company 1-800-PRODIGY Message-ID: <3shm8a$1rg6@usenetp1.news.prodigy.com> References: <3rqkc7$9sk$1@mhadf.production.compuserve.com> X-Newsreader: Version 1.2 > All I know that the software is >called BRF. Do you know which software it is? Or is there is a >way to convert it to PDF format? Or HTML? Here is a example of >the text Looks to me like you may mean BRS. BRS is a text database and retrieval software. The ..XX: at the beginning of some lines are paragraph labels and were defined by the database developer. They vary from database to database. To convert to anything else is a matter of writing something that maps each field into whatever form you want. If you have access to the database people from which this data came, BRS has a capability called PTF (Print Time Formatting) which I have used to output the data into HTML format. They might be able to do the same. From: Mikael Sandberg \ Newsgroups: comp.text.sgml Subject: SGML with Ventura? Date: 24 Jun 1995 20:31:30 GMT Organization: SWIPnet AB Message-ID: <3shsn2$95@mn5.swip.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.2b1 (Windows; I; 16bit) I wonder if someone knows if Corel will support SGML editor with the Corel Ventura publishing? Today I use Ventura 4.2 but soon I have to start with ADEPT Editor with SGML. Since I feel more fore traditional text editing like Ventura, I wish in the future Ventura will be able to handle SGML as ADEPT Editor do. Regards, Mikael From: djones@imagen.net (Dwight Jones) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 24 Jun 1995 21:31:27 GMT Organization: Imagen Communications Inc. Message-ID: <3si07f$dc5@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> \ Mime-Version: 1.0 X-Newsreader: WinVN 0.99.3 >As far as I can see, assuming that RTC isn't just a spoof, it seems to >combine the worst aspects of pdf and html while offering nothing new and >arriving too late. > >Malcolm bebb@ferndown.ate.slb.com >Tech Pubs bebb@embetech.demon.co.uk Au contraire, RTC has advantages over both formats, not the least of which are: 1) Smallest file sizes by far 2) Based on the RTF standard found in all major wordprocessors 3) Free editor and reader 4) HTML compatible. Is PDF? 5) Easiest composition- Major wordprocessors 6) Price- $190/server 7) Convertible back to RTF- important 8) Direct sound support in real time (1Kb/sec) 9) Compressed BMP support 10)Smallest reader by an order of difference There are others, but you will find that a lot of people value a cleaner implementation than PDF or the floating HTML version X.X. Perhaps the biggest difference is that Imagen stays within the 28.8 Kb envelope, yet still can support any variation of motion video, but simply requesting those files via hyperlinks, then directing them to Win95's decompression. If elegance counts, RTC is unsurpassed. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 25 Jun 1995 10:54:26 GMT Organization: Naggum Software; +47 2295 0313 Expires: 9 Jul 1995 10:54:26 GMT Message-ID: <19950625T105426Z@naggum.no> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> \ <3si07f$dc5@googol.bctel.net> [Dwight Jones] | Au contraire, RTC has advantages over both formats, not the least of | which are: | | 2) Based on the RTF standard found in all major wordprocessors my limited experience with RTF tells me that it not _one_ thing, but a set of unlabeled and nigh indistinguishable but related languages that differ between versions of the same word processor, and differ between each word processor. things may have improved in the past two years, but it is still a serious exaggeration to refer to RTF as "the RTF standard": it is neither "the" nor "standard". | There are others, but you will find that a lot of people value a | cleaner implementation than PDF or the floating HTML version X.X. I'm not tracking the HTML development, nor using HTML myself, but this is grossly unfair. as far as I hear people complain about HTML, it is that browsers have implemented features piecemeal and inconsistently, not that the released versions of HTML itself are any more floating than anything else in this business. moreover, it is particularly strange to see that you think HTML floats while you seem to think that RTF is just one, solid thing. | If elegance counts, RTC is unsurpassed. I don't want to take away from your work, which sounds interesting, but I would like to see other people than its creators proclaim its quality in superlatives. Lisp programmers such as myself may also have a different view of what constitutes elegance than, say, the C++ community, and since elegance is so hard to demonstrate objectively, it may not be a good idea to just refer to "elegance", but rather to show how it could apply. I'm sorry, but I don't see that demonstrated. also, I must raise an objection to the attitude that all the world's a PC with Windows. you have argued that one cannot ignore "85%" of the market. I argue that to use that as an argument to ignore "15%" of the market is so arrogant that it detracts from any other merits of a proposed solution. furthermore, the "85%" figure is vastly exaggerated. we're not talking about all the computers shipped in the world, but about users on the Internet. for PC's this figure is less than 50% of the users, soundly outnumbered by huge installations of Unix systems of diverse makes with hundreds of users per computer. regardless of the client computers, the servers remain Unix machines with _at least_ 85% "market share". should we ignore the PC with Windows? even with its still high market share for a single platform, it is unlikely that the servers on which the WWW rests will accept a format that is unreadable on the server. with the current trend in the U.S. Congress towards making Internet service providers responsible for the contents of the material shipped across their equipment, it is very unlikely that server owners will want to have files on their systems that can only be inspected visually on other equipment. I expect that the prevalence of images will be the first "victim" of this trend, and although I object strongly to the idea that service providers be responsible for the content shipped across their equipment just because they're easier targets for the legal system, it would be amusing if the early result would be a cleanup of the bandwidth-wasting image-orientedness of the WWW. anyway, given this climate, aggressive source compression of the material doesn't seem like a very good idea. on the topic of compression, I find the DOS/Windows-world's preponderance towards non-layered solutions because of the lack of multitasking on the host operating system to be alarming. we have decades of experience and research in computer science that show that modularization and layering is a very good idea for large systems (and the Internet is large). to find that DOS/Windows vendors return to the 1950's in their design philosophies because they think every possible resource is so "scarce" that no good idea will survive their "bit scare" is very depressing. under Unix systems, it is the view that a tool should do one job and do it well, and we have an archiving utility called "tar" that does not compress, we have a very good compression utility called "gzip" that works on only one file, and these are used by modern FTP servers to allow on-the-fly packaging of whole directories to be transferred as a single file. from my experience in running the SGML archive, I find that the very concept of such on-the-fly transformation simply does not register with DOS/Windows users because their world-view _precludes_ such layered solutions. now, here's an idea: the document as stored on either end will be in an uncompressed, general format. on the wire, it will be compressed by the protocol engine if requested by the receiver. of course, the compressed image may be cached by the server so as not to waste CPU-time if at a premium. the purpose of this separation is that decompression also takes time, and on a very fast network, having to decompress a document for all viewers will have severe impact on the speed of the browser, especially if displaying on X terminals and running on a loaded server system. this idea stands in sharp contrast to the design that I see in RTC which is prematurely optimized for a Windows PC on the other end of a 28,800 bps dialup link at the cost of machines with full Internet connection in the multimegabit range, at the cost of machines that don't run a particular "operating system", etc. making the compression a "feature" of the format is just plain wrong, and the arguments that you use to defend it aren't very good in a network that has been built on the premise of connecting all kinds of computers over all kinds of connections, not just one kind of machine over one kind of connection, whatever the market share of either. BTW, this idea is not novel in any way -- it's being used in many ways, including modems that talk several times faster to their computers than to the wire and do on-the-fly compression. e.g., my connection which is a nominal 14,400 bps leased line, but which averages 32kbps throughput because the modem-computer links are 38,400 bps. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: Kai Grossjohann \ Newsgroups: comp.text.sgml Subject: Re: Novice question -- Reusing a DTD? Date: 25 Jun 1995 17:49:42 +0200 Organization: Universitaet Dortmund, CS Dept, Chair 6 Message-ID: \ References: \ <3s9nh5$11r@crl2.crl.com> \ \ Reply-To: Kai Grossjohann \ In-reply-to: Horst Szillat's message of Fri, 23 Jun 1995 17:48:23 +0200 To: Horst Szillat \ X-Newsreader: (ding) Gnus v0.87 >>>>> "Horst" == Horst Szillat \ writes: Horst> I am not shure if it is really a solution of your problems Horst> but perhaps SUBDOC is an idea. Thanks a lot for this hint. SUBDOC won't quite do what I want, however. More explanation in email, if you like. \\kai{} -- Life is hard and then you die. From: grmeyer@mcs.com (Gordon R. Meyer) Newsgroups: comp.text.sgml Subject: Re: PDF and SGML Date: Sun, 25 Jun 1995 09:59:16 -0600 Organization: Our Lady of Chaos and Silicon, inc. Message-ID: \ References: \ In article \, masonjd@ornl.gov (James David Mason) wrote: > > Take a document printed in multiple columns and distill it into PDF. Bring it > up in an Acrobat viewer. You will see if your system is slow enough to see > things happening) that the columns on the screen fill in the order in which > the composition system built them--normally down column 1, then down column 2, > etc. Now pick some phrase that you can see that wraps lines in either column > and search for it. You won't get the hit. However, you can get a hit if you > construct a search string consisting of characters at the end of line n in the > last column and the beginning of line n+1 in the first column--even though no > such string exists in the source document!! You can also get hits on jumps > across the gutters between columns. Doctor! Doctor! It hurts when I do this! Then don't do it. Okay, bad joke, but as a producer of information I appreciate the tip. Since .05% of my material is formatted in multiple-columns I guess I won't lose any sleep over this bug. Come to think of it, as a consumer of information most of the PDFs I've read aren't in columns either. I guess it depends on the traditions of your discipline. Thanks for the tip. For now, I'll stick with PDF... -- Gordon Meyer \ Windows 95 = Mac 89 From: grmeyer@mcs.com (Gordon R. Meyer) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: Sun, 25 Jun 1995 10:04:42 -0600 Organization: Our Lady of Chaos and Silicon, inc. Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> <3sc5tr$je0@crocus.csv.warwick.ac.uk> > : There are plans for a Mac and a Unix version to be > : implemented down the line. As for Dos, once Windows `95 hits > : the streets, Dos is as dead as my old Sinclair ZX-80. > Hey! I'm writing this response on a ZX-81, so there! :-) Seriously, if RTC's advantage is that's a leaner alternative to PDF then it ought to work on DOS! Acrobat Reader/DOS is a bit of a resource hog...so here's your chance to shine. But, quite honestly, until there are Mac and Unix readers available. Well, I'm not interested. -- Gordon Meyer \ Windows 95 = Mac 89 From: aelman@cs.stanford.edu (Adam Elman) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: Sun, 25 Jun 1995 08:47:15 -0800 Organization: Stanford University Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3semce$de0@googol.bctel.net> In article <3semce$de0@googol.bctel.net>, djones@imagen.net (Dwight Jones) wrote: > > > >I tried to get the free viewer, but found us Mac users left in the cold. > >Seems UNIX and DOS users are out of luck too. I guess "blown away" can be > >taken literally... > > > >It is intended for the Win95 front end. It makes use of compressed sound > and the JPEG compression/decompression found in Win95. Unfortunately, we > don't see the object oriented OS promised by Apple (Taligent) as being > available yet or we could provide an object. Uh, JPEG compression/decompression has been available on the Mac as part of QuickTime for at least two years, probably three or more. And the "object-oriented" argument is ludicrous, considering that MacOS is no less "object-oriented" than is Windows 95. > > We build RTC for speed and the 85% of the market that will run Win95. As > the company grows we shall expand our product base. > > Apologies > Dwight Jones Funny thing is, I always thought that the whole point of having a platform-independent page description language was to be, well, platform-independent. Why would I use your product over, say, ZIPped Microsoft Write files? -- Adam Elman aelman@cs.stanford.edu \ From: Betty Harvey \ Newsgroups: comp.text.sgml Subject: Re: PDF and SGML Date: 25 Jun 1995 21:13:27 GMT Organization: Electronic Commerce Connection Message-ID: <3skjhn$ag@news.cais.com> References: \ \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.2b1 (Windows; I; 16bit) grmeyer@mcs.com (Gordon R. Meyer) wrote: >In article \, masonjd@ornl.gov (James David >Mason) wrote: > >> >> Take a document printed in multiple columns and distill it into PDF. Bring it >> up in an Acrobat viewer. You will see if your system is slow enough to see >> things happening) that the columns on the screen fill in the order in which >> the composition system built them--normally down column 1, then down column 2, >> etc. Now pick some phrase that you can see that wraps lines in either column >> and search for it. You won't get the hit. However, you can get a hit if you >> construct a search string consisting of characters at the end of line n in the >> last column and the beginning of line n+1 in the first column--even though no >> such string exists in the source document!! You can also get hits on jumps >> across the gutters between columns. > >Doctor! Doctor! It hurts when I do this! > >Then don't do it. > >Okay, bad joke, but as a producer of information I appreciate the tip. >Since .05% of my material is formatted in multiple-columns I guess I won't >lose any sleep over this bug. Come to think of it, as a consumer of >information most of the PDFs I've read aren't in columns either. I guess >it depends on the traditions of your discipline. > >Thanks for the tip. For now, I'll stick with PDF... >-- I would take a guess and say that 90% of all Government documents (technical manuals, administrative manuals, memos, etc.) are multiple-columns. With efforts currently underway to convert 'massive quantities' of legacy data into PDF this is not an insignificant bug. Also, most publications, reference manuals, magazines, newspapers, are multiple columns. When PDF is your only media a bug of this type could be potentially disasterous. I have also noticed, from the PDF documents that I have had the opportunity to view, that the visual quality of the document is poor. Visually they remind me of the way documents looked with CGA. Betty <><><><><><><><><><><><><><><><><><><><><><><><><><><><><> Betty Harvey harvey@eccnet.com <><><><><><><><><><><><><><><><><><><><><><><><><><><><><> From: milor001@maroon.tc.umn.edu (R A Milowski) Newsgroups: comp.text.sgml Subject: Re: Compressed RTF (RTC) to challenge PDF on the WWW Date: 25 Jun 1995 16:14:56 -0500 Organization: University of Minnesota Message-ID: <3skjkg$7rt@maroon.tc.umn.edu> References: <3sa0n5$mre@googol.bctel.net> <3sbt6k$ss8@maroon.tc.umn.edu> <3sehsq$9gg@miles33.co.uk> In article <3sehsq$9gg@miles33.co.uk>, Tony Jones \ wrote: >It has something to do with SGL in that acceptance as a de facto >Web publishing standard will reduce the drive for implementing >SGML publishing solutions > Ok, but this is marketing junk that isn't even marked as such. Its connection in the post's current state is still *null*. If it were worded as: "Will RTC replace SGML?" than it would be appropriate. I see alot of: SGML = HTML = PDF (and now) = RTC and it is just simply *not* true. ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu Copernican Solutions Incorporated (612) 825 - 4132 -- ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu Copernican Solutions Incorporated (612) 825 - 4132 From: trujillo@mail.utexas.edu (Juan A. Trujillo) Newsgroups: comp.text.sgml Subject: Small TEI sample ? Date: 25 Jun 1995 21:29:03 GMT Organization: Spanish & Portuguese, UT-Austin Message-ID: <3skkev$8up@geraldo.cc.utexas.edu> Reply-To: trujillo@mail.utexas.edu X-Newsreader: IBM NewsReader/2 v1.09 I've been exploring the TEI DTD docs for a couple of weeks, and have tried this weekend to put together a test document (by hand) to see if SGML will really work for my project (transcription and linguistic annotation of a 45,000-word corpus of Spanish language legal documents). As the situation stands, sgmls gives me 7 pages of errors for 15 lines of text. No Joy. Does anyone know where I might find a relatively small TEI-encoded text to pick apart? (The 5.6-meg SGML-marked TEI guidelines is too big for this). ALSO--Is anybody doing SGML with OS/2? Any tools available for the platform? (BookManager Build/2 SGML is still in beta, right?) Thanks. <><><><><><><><><><><><><><><><><><><><>|<><><><><><><><><><><> Juan A. Trujillo | El omne que tiene estado, onra Dept. of Spanish & Portuguese | e gran poder, lo que para si non The University of Texas at Austin | quiere, non lo deve a otros fazer <><><><><><><><><><><><><><><><><><><><>|<><><><><><><><><><><> Newsgroups: comp.text.sgml From: Steve Pepper \ Subject: Re: SGML with Ventura? Message-ID: <1995Jun25.215428.25442@falch.no> Organization: Falch Infotek, Oslo, Norway References: <3shsn2$95@mn5.swip.net> Date: Sun, 25 Jun 1995 21:54:28 GMT Mikael Sandberg \ writes: >I wonder if someone knows if Corel will support SGML editor with >the Corel Ventura publishing? Version 6.0 of Corel Ventura goes into beta testing in September. It will have SGML support. How much, I don't know, but I believe they are working with InContext on the implementation, which should vouch for an interesting product... Stayed tuned! Steve -- pepper@falch.no -------------------------------------------------------------- falch infotek a.s, postboks 130 kalbakken, n-0902 oslo, norway tel +47 2290 2733 fax +47 2290 2599 http://www.falch.no/ ^^^^^^^^^^^^ NEW ^^^^^^^^^^^^ From: gerald@cs.ualberta.ca (Gerald Oskoboiny) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 25 Jun 1995 23:53:37 GMT Organization: University of Alberta Message-ID: <3sksu1$h19@scapa.cs.ualberta.ca> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> \ <3si07f$dc5@googol.bctel.net> Reply-To: gerald@cs.ualberta.ca Originator: gerald@cab114.cs.ualberta.ca djones@imagen.net (Dwight Jones) writes: >>As far as I can see, assuming that RTC isn't just a spoof, it seems to >>combine the worst aspects of pdf and html while offering nothing new and >>arriving too late. > >Au contraire, RTC has advantages over both formats, not the least of which >are: : >4) HTML compatible. Is PDF? Define "compatible". Can I view RTC files with Lynx? Do RTC files validate against the HTML specification? I tried to find this information on your Web site, but couldn't because your product's documentation isn't viewable with any of the 7 Web browsers I have access to. >There are others, but you will find that a lot of people value a cleaner >implementation than PDF or the floating HTML version X.X. What do you mean by "floating HTML version X.X"? HTML 2.0 is stable, and widely implemented. I can view HTML 2.0 compliant files on a wider variety of platforms than any other document format, except maybe plain text. HTML 3.0 is evolving nicely, and is beginning to be supported by some browsers. I don't see any problems with HTML "versions", except maybe for documents created by people with poor HTML authoring skills. Gerald -- Gerald Oskoboiny \ http://ugweb.cs.ualberta.ca/~gerald/ From: boo@primenet.com (Walter Ian Kaye) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: Mon, 26 Jun 1995 00:39:02 -0700 Organization: Getting organized... a little more... a little more... Message-ID: \ References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> \ <3si07f$dc5@googol.bctel.net> <3sksu1$h19@scapa.cs.ualberta.ca> X-Web-Site: http://www.primenet.com/~boo/ X-FTP-Site: ftp://ftp.primenet.com/users/b/boo/ X-Quote: More insomniacs get their news from World News Now than from any other source! X-Newsreader: Value-Added NewsWatcher 2.0b27.1+ In article <3sksu1$h19@scapa.cs.ualberta.ca>, gerald@cs.ualberta.ca wrote: > djones@imagen.net (Dwight Jones) writes: > > >There are others, but you will find that a lot of people value a cleaner > >implementation than PDF or the floating HTML version X.X. > > What do you mean by "floating HTML version X.X"? HTML 2.0 is stable, and > widely implemented. > Gerald But HTML 2.0 wasn't even finalized before this month... # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # Walter Ian Kaye: (602) 942-6390 FoxPro/Excel Programmer; Guitarist # # Correspond to: boo@primenet.com, boodlums@genie.com # # BinHex files: boo@primenet.com WWW: http://www.primenet.com/~boo/ # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # From: djbpitt+@pitt.edu (David J. Birnbaum) Newsgroups: comp.text.sgml Subject: Re:Small TEI sample and SGML for OS/2 Date: 26 Jun 1995 11:44:30 GMT Organization: University of Pittsburgh Message-ID: <3sm6iu$lf@usenet.srv.cis.pitt.edu> References: <3skkev$8up@geraldo.cc.utexas.edu> X-Newsreader: NetSuite News for OS/2 [version: 4.0] :>Does anyone know where I might find a relatively small TEI-encoded text to :>pick apart? Lou Burnard has provided the following: -----[begin test.tei]----- \ \ \ ]> \ \ \ \\The world's shortest TEI document.\ \\

An unpublished document.\ \\

No source: electronic form is the original.\ \ \ \\\

The world's shortest TEI document.\\ \ -----[end test.tei]----- :>ALSO--Is anybody doing SGML with OS/2? Any tools available for the :>platform? I have compiled James Clark's SP (version 0.4) for OS/2. It is available by anonymous ftp at ftp.pitt.edu in dept/slavic/download/sgml/sp_os2.zip. Cheers, David ============================================================= David J. Birnbaum Voice: 412-687-4653 3955 Bigelow Blvd, #802 Fax: 412-624-9714 Pittsburgh, PA 15213 USA Email: djbpitt+@pitt.edu URL: http://www.pitt.edu/~djbpitt/ From: tad@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: Implementation of SGML Date: 26 Jun 1995 14:00:38 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <3smei6$3ej@cliffy.lfwc.lockheed.com> References: <3seh03$c4@ixnews2.ix.netcom.com> <19950624T135246Z@naggum.no> X-Newsreader: TIN [version 1.2 PL1] Erik Naggum (erik@naggum.no) wrote: : I'm not sure I can name it, but a book is soon coming out on International : Thompson Publishing that I think should satisfy your needs for information : on Real Life SGML. I have reviewed this book, and I was deeply impressed. Please remember to post here the name of the book when it is published. Thanx, -- --------------------------------------------------------------------- | | Tad McClellan, Logistics Specialist (IETMs and SGML guy) | Lockheed Martin Tactical Aircraft Systems | P.O. Box 748 MZ 1025 | Fort Worth, TX 76101 | | voice: (817) 763-6314 | email: mcclellantj@lfwc.lockheed.com | --------------------------------------------------------------------- | | ... Because I command the sheriff of the county to lead you away to | some remote spot, swing you by the neck from a knotting bough of | some sturdy oak, and let you hang until dead. | | And then, Jose Manuel Miguel Xaviar Gonzales, I further command such | officer retire quickly from your dangling corpse, that vultures may | descend from the heavens upon your filthy body until nothing shall | remain but bare, bleached bones of a coldblooded, bloodthirsty, | throatcutting, murdering Son of a Bitch. | | Judge Roy Bean (the Hanging Judge) --------------------------------------------------------------------- From: Ted Freeman <75701.1760@CompuServe.COM> Newsgroups: comp.text.sgml Subject: ISO 12083 (articles) Date: 26 Jun 1995 14:15:47 GMT Organization: Allen Press, Inc. Message-ID: <3smfej$ic$1@mhafm.production.compuserve.com> I have the ISO 12083:1993 DTD for articles. Where can I get supporting documentation for this standard--with full descriptions of elements and examples of their use? Is the following book what I want, and if so, where do I send for it and what am I getting? International Organization for Standardization (ISO). ISO 12083:1993(E) Information and documentation - Electronic manuscript preparation and markup. First edition, 1994-01-15. Geneva: International Organization for Standardization, 1994. 96 pages. Thanks for any advice. -- Ted Freeman E-Mail: tfreeman@allenpress.com Allen Press, Inc. Lawrence, Kansas From: datkins698@aol.com (DAtkins698) Newsgroups: comp.text.sgml Subject: BookMaster to SGML Conversion Date: 26 Jun 1995 10:24:34 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3smfv2$ru2@newsbf02.news.aol.com> Reply-To: datkins698@aol.com (DAtkins698) Does anyone know of any publicly available tools to convert BookMaster files to SGML? Any help would be greatly appreciated. I really need this information ASAP. Thanks, Diana Atkins DAtkins698 From: btuten@curly.cc.emory.edu (Belle S. Tuten) Newsgroups: comp.text.sgml Subject: TEI Lite Date: 26 Jun 1995 11:00:07 -0400 Organization: Emory University Message-ID: <3smi1n$qhl@curly.cc.emory.edu> X-Newsreader: TIN [version 1.2 PL2] Can anyone tell me where to get an electronic version of TEILite DTDs for markup? -- Belle Tuten Emory University btuten@emory.edu From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: Re: SGML and double-byte character sets Date: 26 Jun 1995 16:41:49 GMT Organization: SEMA Group Belgium Message-ID: <3smo0d$3u0@news.Belgium.EU.net> References: <3scjuk$grb@grok.provo.Novell.COM> <3sdti8$7is@news.Belgium.EU.net> <19950624T140535Z@naggum.no> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) erik@naggum.no (Erik Naggum) writes: >[Jacques Deseyne] > >| It is possible to use a non-ASCII concrete syntax (although not in the >| SGML declaration) for your markup. > >this implies that the SGML declaration must be in ASCII, which is _wrong_. >the SGML declaration must only be in some character set that the computer >on which the parser runs understands, known as the "system character set". >Martin Bryan is the source of the _extremely_ unfortunate and unsupported >belief that the SGML declaration must be in ASCII. this belief has been >thoroughly repudiated here in comp.text.sgml previously, and I'm sorry to >see that the SEMA Group still caters to this belief. > Indeed, from a formal point of view, the term "ASCII" should not be used, but rather "ISO 646 IRV". Let me quote both ISO 8879:1986 and Ch. Goldfarb, "The SGML Handbook" : ISO 8879:1986, p. 47 Goldfarb[1990], p. 451 Clause 13 --------- (Production 171) "The reference concrete syntax must be used in the SGML declaration, regardless of the concrete syntax used in the remainder of the document." and : ISO 8879:1986, p. 53 Goldfarb[1990], p. 476 Clause 14 --------- .. "NOTE - The syntax-reference character set of the reference concrete syntax is ISO 646 IRV. That set consists of characters ..." I don't see how you can interpret these passages to conclude that the SGML declaration must be in some (does that mean "whatever" ?) character set that the computer the parser runs on understands. What is meant by "understanding" ? Let me also point out that Martin Bryan is not formally an employee of any of the branches of SEMA Group. We do maintain a good relationship, however. Do we have to include a disclaimer like "opinions expressed are mine, but do not necessarily conflict with my employer's" with every posting ? We would appreciate if Erik Naggum could point at the relevant news items having "thoroughly repudiated" this "belief" (Heresy ? Do we smell the pyres of the Inquisition here ?); they must be somewhere in his archive. >it is possible to use _any_ character set whatsoever for the entire SGML >document. the SGML declaration will specify the character set used. when >an SGML document is translated from one character set to another, the SGML >declaration will certainly aid the process, but it must also be translated >at a more semantic level to reflect the new character set. that is, the >SGML declaration does _not_ serve the purpose of document interchange that >some like to impute to it, however much we would like it to do so. > The SGML declaration will indeed specify the character set used. How is this specification encoded ? I mean, what character set coding can be used to specify which character set coding you are going to use ? You have to agree on one formalism, and it looks like that ISO 646 IRV serves that purpose. Otherwise, you provide arguments to the ODA proponents stating that with SGML, no exchange is possible without agreements between sender and receiver; agreements not included in the exchanged information, that is. Note 1 after production 171 in Clause 13 states that "The SGML declaration is intended for human consumption (in printed form!) as well as for machine processing, as it enables the recipient of a document to determine whether a system can process it "as is", whether character translation or other algorithmic conversion is needed ...". If humans have to consume it in printed form, there is certainly no discussion about character codesets. ------------------------------------------------------ Jacques Deseyne \ SEMA Group belgium Stallestraat 96 B-1180 Brussels Belgium info-markit@sema.be info-sgml@sema.be ------------------------------------------------------ From: gerald@cs.ualberta.ca (Gerald Oskoboiny) Newsgroups: comp.text.sgml Subject: Re: Help using James Clark's SP parser Date: 26 Jun 1995 17:03:03 GMT Organization: University of Alberta Message-ID: <3smp87$oq9@scapa.cs.ualberta.ca> References: \ Reply-To: gerald@cs.ualberta.ca Originator: gerald@gsb014.cs.ualberta.ca jsc@tabitha.atype.com writes: >Is there any documentation of any sort relevant to James Clark's SP SGML >parser? I've been grovelling through the code and examples, but this >could take a while and any documentation at all would be a great help. This package is documented at: http://www.jclark.com/sp.html and Nelson Beebe has some notes on SP at: http://www.math.utah.edu/~beebe/sp-notes.html >I just want something to convert HTML to a tree-like data structure. Do you want to convert an HTML document to a tree-like structure, or the HTML DTD? If it's the DTD, this is done well with the dtd2html package, which is documented at: http://www.oac.uci.edu/indiv/ehood/dtd2html.doc.html and there's a sample tree at: http://www.oac.uci.edu/indiv/ehood/html2.0/DTD-TREE.html Gerald -- Gerald Oskoboiny \ http://ugweb.cs.ualberta.ca/~gerald/ Newsgroups: comp.text.sgml From: dbecke@info.tp.eps0 (David Becke, #?) Subject: Re: DTD to meta-DTD conversion? Message-ID: <1995Jun26.182018.24408@dhnews.dehavilland.ca> Sender: news@dhnews.dehavilland.ca (DHNEWS NEWS Administrator) Reply-To: dbecke@info.tp.eps0 Organization: Bombardier Regional Aircraft Division References: <3s94iv$h9r@ruulch.let.ruu.nl> Date: Mon, 26 Jun 1995 18:20:18 GMT If what you are looking for is a product that will generate a graphic view of your DTD then there are several avcailable for the PC platform; Near & Far by MicroStar Software Inc. SGML Companion by Publishing Development AB christian@pubdev.se DTD Viewer by ZIFtech Computer Systems Inc. From: lalthen@wtcd.DaytonOH.ATTGIS.COM (Lee Althen) Newsgroups: comp.text.sgml Subject: SGML Author, MicroSift Keywords: SGML Author Message-ID: <1814@wtcd.DaytonOH.ATTGIS.COM> Date: 26 Jun 95 18:55:03 GMT Organization: AT\&T Global Information Solutions - Dayton, OH USA Does anyone have any experience building applications with SGML Author? In particular, has anyone done anything with SGML Author using DocBook DTD? Our author community has been searching for a tool to use for authoring in SGML for sometime now. They, unfortunately, believe it must be WYSIWYG and of MS origin. Hence, they are hoping SGML Author will be their answer. I tried posting a similar article (request) last week but evidently the network gremlins digested it rather than let it be posted. I hope this one works and does not turn out to be a duplicate posting. If it's a duplicate, accept my appologies in advance. Lee Lee E. Althen (WHQ-2) Lee.Althen@wtcd.DaytonOH.ATTGIS.COM Information Products Division ...!uunet!ncrcom!wtcp!wtcd!lalthen AT\&T GIS, So Patterson Bldv FAX: (513) 445-2814 Dayton, OH 45479 Tele: (513) 445-3487 -- Lee E. Althen (WHQ-2) Lee.Althen@wtcd.DaytonOH.ATTGIS.COM Information Products Division ...!uunet!ncrcom!wtcp!wtcd!lalthen AT\&T GIS, So Patterson Bldv FAX: (513) 445-2814 Dayton, OH 45479 Tele: (513) 445-3487 From: postmaster@aol.com (Steve Chase) Newsgroups: comp.sys.mac.hardware,comp.sys.mac.hypercard,comp.sys.mac.misc,comp.sys.mac.system,comp.sys.mac.wanted,comp.sys.next.advocacy,comp.sys.next.hardware,comp.sys.next.misc,comp.sys.next.programmer,comp.sys.next.software,comp.sys.next.sysadmin,comp.sys.palmtops,comp.sys.sun.admin,comp.sys.sun.apps,comp.sys.sun.hardware,comp.sys.sun.misc,comp.sys.sun.wanted,comp.sys.tandy,comp.sys.unisys,comp.text.frame,comp.text.sgml,comp.text.tex,comp.theory,comp.unix.admin,comp.unix.aix,comp.unix.misc,comp.unix.osf.osf1,comp.unix.programmer,comp.unix.questions,comp.unix.shell Subject: - FREE PUNK ROCK in alt.binaries.sounds.music Date: Mon, 26 Jun 95 20:48:48 GMT Organization: The SenseMedia Network News - http://sensemedia.net/news - info@sensemedia.net - smn Message-ID: <3sn6cc$g69@Sequoia.picosof.com> X-Newsreader: News Xpress Version 1.0 Beta #3 For the past three months 2-3 punk songs have been posted each and every day to alt.binaries.sounds.music. These are name brand punk bands and not a bunch of shitty generic bands in case you're wondering. All sound files are CD-quality MPEG's. Free MPEG software is available for: IBM PC (DOS), Sun, HP, SGI, PC running Linux, and PC or NeXT running NeXTSTEP. 386DX or better and 16-bit sound card required. Newsgroups: comp.text.sgml,comp.text.tex,fj.comp.text,fr.comp.text.tex From: dichter@netcom.com (Carl Dichter) Subject: Documentation tools for multi-platform? Message-ID: \ Organization: Cross Platform Strategies Date: Mon, 26 Jun 1995 21:06:36 GMT Sender: dichter@netcom.netcom.com We have an article in the next issue about cross platform documentation strategies. We'd like to mention commercial tools that would be helpful to our readers. Perhaps you use, sell, or know of documentation tools that work on multiple platforms, or have features (like conditional text) that are helpful when writing manuals for multiple platforms. We are also interested in add-ons packages for major word processors that help in cross platform use. I am aware of FrameMaker (and FrameBuilder), Word Perfect, MS Word (adapted/emulated on Sun and HP-UX), Nroff, and TEX. What other pertinant products are there? -- Carl Dichter dichter@netcom.com Editor of Cross Platform Strategies. Engineer at Motorola. Co-author (with Mark Pease) of Software Engineering with Perl: Toolsmithing and Prototyping for Better Software -- Sooner. From: streich@odie.austin.wireline.slb.com (Robert Streich) Newsgroups: comp.text.sgml Subject: Re: conloc Date: 26 Jun 1995 23:05:26 GMT Organization: Schlumberger Austin Research Message-ID: \ References: \ In-reply-to: Paul_Hermans@protext.be's message of Sat, 24 Jun 1995 11:45:03 +0100 In article \ Paul_Hermans@protext.be (Paul Hermans) writes: > Are there any tools in the marketplace yet that support the content > location attribute of the all-id attribute list form of the base module of > HyTime ? Don't we wish! The idea of breaking up the entity structure of my document just because I want to reuse pieces of it makes me shudder. Some of the DBMS vendors even offer you that "service," even if you don't own the document that you want to use a piece of. Thoughtful of them isn't it! Yecchhh. There are a couple tools that I think could easily do it, they just need to have more people request it. Pinnacles has a similar concept in their DTD so that seems to be helping bring some pressure to bear. You actually need two pieces to make it work reasonably: conloc and nameloc. Without nameloc, you can't reuse anything outside of the current document. -- Robert Streich streich@austin.sar.slb.com Schlumberger 512-331-3318 (voice) Austin Research 512-331-3760 (fax) -- Robert Streich streich@austin.sar.slb.com Schlumberger 512-331-3318 (voice) Austin Research 512-331-3760 (fax) From: dwm@shell.portal.com (Morris) Newsgroups: comp.text.sgml Subject: Re: Browser for HTM or SGML files on a LOCAL PC Date: 26 Jun 1995 16:30:52 -0700 Organization: Portal Communications (shell) Message-ID: <3snfvc$egd@jobe.shell.portal.com> References: <3ropjq$hmv@imp.demon.co.uk> <3s914k$2cc@newsgate.dircon.co.uk> Several of the 'common' PC / Windows browsers may be used to view local files w/o a network attached. For netscape for example, they offer a dummy winsock.dll which must appear in the path. Some others just work. Of course, this is HTML only, but there are no conformance issues beyond those of the net which already exist. Dave Morris From: Darrin Bogue \ Newsgroups: comp.text.sgml Subject: Trademark Entity? Date: 27 Jun 1995 00:17:32 GMT Organization: zNET Message-ID: <3snims$fiv@news.znet.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) Hi! I was wondering if there is a code for trademark... I can't find it anywhere. Thanks. please respond to: daver@digpath.com From: Betty Harvey \ Newsgroups: comp.text.sgml Subject: Re: DTD analysis tool Date: 27 Jun 1995 02:32:33 GMT Organization: Electronic Commerce Connection Message-ID: <3snqk1$8u3@news.cais.com> References: <3sfvvn$v2@nnrp1.primenet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.2b1 (Windows; I; 16bit) quixote@Primenet.Com (Donald A. Hosek) wrote: >DOes anyone have any tools for analyzing DTDs? A couple things that would >be handy would be > - a tool to list the minimum required document frame given a DTD > - a tool that takes the above one step further by indicating optional > tags, stopping when it detects a recursion possibility (this would > be a bit tougher to represent the output of, although even the first > will require some nastiness). > >Any suggestions? > Look at Near and Far. You aren't going to find any 'nastier' DTD's than some developed for DoD. Near and Far allows you to really understand the DTD visually. I have used it quite a bit for a 'sanity check'. I have used Near and Far output in meetings to demonstrate the complexity of SGML and DoD DTD's. It is interesting to watch the faces of people when you take printed output of Near and Far representation of a DTD and spread it out on a conference table and it falls off both ends! I really haven't found a better tool. However, Earl Hood's DTD2HTML is also a very good tool for looking and learning about DTD's. His DTD2TREE program is also very useful for viewing a DTD visually. <><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><> Betty Harvey |The Hard Travelers Cystic Fibrosis Benefit harvey@eccnet.com | Concert with Kenny Rogers | http://www.eccnet.com/cf/ <><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><> From: Neil McNaughton <100045.3625@CompuServe.COM> Newsgroups: comp.text.sgml Subject: SGML FAQ Date: 27 Jun 1995 06:33:36 GMT Organization: TDR - The Data Room Message-ID: <3so8o0$bc8$1@mhadf.production.compuserve.com> I posted a request for info on SGML a while back vis 1.)Is there a FAQ on SGML 2.)What are the main differences between SGML and HTML, is the second a subset of the first, and what does SGML get you that HTML doesn't. So far, no response. So I now have a few more questions vis bis.. 1.) Is comp.text.sgml some sort of a secret society? 2.) Are the above questions too hard or too easy? Hope someone can help. Neil -- Neil McNaughton 100045.3625@compuserve.com From: margit@tordalk.ims.uni-stuttgart.de (Margit Hippelein) Newsgroups: comp.text.sgml Subject: FAQ Date: 27 Jun 1995 07:48:51 GMT Organization: IMS, University of Stuttgart, Germany Message-ID: \ Reply-To: margit@ims.uni-stuttgart.de Is there a FAQ for this group? How can I get it? Thanks, Margit -- Newsgroups: comp.text.sgml From: Steve Pepper \ Subject: Re: SGML Author, MicroSift Message-ID: <1995Jun27.075542.27005@falch.no> Keywords: SGML Author Organization: Falch Infotek, Oslo, Norway References: <1814@wtcd.DaytonOH.ATTGIS.COM> Date: Tue, 27 Jun 1995 07:55:42 GMT lalthen@wtcd.DaytonOH.ATTGIS.COM (Lee Althen) writes: >Does anyone have any experience building applications with >SGML Author? In particular, has anyone done anything with >SGML Author using DocBook DTD? Our author community has been >searching for a tool to use for authoring in SGML for sometime >now. They, unfortunately, believe it must be WYSIWYG and of MS >origin. Hence, they are hoping SGML Author will be their answer. I haven't used SGML Author myself, but I would suggest you take a good look at Microstar's Near & Far Author for Microsoft Word before opting for SGML Author. From what I have seen, SGML Author is more a conversion system, whereas Near & Far Author turns MS Word into a true SGML editor. Another alternative is Tag Wizard from NICE technologies. More details from: cade@microstar.com nicetech@netcom.com Regards, Steve -- pepper@falch.no -------------------------------------------------------------- falch infotek a.s, postboks 130 kalbakken, n-0902 oslo, norway tel +47 2290 2733 fax +47 2290 2599 http://www.falch.no/ From: martyn davison \ Newsgroups: comp.text.sgml Subject: Re: SGML Author, MicroSift Date: Tue, 27 Jun 1995 08:22:34 GMT Organization: OMI Logistics Ltd Message-ID: <271743554wnr@omimjd.demon.co.uk> References: <1814@wtcd.DaytonOH.ATTGIS.COM> Reply-To: martyn@omimjd.demon.co.uk X-NNTP-Posting-Host: omimjd.demon.co.uk X-Broken-Date: Tuesday, Jun 27, 1995 08.22.34 X-Newsreader: Newswin Alpha 0.7 In article: <1814@wtcd.DaytonOH.ATTGIS.COM> lalthen@wtcd.DaytonOH.ATTGIS.COM (Lee Althen) writes: > > Does anyone have any experience building applications with > SGML Author? In particular, has anyone done anything with > SGML Author using DocBook DTD? Our author community has been > searching for a tool to use for authoring in SGML for sometime > now. They, unfortunately, believe it must be WYSIWYG and of MS > origin. Hence, they are hoping SGML Author will be their answer. >> > -- > Lee E. Althen (WHQ-2) Lee.Althen@wtcd.DaytonOH.ATTGIS.COM > Information Products Division ..!uunet!ncrcom!wtcp!wtcd!lalthen > AT\&T GIS, So Patterson Bldv FAX: (513) 445-2814 > Dayton, OH 45479 Tele: (513) 445-3487 > Noted your request about SGML Author from Microsoft. We had it installed but it screwed up W4W. At best it is an SGML converter. It is far better to work in native SGML. We use the GriF product which is WYSIWYG. Ok it needs configuring but it is very powerful indeed. The built in functionality is extremely good and with the GATE application its base functionality is almost limitless. The down side is that it is more expensive than the Microsoft option but it is _true_ SGML and forces the user to adhere to the DTD. -- Martyn Davison OMI Logistics Ltd UK. Email martyn@omimjd.demon.co.uk From: gelpo@news.jyu.fi (DigPub) Newsgroups: comp.text.sgml Subject: Docbook DTD -> LaTeX/HTML Date: 27 Jun 1995 09:43:24 GMT Organization: University of Jyvaskyla, Finland Message-ID: <3sojrs$jbb@mordred.cc.jyu.fi> X-Newsreader: TIN [version 1.2 PL2] Hello Does somebody know if there is any complete conversion program from Docbook DTD to LaTeX or/and HTML? Thanks in advance. --------------------------------------------------------------------------- Riitta Kunnari of. : MaE 112 Virpi Lyytikainen tel : 941-603050 Tiina Pietinen email: gelpo@jytko.jyu.fi University of Jyvaskyla, PL 35, 40351 JYVASKYLA, FINLAND ------------------------------------------------------------------------- Newsgroups: comp.text.sgml From: jta_swets@hgl.signaal.nl (Swets J.T.A.) Subject: Re: Implementation of SGML Message-ID: <1995Jun27.100543.8483@hgl.signaal.nl> Sender: usenet@hgl.signaal.nl (USENET News System) Organization: Hollandse Signaal apparaten BV X-Newsreader: TIN [version 1.2 PL2] References: <3seh03$c4@ixnews2.ix.netcom.com> <19950624T135246Z@naggum.no> Date: Tue, 27 Jun 1995 10:05:43 GMT Erik Naggum (erik@naggum.no) wrote: : my impression is that adoption of SGML is moving from the publishing end : closer to the source of the information. I have gone so far as to say that : SGML is nearly worthless if used only for publishing. I think Erik means that adoption of SGML more stresses the aspect of information management (if the DTDs are content-based, i.e. stress on WHAT information is to be entered, enforcing the content), but what about the profit of - application and platform independence (no longer conversions between proprietary formats), - consistence of style (as presentation semantics will be added seperately). These have also some value. Can you elaborate on the statement that SGML is nearly worthless if used only for publishing? Regards, Jan Swets =============================== Unclassified ============================== J.T.A. Swets Dept. DSA-PIM email: jta_swets@hgl.signaal.nl Research, Development & Technology tel: (+31)-(0)74-482539 (work) Document Technology Consultant Fax: (+31)-(0)74-484009 -------------------------------- Hollandse Signaalapparaten BV PO Box 42, 7550 GD Hengelo The Netherlands. =============================== Unclassified ============================== From: Daniel Waz <76633.3636@CompuServe.COM> Newsgroups: comp.text.sgml Subject: Opinions on the Rainbow DTD Date: 27 Jun 1995 13:10:43 GMT Organization: Cardinals Soccer Club, Bowie Soccer United Message-ID: <3sp00k$ema$1@mhadf.production.compuserve.com> [Daniel Waz] I would like your opinions on the Rainbow DTD, both the good, but more importantly the areas of weakness. How would you rank it as a conversion DTD? Are there better ones available? Thanks. -- Daniel Waz From: i07m@p101.informatik.Uni-Bremen.DE (Kai Hofmann) Newsgroups: comp.text.sgml Subject: Re: FAQ Date: 27 Jun 1995 13:14:09 GMT Organization: Universitaet Bremen Message-ID: <3sp071$l4n@p235.informatik.uni-Bremen.de> References: \ Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit In article \, margit@tordalk.ims.uni-stuttgart.de (Margit Hippelein) writes: |> |> Is there a FAQ for this group? How can I get it? |> |> Thanks, |> |> Margit |> -- Try: ftp://sgml1.ex.ac.uk/pub/SGML -- Kai Hofmann EMail: i07m@zfn.uni-bremen.de FB 3, Informatik i07m@informatik.uni-bremen.de Univ. Bremen IRC: PowerStat From: djones@imagen.net (Dwight Jones) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 27 Jun 1995 15:06:06 GMT Organization: Imagen Communications Inc. Message-ID: <3sp6ou$6qd@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> \ <3si07f$dc5@googol.bctel.net> <19950625T105426Z@naggum.no> Mime-Version: 1.0 X-Newsreader: WinVN 0.99.3 result would be a cleanup of the bandwidth-wasting image-orientedness >of the WWW. anyway, given this climate, aggressive source compression of >the material doesn't seem like a very good idea. There can be no argument against the value of compression. > >on the topic of compression, I find the DOS/Windows-world's preponderance >towards non-layered solutions because of the lack of multitasking on the >host operating system to be alarming. You see directory-level compression in NT 3.51, but granted that is not a client OS. >now, here's an idea: the document as stored on either end will be in an >uncompressed, general format. on the wire, it will be compressed by the >protocol engine if requested by the receiver. >this idea stands in sharp contrast to the design that I see in RTC which is >prematurely optimized for a Windows PC on the other end of a 28,800 bps >dialup link at the cost of machines with full Internet connection in the >multimegabit range, at the cost of machines that don't run a particular >"operating system", etc. making the compression a "feature" of the format >is just plain wrong, When an RTF file with images is 750K, and the same file compressed is 20K, it can't be wrong. The previous file isn't even real for any purpose in communications, unless the Dean set you up with a personal T1. Having said that, I do like your idea very much that the uncompressed RTF file could be posted, except that we could post it without images. That would indeed be a very small file, like an HTMl file without graphics. An excellent idea. We find that most wordprocessors can read other-format RTF as long as there are no images very readily, so we could at trade some attractive documents this way. It is Imagen's hope that with progressive access to growing resources, that we can indeed support most platforms, and aid the standardization of RTF by reliably compressing all variants. That could unify the format somewhat and provide a lingue franca that is fast and accessible to all. From: bducharme@riatax.com (Bob DuCharme) Newsgroups: comp.text.sgml Subject: Re: Small TEI sample ? Date: Tue, 27 Jun 1995 15:08:53 GMT Organization: Thomson Professional Publishing Message-ID: <3sp6ve$ivl@news.tpp.com> References: <3skkev$8up@geraldo.cc.utexas.edu> X-Newsreader: Forte Free Agent v0.55 trujillo@mail.utexas.edu (Juan A. Trujillo) wrote: >Does anyone know where I might find a relatively small TEI-encoded text to >pick apart? The Oxford Text Archive's ftp site at ota.ox.ac.uk has a lot of TEI-encoded stuff. When I needed some small yet interesting instances, I took the first couple books of Milton's "Paradise Lost" and the first couple chapters of "Huckleberry Finn." Bob DuCharme | "The elements be kind to thee, and make bobducharme@acm.org | Thy spirits all of comfort!" | Anthony and Cleopatra, III ii From: djones@imagen.net (Dwight Jones) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 27 Jun 1995 15:15:37 GMT Organization: Imagen Communications Inc. Message-ID: <3sp7ap$6qd@googol.bctel.net> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> \ <3si07f$dc5@googol.bctel.net> <3sksu1$h19@scapa.cs.ualberta.ca> Mime-Version: 1.0 X-Newsreader: WinVN 0.99.3 >>4) HTML compatible. Is PDF? > >Define "compatible". Can I view RTC files with Lynx? >Do RTC files validate against the HTML specification? Admittedly, that is a bit of a stretch. What we mean is that RTC supports Hyperlinks, hot graphics etc. with the same syntax, and will also directly execute those links without the browser via winsockets. . > >>There are others, but you will find that a lot of people value a cleaner >>implementation than PDF or the floating HTML version X.X. > >What do you mean by "floating HTML version X.X"? HTML 2.0 is stable, and >widely implemented. I can view HTML 2.0 compliant files on a wider variety >of platforms than any other document format, except maybe plain text. >Gerald Oskoboiny \ http://ugweb.cs.ualberta.ca/~gerald/ It seems that Netscape, with the majority of the market, is implementing motion video capability, and onboard PDF. Is PDF part of the HTML X.X spec? It's bandwidth boorishness and needs ISDN or faster to make any sense at all. What Netscape wants they seem to get. It's interesting that the video that they chose to support is not the same that Microsoft will include in their Windows operating systems. We'll see how Netscape holds that position over time. Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 27 Jun 1995 11:27:55 -0400 From: "Claude L. Bullard" \ Message-ID: <9506271527.AA25261@source.asset.com> Subject: CORBA Can someone provide a network address or Web site for information on the CORBA specifications (2.0?)? Thanks in advance. Len Bullard Loral Defense Systems - East From: grmeyer@mcs.com (Gordon R. Meyer) Newsgroups: comp.text.sgml Subject: Re: PDF and SGML Date: Tue, 27 Jun 1995 10:34:21 -0600 Organization: Our Lady of Chaos and Silicon, inc. Message-ID: \ References: \ \ <3skjhn$ag@news.cais.com> In article <3skjhn$ag@news.cais.com>, Betty Harvey \ wrote: > I have also noticed, from the PDF documents that I have had the > opportunity to view, that the visual quality of the document > is poor. Visually they remind me of the way documents looked > with CGA. > Alot of this can be fixed by embedding fonts in the PDF, but in general I agree that the onscreen fidelity isn't what it should be. However, the printed output is superb, and by zooming in I can compensate for the initial fidelity issue. Of the two mark-up language online solutions I've seen first-hand (Bookmanager and Dynatext) I found the font-substitution they did to be much more annoying. -- Gordon Meyer \ Windows 95 = Mac 89 Newsgroups: comp.text.sgml Approved: erik@naggum.no Message-ID: <9506271942.AA1501@notes.microstar.com> From: Ray Luoma \ Date: 27 Jun 1995 12:39:05 -0400 References: <3shsn2$95@mn5.swip.net> Subject: Re: SGML with Ventura? >I wonder if someone knows if Corel will support SGML editor with >the Corel Ventura publishing? The upcoming Version 6.0 of Corel Ventura will also use technology licensed from Microstar Software in Ottawa, the makers of NEAR & FAR. NEAR & FAR is a graphical tree-structure DTD editor. Get in touch if you want more information. Ray /\\ /\\ Ray Luoma - rluoma@microstar.com \\/ \\/ Computer /\\ /\\ Aided Microstar Software Ltd. Phone: +1 613 727-5696 \\/ \\/ Document 34 Colonnade Road North Fax: +1 613 727-9491 /\\ /\\ Engineering Nepean Ontario Can/US: 1 800 267-9975 \\/ \\/ CANADA K2E 7J6 Info: cade@microstar.com From: grmeyer@mcs.com (Gordon R. Meyer) Newsgroups: comp.text.sgml Subject: Re: BookMaster to SGML Conversion Date: Tue, 27 Jun 1995 10:39:31 -0600 Organization: Our Lady of Chaos and Silicon, inc. Message-ID: \ References: <3smfv2$ru2@newsbf02.news.aol.com> In article <3smfv2$ru2@newsbf02.news.aol.com>, datkins698@aol.com (DAtkins698) wrote: > Does anyone know of any publicly available tools to convert BookMaster > files to SGML? Any help would be greatly appreciated. I really need this > information ASAP. I doubt you'll find exactly this, the BookMaster file is a compressed, binary file. You might be better off looking for a GML to SGML conversion (in other words, use the *source* files in the BookMaster process, not the output.) -- Gordon Meyer \ Windows 95 = Mac 89 From: clollin@bgnet.bgsu.edu (Christopher Lollini) Newsgroups: comp.text.sgml Subject: DSSSL Date: 27 Jun 1995 17:01:31 GMT Organization: Bowling Green State University Message-ID: \ I'm having a difficult time finding information on DSSSL. Can anyone point me in the right direction? From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: Re: MS SGML Author Application Building Experience Date: 27 Jun 1995 17:14:30 GMT Organization: SEMA Group Belgium Message-ID: <3spe9m$6hf@news.Belgium.EU.net> References: <1813@wtcd.DaytonOH.ATTGIS.COM> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) lalthen@wtcd.DaytonOH.ATTGIS.COM (Lee Althen) wrote: >I have recently received a copy of Micosoft SGML Author. After >experimenting with it and reading the rather sparse System Administrator >Guide (booklet), I am of the opinion that a rather serious development >effort is required in order to have a useful tool for authoring with anything >but the simplest of DTDs. The sample partial implementation of the CALS >DTD included with the MS SGML Author, I believe, is an example of the >development required. Does anyone have experience building / using this >tool? Has anyone implemented the tool for use with DocBook? Where can I >find some expertise to help with building an application? > We have set up a demo application for the European Council last February. With a quite simple DTD, a rudimentary conversion environment took almost one week to set up. In that test, we didn't even bother about constructing an .INI file with the combinatorial number of contextual information items in order to enable the "Next Legal Style" listbox. Indeed, we also estimate that a quite serious effort is needed to make authors work with this tool; too much work, in our opinion. Take maybe a look at Microstar's Near & Far Author which is in beta-test and builds most mappings itself; TagWizard is also a good alternative (in case you want to stay with WinWord). Maybe people at Avalanche (now Interleaf, Waltham, MA.) can point at "real-life" experience; they are behind the technology (there's no coincidence being a lot of AV*.DLL files in SGML Author). ------------------------------------------------------ Jacques Deseyne \ Sema Group Belgium Stallestraat 96 B-1180 Brussels Belgium info-markit@sema.be info-sgml@sema.be ------------------------------------------------------ From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: Re: SGML with Ventura? Date: 27 Jun 1995 17:20:26 GMT Organization: SEMA Group Belgium Message-ID: <3spekq$6hf@news.Belgium.EU.net> References: <3shsn2$95@mn5.swip.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) Mikael Sandberg \ wrote: >I wonder if someone knows if Corel will support SGML editor with >the Corel Ventura publishing? >Today I use Ventura 4.2 but soon I have to start with ADEPT Editor with >SGML. >Since I feel more fore traditional text editing like Ventura, I wish in >the future Ventura will be able to handle SGML as ADEPT Editor do. >Regards, Mikael > With the current version CorelVentura 5.0, you can import (and export ?) SGML documents. This is done through Zandar's TagWrite, which is, however, not a true SGML tool, only a tagging/untagging application without a parser (the "template" designer has to build in by hand the complete DTD logic; this can be a cumbersome task). The future version 6.0 will incorporate technology from InContext and from Microstar, as I have read somewhere (I thought it was in the May issue of \). Beta-testing would start in september. ------------------------------------------------------ Jacques Deseyne \ Sema Group Belgium Stallestraat 96 B-1180 Brussels Belgium info-markit@sema.be info-sgml@sema.be ------------------------------------------------------ From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: Re: conloc Date: 27 Jun 1995 17:23:09 GMT Organization: SEMA Group Belgium Message-ID: <3spept$6hf@news.Belgium.EU.net> References: \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) Paul_Hermans@protext.be (Paul Hermans) wrote: >Are there any tools in the marketplace yet that support the content >location attribute of the all-id attribute list form of the base module of >HyTime ? > The HyTime engine built in the MIPS project supports conloc. We don't see how a HyTime application could possibly NOT support conloc. The engine is not really in the marketplace YET. It was demonstrated at SGML Europe '95 in Gmunden. ------------------------------------------------------ Jacques Deseyne \ Sema Group Belgium Stallestraat 96 B-1180 Brussels Belgium info-markit@sema.be info-sgml@sema.be ------------------------------------------------------ Newsgroups: comp.text.sgml From: ak116@FreeNet.Carleton.CA (Bonnie Robinson) Subject: _Practical SGML_ Message-ID: \ Sender: ak116@freenet.carleton.ca (Bonnie Robinson) Reply-To: ak116@FreeNet.Carleton.CA (Bonnie Robinson) Organization: The National Capital FreeNet Date: Tue, 27 Jun 1995 18:32:05 GMT I am in the process of learning SGML and would like a copy of the second edition of _Practical SGML_. No bookstores in Toronto or Ottawa carry it, and when I tried to order it, I learned that it is not available from the publisher until October! If anyone has a copy that they don't want, or can track it down for me, I would be eternally grateful. Bonnie Robinson ak116@freenet.carleton.ca From: truly@lunemere.com (Truly Donovan) Newsgroups: comp.text.sgml Subject: Re: BookMaster to SGML Conversion Date: Tue, 27 Jun 1995 19:05:01 Organization: La Lunemere Message-ID: \ References: <3smfv2$ru2@newsbf02.news.aol.com> \ X-Newsreader: Trumpet for Windows [Version 1.0 Rev A] In article \ grmeyer@mcs.com (Gordon R. Meyer) writes: >In article <3smfv2$ru2@newsbf02.news.aol.com>, datkins698@aol.com >(DAtkins698) wrote: >> Does anyone know of any publicly available tools to convert BookMaster >> files to SGML? Any help would be greatly appreciated. I really need this >> information ASAP. >I doubt you'll find exactly this, the BookMaster file is a compressed, >binary file. You might be better off looking for a GML to SGML conversion >(in other words, use the *source* files in the BookMaster process, not the >output.) I think you've confused BookMaster with BookManager. BookMaster files are source files. Truly Donovan From: avterry@PROBLEM_WITH_INEWS_GATEWAY_FILE (Terry Brien) Newsgroups: comp.text.sgml Subject: Join SGML Newswire! Date: 27 Jun 1995 21:04:03 GMT Organization: SuperNet Inc. (303)-296-8202 Denver Colorado Message-ID: <3spro3$536@news-4.csn.net> X-Newsreader: TIN [version 1.2 PL2] The listserv 'sgmlinfo' is alive and well and sponsored by Interleaf Avalanche. Topics range from DTDs, CALS table models, discussions, product reviews, SGML articles, and upcoming events. Occasionally posts from this group may appear. Comments, questions, and suggestions are also welcomed by the newswire manager. Thanks, Terry To join, send email to 'sgmlinfo@avalanche.com' with the body SUBSCRIBE 'yourname'. From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: SGML and double-byte character sets Date: 27 Jun 1995 21:52:18 GMT Organization: ACM Network Services Message-ID: <3spuii$f3u@hopper.acm.org> References: <3scjuk$grb@grok.provo.Novell.COM> <3sdti8$7is@news.Belgium.EU.net> <19950624T140535Z@naggum.no>,<3smo0d$3u0@news.Belgium.EU.net> Reply-To: davep@ACM.ORG In article <3smo0d$3u0@news.Belgium.EU.net>, Jacques Deseyne \ writes: >erik@naggum.no (Erik Naggum) writes: > >>[Jacques Deseyne] >> >>| It is possible to use a non-ASCII concrete syntax (although not in the >>| SGML declaration) for your markup. >> >>this implies that the SGML declaration must be in ASCII, which is _wrong_. >>the SGML declaration must only be in some character set that the computer >>on which the parser runs understands, known as the "system character set". >>Martin Bryan is the source of the _extremely_ unfortunate and unsupported >>belief that the SGML declaration must be in ASCII. this belief has been >>thoroughly repudiated here in comp.text.sgml previously, and I'm sorry to >>see that the SEMA Group still caters to this belief. >Indeed, from a formal point of view, the term "ASCII" should not be used, >but rather "ISO 646 IRV". That's a red herring. Neither ASCII nor ISO 646 are mandated as the document character set. What is mandated is that the entire document, from the initial "\"The reference concrete syntax must be used in the SGML declaration, regardless >of the concrete syntax used in the remainder of the document." >"NOTE - The syntax-reference character set of the reference concrete syntax is >ISO 646 IRV. That set consists of characters ..." > >I don't see how you can interpret these passages to conclude that the SGML >declaration must be in some (does that mean "whatever" ?) character set that >the computer the parser runs on understands. Once again, class, and pay attention this time: A character (at least for ISO 8879) is an abstraction; a character set is an association of bit-strings with characters, and we associate small nonnegative integers with the bit-strings. There are two character sets involved with SGML documents: the document character set, in which the document is encoded, and the syntax character set, whose only function is to associate small integers with characters so that the concrete syntax can talk about them conveniently. The two are kept separate so that one can, for example, use the reference concrete syntax even though the document character set is EBCDIC--but the SGML declaration that references that syntax is written in EBCDIC. Using ISO 646 as the syntax-reference character set simply means you can still refer to a carriage return as 13 and a line feed as 10, etc., *when describing the concrete syntax*. >Note 1 after production 171 in Clause 13 states that "The SGML declaration >is intended for human consumption (in printed form!) as well as for machine >processing, as it enables the recipient of a document to determine whether a >system can process it "as is", whether character translation or other algorithmic >conversion is needed ...". Precisely--you have to be told the character set before you can read the SGML declaration. We can wish it were otherwise, but it ain't. Hope this clears up some of the confusion. Dave Peterson SGMLWorks! davep@acm.org From: ecampbel@iol.ie (Eoin Campbell) Newsgroups: comp.text.sgml Subject: Re: BookMaster to SGML Conversion Date: 27 Jun 1995 21:54:04 GMT Organization: Ireland On-Line Message-ID: <3spuls$8qj@barnacle.iol.ie> References: <3smfv2$ru2@newsbf02.news.aol.com> X-Newsreader: WinVN 0.92.6+ In article <3smfv2$ru2@newsbf02.news.aol.com>, datkins698@aol.com (DAtkins698) says: > >Does anyone know of any publicly available tools to convert BookMaster >files to SGML? Any help would be greatly appreciated. I really need this >information ASAP. > >Thanks, >Diana Atkins >DAtkins698 I think a company called Softline International do this. They also to BookMaster to Frame filters, and if you find the guide to available Frame add-ons, they are mentioned in it. If you can't find them, let me know. From: fritz@hc.ti.com (Fritz Whittington) Newsgroups: comp.text.sgml Subject: Re: SGML Author, MicroSift Date: 27 Jun 1995 22:04:11 GMT Organization: TI Semiconductor Process and Device Center Message-ID: <3spv8r$mac@cauldron.spdc.ti.com> References: <1814@wtcd.DaytonOH.ATTGIS.COM> <271743554wnr@omimjd.demon.co.uk> X-Newsreader: NN version 6.5.0 #8 (NOV) martyn davison \ writes: >(Lee Althen) writes: >> Does anyone have any experience building applications with >> SGML Author? In particular, has anyone done anything with >> SGML Author using DocBook DTD? Our author community has been >> searching for a tool to use for authoring in SGML for sometime >> now. They, unfortunately, believe it must be WYSIWYG and of MS >> origin. Hence, they are hoping SGML Author will be their answer. >Noted your request about SGML Author from Microsoft. We had it installed but >it screwed up W4W. At best it is an SGML converter. It is far better to work >in native SGML. >We use the GriF product which is WYSIWYG. Ok it needs configuring but it is >very powerful indeed. The built in functionality is extremely good and with >the GATE application its base functionality is almost limitless. The down >side is that it is more expensive than the Microsoft option but it is _true_ >SGML and forces the user to adhere to the DTD. I don't want to get into a "which product is best" war here, but I feel I should point out a few things. Unless I'm mistaken, SGML Author was first released marked *only* for the US version of Word. If you in the UK applied it to a UK version of Word, I suspect that it would mess it up. (If you did get a UK version of SGML Author, then I don't know why it didn't work.) Secondly, "At best it is an SGML converter." YES! It allows you to actually print a document in a form pleasant to the majority of people, while maintaining an SGML form that may be useful for other purposes. I have installed it and used it. It works well for me. I have not tried to do so, but you can supply your own DTD, define your own Word styles for each element, and provide the rules for mapping each element in SGML to a style or format in Word. You may start with an SGML document, and import it to Word for viewing and printing. Or, you may start with a Word document, apply the appropriate template, apply styles to selected parts of the text, then save an SGML version. As you do so, the SGML will be syntax-checked against the DTD, and an error file created which is your original Word document with the errors flagged by explanatory text. You can always, if you wish, take the SGML text and work on it with a plain text editor, or an advanced SGML-aware editor. It comes with a small toy "sample" DTD and mapping file, and with two real DTDs: CALS and ATASB. Both are widely used in the US, and represent the most common cases where SGML documentation is required by government contract or regulations. Unfortunately, neither is really all that great for writing books. I suspect that either MS or some ISVs intend to fill this gap by selling add-on DTDs and styles and mapping file packages. -- Fritz Whittington Texas Instruments, P.O. Box 655474, MS 446 Dallas, TX 75265 Shipping address: 13510 North Central Expressway, MS 446 Dallas, TX 75243 fritz@ti.com Office: +1 214 995 0397 FAX: +1 214 995 6194 Since I am not an official TI spokesperson, these opinions contain no spokes. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: FAQ Date: 27 Jun 1995 22:48:14 GMT Organization: Naggum Software; +47 2295 0313 Expires: 11 Jul 1995 22:48:14 GMT Message-ID: <19950627T224814Z@naggum.no> References: \ <3sp071$l4n@p235.informatik.uni-Bremen.de> [Margit Hippelein] | Is there a FAQ for this group? How can I get it? [Kai Hofmann] | Try: ftp://sgml1.ex.ac.uk/pub/SGML or ftp.ifi.uio.no:/pub/SGML. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 28 Jun 1995 00:11:03 GMT Organization: Naggum Software; +47 2295 0313 Expires: 12 Jul 1995 00:11:03 GMT Message-ID: <19950628T001103Z@naggum.no> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> \ <3si07f$dc5@googol.bctel.net> <19950625T105426Z@naggum.no> <3sp6ou$6qd@googol.bctel.net> [Dwight Jones] | There can be no argument against the value of compression. oh? compression is an answer to a specific question, namely the ratio of available CPU power to available bandwidth that is vastly in favor of the CPU power. this is not a universal problem, but rather a problem specific to a particular subcommunity on the Internet _today_. my problem with your attitude is that I don't want you to make tradeoffs for me by tying me to a format that forbids smart allocation of my resources. | When an RTF file with images is 750K, and the same file compressed is | 20K, it can't be wrong. The previous file isn't even real for any | purpose in communications, unless the Dean set you up with a personal | T1. the Internet technical community have been discussing the problems of gigabit networks for some time, now. T1 is _slow_, mate. a 100 Mbps FDDI ring is being saturated somewhere, right now. obviously, you don't want the RTC files to be used inside a high-speed (> 30Mbps) limited area network (LAN, campus network, MAN, whatever), only for PC's sitting behind feeble dial-up links. I'm on principle strongly opposed to fragmentation of a user community, especially one knit together by so much good will and solid technology and foresight as the Internet. | Having said that, I do like your idea very much that the uncompressed | RTF file could be posted, except that we could post it without images. the desirability of compression is a function of the CPU/bandwidth ratio. you should realize that in very many cases, the CPU/bandwidth ratio makes compression completely insane. it's sometimes faster to send an image across a very high-speed wire as full 24-bit ultra-high-resolution bitmaps than to force the receiving display engine to decode a bit-scared format. I mean, look at JPEG, whose _decoding_ is so complex it requires several times more memory than the resulting image, and can take minutes of CPU time on a large high-resolution image on high-end machines. sure, it saves a lot of bits, but you have to calculate how many bits you can move over the same wire in the time it takes to decode the compressed format. if you find that you can move more bits across the wire in that time than you save by shipping a compressed format that must be decoded, it is _stupid_ to use a compressed format. the modem standards developers got it right: they allowed people on slow links to compress the data there and then, at the cost of stuffing the modems with more CPU-power than your average computer. of course, this compression is probably inferior to what we get from special-purpose compression that changes for every application (and that's how I see RTC). however, achieving 64kbps transfer rate over a 14.4kbps link isn't so bad for a general- purpose compression algorithm. I'd surmise that you could get 128kbps over those 28.8kbps links, and indeed, that's what some of those PC's and modems can actually do. (probably 115kbps, though.) if you use ISDN, you can beef up your connection to get 128kbps _without_ any compression, too, and _then_ think about what you could get through that "telephone" wire! (note: this is technology available _today_, _cheap_.) | It is Imagen's hope that with progressive access to growing resources, | that we can indeed support most platforms, and aid the standardization | of RTF by reliably compressing all variants. That could unify the | format somewhat and provide a lingue franca that is fast and accessible | to all. I for one don't want it. I do understand that you're trying to solve an acute problem that you have been smart enough to see and also to find and build a solution for. these are good qualities in a market full of people with mostly problem-creating "solutions". however, the solution is limited in scope, and also _limiting_, because if it becomes popular, it will also become irrelevant when the cable TV providers discover their fat "clues receivable" accounts and start to use those bandwidth-wasting coax cables they have strung all over the country for something more than uncompressed boobtubesoup. and when it becomes irrelevant, we will have a choice: lose the old information that people put up on the Net, or convert it to the _next_ limited and limiting solution. we don't have to choose between just those two options, and SGML is trying to help us all realize that we have more options, by being a little more abstract, a little more willing to invest in the future, and a little more willing to resist the temptation to take the path of least resistance and immediate gratification. I know, I'm preaching. "narrow path" and all that boring stuff. doesn't mean it's wrong, just that I'm being a little impatient with going through the same old stuff every other quarter. not your fault -- the SGML crowd isn't exactly publishing their ideas and solutions where the PC crowds are reading about the next hyped-up non-solution to their many non-problems. "the billigates constant: a cosmic constant, the value of which is kept a closely guarded company secret by those who have discovered it, that states how many times a given idea must be reinvented before it can be considered fully invented." (also an Original Naggum, 1995 vintage.) #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: John Lamp \ Newsgroups: comp.text.sgml Subject: Re: DSSSL Date: 28 Jun 1995 02:11:24 GMT Organization: University of Tasmania, Department of Computer Science Message-ID: <3sqdoc$ga4@franklin.cc.utas.edu.au> References: \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Macintosh; I; 68K) X-URL: news:clollin-2706951302560001@m253-119.bgsu.edu clollin@bgnet.bgsu.edu (Christopher Lollini) wrote: >I'm having a difficult time finding information on DSSSL. Can anyone point >me in the right direction? I have links to the US and Australian DSSSL and CALS projects on http://lamp.cs.utas.edu.au/net.html Cheers John -- _--_|\\ John Lamp, originating in Hobart, Tasmania / \\ Phone: 002 20 2375 - Fax: 002 20 2913 \\_.--._/ email: John.Lamp@cs.utas.edu.au v <--<< http://lamp.cs.utas.edu.au/jw_lamp.html From: melhaven@mail.rcinet.com Newsgroups: Lee.Althen@daytonoh.attgis.com,comp.text.sgml Subject: Re: MS SGML Author Application Building Experience Date: 28 Jun 1995 03:57:32 GMT Organization: Dayton Microcomputer Association; Dayton, Ohio, USA Message-ID: <3sqjvc$g2f@sally.dma.org> References: <1813@wtcd.DaytonOH.ATTGIS.COM> Reply-To: melhaven@mail.rcinet.com X-Newsreader: IBM NewsReader/2 v1.09 In <1813@wtcd.DaytonOH.ATTGIS.COM>, lalthen@wtcd.DaytonOH.ATTGIS.COM (Lee Althen) writes: >I have recently received a copy of Micosoft SGML Author. After >experimenting with it and reading the rather sparse System Administrator >Guide (booklet), I am of the opinion that a rather serious development >effort is required in order to have a useful tool for authoring with anything >but the simplest of DTDs. The sample partial implementation of the CALS >DTD included with the MS SGML Author, I believe, is an example of the >development required. Does anyone have experience building / using this >tool? Has anyone implemented the tool for use with DocBook? Where can I >find some expertise to help with building an application? > >Thanks in advance, > >Lee Althen > >Lee E. Althen (WHQ-2) Lee.Althen@wtcd.DaytonOH.ATTGIS.COM >Information Products Division ...!uunet!ncrcom!wtcp!wtcd!lalthen >AT\&T GIS, So Patterson Bldv FAX: (513) 445-2814 >Dayton, OH 45479 Tele: (513) 445-3487 > >-- >Lee E. Althen (WHQ-2) Lee.Althen@wtcd.DaytonOH.ATTGIS.COM >Information Products Division ...!uunet!ncrcom!wtcp!wtcd!lalthen >AT\&T GIS, So Patterson Bldv FAX: (513) 445-2814 >Dayton, OH 45479 Tele: (513) 445-3487 - - Using SGML may be a good idea, but is it practical? I believe you would be better of looking at HTML version 3 before you go down the SGML route. Information on HTML can be found in http://www.w3.org - My reasoning is that HTML 3 may give you all the capabilities you need, it may be easier to find an editor, and viewers. SGML stand alone products are not cheap. You also have to develop the DTD, and have your unique DTD baggage to carry with the document. Regards, Ken Newsgroups: wpi.techwriting,bit.listserv.techwr-l,misc.books.technical,comp.text,comp.text.pdf,comp.text.sgml,comp.text.desktop,comp.text.interleaf From: dichter@netcom.com (Carl Dichter) Subject: Documentation tools for cross platform Message-ID: \ Organization: Cross Platform Strategies Date: Wed, 28 Jun 1995 05:02:09 GMT Sender: dichter@netcom6.netcom.com We have an article in the next issue about cross platform documentation strategies. We'd like to mention commercial tools that would be helpful to our readers. Perhaps you use, sell, or know of documentation tools that work on multiple platforms, or have features (like conditional text) that are helpful when writing manuals for multiple platforms. We are also interested in add-ons packages for major word processors that help in cross platform use. Our list so far includes: FrameMaker (and FrameBuilder) Interleaf Word Perfect Adobe Acrobat Electronic Book Technologies: DynaText, DynaTag, InStd, DynaWeb MS Word (adapted/emulated on Sun and HP-UX) Nroff, Groff, etc. TEX, GnuTEX, etc. What other pertinant products are there? Please reply by email or send information to the address below. Carl Dichter, Editor Cross Platform Strategies 1960 W. Ray Road #1-A25 Chandler, AZ 85224 -- Carl Dichter dichter@netcom.com Editor of Cross Platform Strategies. Engineer at Motorola. Co-author (with Mark Pease) of Software Engineering with Perl: Toolsmithing and Prototyping for Better Software -- Sooner. From: Martin Hollender \ Newsgroups: comp.text.sgml Subject: Re: CORBA Date: 28 Jun 1995 06:14:36 GMT Organization: EURISCO Message-ID: <3sqs0c$4ch@naurouze.cert.fr> References: <9506271527.AA25261@source.asset.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Macintosh; I; PPC) X-URL: news:9506271527.AA25261@source.asset.com "Claude L. Bullard" \ wrote: >Can someone provide a network address or Web site for information on the >CORBA specifications (2.0?)? http://www.acl.lanl.gov/sunrise/DistComp/Objects/corba.html seems to be an interesting starting point Martin Hollender eurisco - European Institute of Cognitive Sciences and Engineering From: Fred K Weil \ Newsgroups: comp.text.sgml Subject: Re: _Practical SGML_ Date: Wed, 28 Jun 1995 06:46:05 GMT Organization: Private Internet Node Message-ID: <806461876wnr@fred1.demon.co.uk> References: \ Reply-To: fredw@fred1.demon.co.uk X-NNTP-Posting-Host: fred1.demon.co.uk X-Broken-Date: Wednesday, Jun 28, 1995 06.46.05 X-Newsreader: Newswin Alpha 0.7 In article: \ ak116@FreeNet.Carleton.CA (Bonnie Robinson) writes: > > I am in the process of learning SGML and would like a copy of the second > edition of _Practical SGML_. No bookstores in Toronto or Ottawa carry it, > and when I tried to order it, I learned that it is not available from the > publisher until October! Try direct order to Klewer Publishing +31 78 524400. (NB Europe therfore allow for the time zone difference, phone early in your working day) I have a copy bought in the UK. Klewer handle mail order by credit card > > If anyone has a copy that they don't want, or can track it down for me, I > would be eternally grateful. > > Bonnie Robinson > ak116@freenet.carleton.ca > > > -- ----------------------------------------------------------------- Fred K Weil EMail fredw@fred1.demon.co.uk ----------------------------------------------------------------- From: Grzegorz Staniak \ Newsgroups: comp.text.pdf,alt.hypertext,comp.text.sgml,alt.fan.mozilla Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 28 Jun 1995 08:02:15 GMT Organization: ZIO UMCS, Pl. Marii Curie-Sklodowskiej 1, 20-031 Lublin, Poland Message-ID: <3sr2a7$kg2@priam.umcs.lublin.pl> References: <3mgo02$mtp@doc.jmu.edu> \ <3r4ad9$t4t@imp.demon.co.uk> \ <3s6j7s$1013@hearst.cac.psu.edu> <3s76f1$5td@googol.bctel.net> \ <3sa7hq$qqc@googol.bctel.net> \ <3si07f$dc5@googol.bctel.net> <3sksu1$h19@scapa.cs.ualberta.ca> \ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.2b1 (Windows; I; 16bit) boo@primenet.com (Walter Ian Kaye) wrote: [...] > > What do you mean by "floating HTML version X.X"? HTML 2.0 is stable, and > > widely implemented. > > > Gerald > >But HTML 2.0 wasn't even finalized before this month... So? HTML 3.0 is still being discussed, and the more stable parts of it are already being implemented in Mosaic, Netscape and the like. The official staus of the standard is one thing, and stability/wide implementation is another. ># # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ># Walter Ian Kaye: (602) 942-6390 FoxPro/Excel Programmer; Guitarist # ># Correspond to: boo@primenet.com, boodlums@genie.com # ># BinHex files: boo@primenet.com WWW: http://www.primenet.com/~boo/ # ># # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # -- ------------------------------- Grzesiek Staniak \ Newsgroups: comp.text.sgml From: Steve Pepper \ Subject: Re: DSSSL Message-ID: <1995Jun28.081805.7707@falch.no> Organization: Falch Infotek, Oslo, Norway References: \ Date: Wed, 28 Jun 1995 08:18:05 GMT clollin@bgnet.bgsu.edu (Christopher Lollini) writes: >I'm having a difficult time finding information on DSSSL. Can anyone point >me in the right direction? http://navysgml.dt.navy.mil/dsssl.html (Navy DSSSL Information) http://www.jclark.com/dsssl/ (James Clark's DSSSL page) http://www.falch.no/~pepper/DSSSL-Lite/ (DSSSL Lite) ftp://infosrv1.ctd.ornl.gov/pub/sgml/WG8/DSSSL (DSSSL standard) Best regards, Steve From: lsnyder@immcms1.redstone.army.mil (Lori Snyder Bolen) Newsgroups: comp.text.sgml Subject: Re: DSSSL Date: Wed, 28 Jun 1995 08:33:05 UNDEFINED Organization: Redstone Arsenal, Alabama Message-ID: \ X-Newsreader: Trumpet for Windows [Version 1.0 Rev B] Can not find this URL: ftp://infosrv1.ctd.ornl.gov/pub/sgml/WG8/DSSSL (DSSSL standard) Any suggestions or thoughts? Lori Bolen From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: Re: SGML Author, MicroSift Date: 28 Jun 1995 13:56:17 GMT Organization: SEMA Group Belgium Message-ID: <3srn21$6m8@news.Belgium.EU.net> References: <1814@wtcd.DaytonOH.ATTGIS.COM> <271743554wnr@omimjd.demon.co.uk> <3spv8r$mac@cauldron.spdc.ti.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) fritz@hc.ti.com (Fritz Whittington) wrote: (Quotes ...) > >I don't want to get into a "which product is best" war here, but I feel >I should point out a few things. > >Unless I'm mistaken, SGML Author was first released marked *only* for >the US version of Word. If you in the UK applied it to a UK version of >Word, I suspect that it would mess it up. (If you did get a UK version >of SGML Author, then I don't know why it didn't work.) > >Secondly, "At best it is an SGML converter." YES! It allows you to >actually print a document in a form pleasant to the majority of people, >while maintaining an SGML form that may be useful for other purposes. > >I have installed it and used it. It works well for me. I have not >tried to do so, but you can supply your own DTD, define your own Word >styles for each element, and provide the rules for mapping each element >in SGML to a style or format in Word. > You did not try it. In our opinion, the things you mention are really hard to do, especially when you compare it to other add-ons to Word. >You may start with an SGML document, and import it to Word for viewing >and printing. Not unless you have defined all appropriate paragraph and character styles PLUS all the mappings between elements-in-context and styles. ------------------------------------------------------ Jacques Deseyne \ Sema Group Belgium Stallestraat 96 B-1180 Brussels Belgium info-markit@sema.be info-sgml@sema.be ------------------------------------------------------ From: Ted Freeman <75701.1760@CompuServe.COM> Newsgroups: comp.text.sgml Subject: ISO 12083 (articles) Date: 28 Jun 1995 14:17:35 GMT Organization: Allen Press, Inc. Message-ID: <3sro9v$9db$1@mhafm.production.compuserve.com> I posted this message several days ago and received no response, so here it goes again. I'm still hopeful. I ftped a copy of the 12083 DTD for articles from the Exeter site (by the way, what does "DIS DTD" mean?) and would like to get the "standard" for it., e.g., something that describes and/or illustrates its use. Is the following publication what I need, and if so, what am I getting? International Organization for Standardization (ISO). ISO 12083:1993(E) Information and documentation - Electronic manuscript preparation and markup. First edition, 1994-01-15. Geneva: International Organization for Standardization, 1994. 96 pages. Thanks for any advice. -- Ted Freeman E-Mail: tfreeman@allenpress.com Allen Press, Inc. Lawrence, Kansas From: Martin Bryan \ Newsgroups: comp.text.sgml Subject: Re: Trademark Entity? Date: Wed, 28 Jun 95 15:17:46 GMT Organization: The SGML Centre Message-ID: <804352666snz@sgml-cen.demon.co.uk> References: <3snims$fiv@news.znet.net> Reply-To: mtbryan@sgml-cen.demon.co.uk X-NNTP-Posting-Host: sgml-cen.demon.co.uk X-Newsreader: Demon Internet Simple News v1.29 In article <3snims$fiv@news.znet.net> dbogus1@znet.com "Darrin Bogue" writes: > Hi! I was wondering if there is a code for trademark... I can't find > it anywhere. Thanks. > > please respond to: daver@digpath.com > > Try \™ from the ISOnum set of Numeric and Special Graphic character set provided in ISO 8879 and ISO TR 9573 -- Martin Bryan @ The SGML Centre, Churchdown, Glos. GL3 2PU, UK (+44 1452 714029) From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: DSSSL Date: 28 Jun 1995 15:36:57 GMT Organization: Naggum Software; +47 2295 0313 Expires: 12 Jul 1995 15:36:56 GMT Message-ID: <19950628T153657Z@naggum.no> References: \ [Lori Snyder Bolen] | Can not find this URL: | | ftp://infosrv1.ctd.ornl.gov/pub/sgml/WG8/DSSSL (DSSSL standard) | | Any suggestions or thoughts? ftp.ifi.uio.no:/pub/SGML/DSSSL (or ftp://ftp.ifi.uio.no/pub/SGML/DSSSL to URL-fans.) #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: mjamesb@bnr.ca (M. James Bartley) Newsgroups: comp.text.sgml Subject: Re: DSSSL Followup-To: comp.text.sgml Date: 28 Jun 1995 16:25:35 GMT Organization: Bell-Northern Research Ltd., Ottawa, Canada Message-ID: <3srvpv$de1@bmerhc5e.bnr.ca> References: \ Reply-To: mjamesb@bnr.ca Originator: mjamesb@bmerh987 In comp.text.sgml, lsnyder@immcms1.redstone.army.mil (Lori Snyder Bolen) writes: > Can not find this URL: > > ftp://infosrv1.ctd.ornl.gov/pub/sgml/WG8/DSSSL (DSSSL standard) It looks good to me: 230- Connected to infosrv1.ctd.ornl.gov. 220 infosrv1 FTP server (OSF/1 Version 5.60) ready. 230 Guest login ok, access restrictions apply. Remote system type is UNIX. Using binary mode to transfer files. ftp> cd pub 250 CWD command successful. ftp> cd sgml 250 CWD command successful. ftp> cd WG8 250 CWD command successful. ftp> cd DSSSL 250 CWD command successful. ftp> ls 200 PORT command successful. 150 Opening ASCII mode data connection for /bin/ls (192.58.194.77,2412). -rw-rw-r-- 1 15667 300 1080 Nov 9 1994 README -rw-rw-r-- 1 15667 300 39627 Nov 7 1994 arabic.eps -rw-rw-r-- 1 15667 300 15284 Nov 7 1994 cl1to3.sgm -rw-rw-r-- 1 15667 300 9992 Nov 7 1994 cldefs.sgm -rw-rw-r-- 1 15667 300 27212 Nov 7 1994 clmodel.sgm -rw-rw-r-- 1 15667 300 3558 Nov 7 1994 clnotat.sgm -rw-rw-r-- 1 15667 300 160742 Nov 7 1994 clstfp.sgm -rw-rw-r-- 1 15667 300 196668 Nov 7 1994 clsttp.sgm -rw-rw-r-- 1 15667 300 15184 Nov 7 1994 colset.cdr -rw-rw-r-- 1 15667 300 22997 Nov 7 1994 colset.eps -rw-rw-r-- 1 15667 300 18102 Nov 7 1994 display.cdr -rw-rw-r-- 1 15667 300 27142 Nov 7 1994 display.eps -rw-rw-r-- 1 15667 300 1111805 Jan 13 13:51 dsssl.pdf -rw-rw-r-- 1 15667 300 3732116 Nov 13 1994 dsssl.ps -rw-rw-r-- 1 15667 300 23784 Nov 7 1994 french.eps -rw-rw-r-- 1 15667 300 13294 Nov 7 1994 inline.cdr -rw-rw-r-- 1 15667 300 21833 Nov 7 1994 inline.eps -rw-rw-r-- 1 15667 300 50172 Nov 7 1994 lev2mod.eps -rw-rw-r-- 1 15667 300 248636 Nov 7 1994 mixed.eps -rw-rw-r-- 1 15667 300 34270 Nov 7 1994 stfpmod.eps -rw-rw-r-- 1 15667 300 18360 Nov 7 1994 syncol.cdr -rw-rw-r-- 1 15667 300 28325 Nov 7 1994 syncol.eps -rw-rw---- 1 15667 300 266384 Nov 7 1994 univers.pfa 226 Transfer complete. ftp> -- M. James Bartley, Bell-Northern Research Ltd. \\ Internet: mjamesb@bnr.ca P.O. Box 3511, Station C, Ottawa, Canada, K1Y 4H7 \\ Phone: 613-763-3556 ** Opinions expressed are mine, not BNR's. - MJB ** \\ Fax: 613-765-5750 "Stupid pigeons!" Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 28 Jun 1995 12:29:10 -0400 From: "Alan Fleischman" \ Message-ID: <9505288043.AA804356950@dts68po.volpe.dot.gov> Subject: Request for information on SGML implementations The Volpe National Transportation Systems Center is assisting a system support organization, primarily performing software maintenance, to develop a plan to transition to an SGML based electronic documentation environment. Part of this effort includes development of a pilot system which will allow one program to provide technical documentation to 140 field sites in electronic form. We would very much like to factor into our project the current thinking of commercial and government organizations that are pursuing similar strategies. Insights with respect to development of effective transition strategies and lessons learned from actual system implementation would be particularly beneficial to us at this stage. Types of insights that we are interested in include: o Repository strategy - proprietary document management vs. SGML o SGML repository schema development and revision - Relational and object oriented technology o Query strategy for SGML repository- application/report generation o Simultaneous management of SGML, raster images and paper documents o Integration of document management system with network operating systems o Automation of document revision processes - library validation of change page, redlining, editing, quality assurance o Experience with automated tools for document analysis From: quixote@Primenet.Com (Donald A. Hosek) Newsgroups: comp.text.sgml Subject: Re: SGML FAQ Date: 28 Jun 1995 19:12:02 GMT Organization: PrimeNet Message-ID: <3ss9i2$g85@nnrp3.primenet.com> References: <3so8o0$bc8$1@mhadf.production.compuserve.com> In article <3so8o0$bc8$1@mhadf.production.compuserve.com>, Neil McNaughton <100045.3625@CompuServe.COM> wrote: >1.) Is comp.text.sgml some sort of a secret society? If it were, and I'm not saying that it is or it isn't, we couldn't tell you anyway. -dh -- Don Hosek dhosek@quixote.com 909-621-1291 fax: 909-625-1342 Quixote Digital Typography http://www.quixote.com/ Publishers of _Serif: The Magazine of Type and Typography_ Newsgroups: comp.text.sgml From: niel@fgh.fgh.oz.au (Niel Sayo) Subject: FAQ for this group Message-ID: \ Organization: FGH Decision Support Systems Pty Ltd Date: Wed, 28 Jun 1995 23:18:22 GMT Hi, I'm a newbie in this newsgroup and I'm quite interested to know more about SGML. Does this newsgroup have a FAQ file somewhere? If so, could someone please tell me which site I could get it from? Thanks in advance! Niel Sayo (niel@fgh.fgh.oz.au) From: John Lamp \ Newsgroups: comp.text.sgml Subject: Re: SGML FAQ Date: 29 Jun 1995 00:21:59 GMT Organization: University of Tasmania, Department of Computer Science Message-ID: <3ssrn7$ige@franklin.cc.utas.edu.au> References: <3so8o0$bc8$1@mhadf.production.compuserve.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Macintosh; I; 68K) To: 100045.3625@CompuServe.COM X-URL: news:3so8o0$bc8$1@mhadf.production.compuserve.com Neil McNaughton <100045.3625@CompuServe.COM> wrote: >I posted a request for info on SGML a while back vis > >1.)Is there a FAQ on SGML This seems to be the main FAQ on comp.text.sgml There is an ftp'able FAQ at ftp://sgml1.ex.ac.uk/pub/SGML and other sites, but it doesn't ever seem to get posted to the group. >2.)What are the main differences between SGML and HTML, is the second a >subset of the first, and what does SGML get you that HTML doesn't. HTML is a specific markup language generated using SGML and the HTML DTD (document type declaration). >So far, no response. > >So I now have a few more questions vis bis.. > >1.) Is comp.text.sgml some sort of a secret society? [the answer to this question has been deleted - see the charter in net.cop] >2.) Are the above questions too hard or too easy? six of one and half a dozen of the other. >Hope someone can help. Seriously, though, we should sort out a basic cron job to get the faq posted regularly. Cheers John -- _--_|\\ John Lamp, originating in Hobart, Tasmania / \\ Phone: 002 20 2375 - Fax: 002 20 2913 \\_.--._/ email: John.Lamp@cs.utas.edu.au v <--<< http://lamp.cs.utas.edu.au/jw_lamp.html From: asengupt@ucs.indiana.edu (Arijit Sengupta) Newsgroups: comp.text.sgml Subject: Re: SGML FAQ Date: 29 Jun 1995 00:51:50 GMT Organization: Indiana University Message-ID: <3sstf6$nut@usenet.ucs.indiana.edu> References: <3so8o0$bc8$1@mhadf.production.compuserve.com> X-Newsreader: TIN [version 1.2 PL2] I believe there were two responses to this one posted. There are no official faq on sgml, a couple unofficial ones are floating around, a couple of them have pointers at http://www.sil.org/sgml/sgml.html : 2.)What are the main differences between SGML and HTML, is the second a : subset of the first, and what does SGML get you that HTML doesn't. Think of html as a program written in C, where SGML is C.... : 1.) Is comp.text.sgml some sort of a secret society? : 2.) Are the above questions too hard or too easy? None of the above, just make sure that you read the newsgroup as well as look at your messages :) Jit. -- _|_|_|_|_| _|_|_| _|_|_|_|_| | asengupt@indiana.edu _| _| _| http://www.cs.indiana.edu/hyplan/asengupt.html _| _| _| | Computer Science, LH215 _| _| _| _| | Indiana University, Bloomington IN47405 _|_|_|ust _|_|_|n _|ime | (812) 855-4318 / (812) 334-2695 From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: MS SGML Author Application Building Experience Date: 29 Jun 1995 01:06:54 GMT Organization: Naggum Software; +47 2295 0313 Expires: 13 Jul 1995 01:06:54 GMT Message-ID: <19950629T010654Z@naggum.no> References: <1813@wtcd.DaytonOH.ATTGIS.COM> <3sqjvc$g2f@sally.dma.org> [Ken Melhaven] | Using SGML may be a good idea, but is it practical? I believe you | would be better of looking at HTML version 3 before you go down the | SGML route. Information on HTML can be found in http://www.w3.org HTML is an application of SGML, so by using HTML 3.0 you do, indeed, go down the SGML route. HTML is a particular solution to a particular problem. SGML is a general solution to a general problem. it will of course be easier to use a particular solution, but it usually does not fit your particular problem. that's why somebody took the SGML route to build HTML 3.0 for one particular set of problems. you seem to approve of this. | My reasoning is that HTML 3 may give you all the capabilities you need, | it may be easier to find an editor, and viewers. SGML stand alone | products are not cheap. true, many of them are actually free. | You also have to develop the DTD, and have your unique DTD baggage to | carry with the document. DTD development is a process of understanding and caring for information. this process can only be ignored at the peril of the information. like ignoring your own health, you won't notice it for a long time, but you might live to regret it for an even longer time. actually, I think it's better to say "application development" instead of DTD development -- the idea is not just to write a DTD that will "accept" your data, but to describe the meaning and structure of your data in a DTD that your data should fit. this is a matter of perspective, and one that I think is being taken more seriously both in the industry and on the Net. SGML is about typing information such that we can reason about and work with progressively more abstract types, as opposed to progressively larger numbers of discrete chunks of (untyped) information. (that's "typing" as in "strong typing", not to be confused with the clerical task of the same name.) #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: Kimmo Elovainio \ Newsgroups: comp.text.sgml Subject: AmiPro Date: 29 Jun 1995 09:26:47 GMT Organization: Technical Research Centre of Finland, Information Technology Message-ID: <3strkn$stm@hemuli.tte.vtt.fi> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit I am wondering if there do not exist any plans for implementing SGML support for AmiPro word processing tool. Could anybody confirm this? -Kimmo Elovainio VTT Information Technology From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: SGML FAQ Date: 29 Jun 1995 09:56:30 GMT Organization: Naggum Software; +47 2295 0313 Expires: 13 Jul 1995 09:56:30 GMT Message-ID: <19950629T095630Z@naggum.no> References: <3so8o0$bc8$1@mhadf.production.compuserve.com> [Neil McNaughton] | I posted a request for info on SGML a while back vis | | 1.)Is there a FAQ on SGML no. see ftp.ifi.uio.no:/pub/SGML for the SGML Repository. you will also find pointers to other useful sites there. | 2.)What are the main differences between SGML and HTML, is the second a | subset of the first, and what does SGML get you that HTML doesn't. SGML is like a programming language. HTML is like a program written in that language. the difference is in the "G". the Standard _Generalized_ Markup Language is not a "markup language", but a markup language generator. the Hypertext Markup Language is such a a "markup language". | So far, no response. the above quoted article is the first article by anyone named McNaughton to the newsgroup since I started archiving it in the fall of 1990. (archive site: ftp.ifi.uio.no:/pub/SGML/comp.text.sgml). this could explain the lack of responses. | So I now have a few more questions vis bis.. | | 1.) Is comp.text.sgml some sort of a secret society? yes. | 2.) Are the above questions too hard or too easy? yes. #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: lab@gateway.grumman.com (Larry Beck) Newsgroups: comp.text.sgml Subject: Change of Status Date: 29 Jun 1995 07:01:38 -0400 Organization: Northrop Grumman, Bethpage NY Message-ID: <3su16i$d38@gateway.grumman.com> As of close of business today, June 29th, 1995, my almost 21 years of association with Grumman Data Systems has ended. As of July 31st, 1995 I will be employed by Atlis Systems Inc. of Rockville MD. LAB From: erik@naggum.no (Erik Naggum) Newsgroups: comp.text.sgml Subject: Re: Join SGML Newswire! Date: 29 Jun 1995 11:13:56 GMT Organization: Naggum Software; +47 2295 0313 Expires: 13 Jul 1995 11:13:55 GMT Message-ID: <19950629T111356Z@naggum.no> References: <3spro3$536@news-4.csn.net> [Terry Brien] | The listserv "sgmlinfo" is alive and well and sponsored by Interleaf | Avalanche. Topics range from DTDs, CALS table models, discussions, | product reviews, SGML articles, and upcoming events. Occasionally | posts from this group may appear. Comments, questions, and suggestions | are also welcomed by the newswire manager. Newswire articles are also archived at the U of Oslo SGML Repository. consult ftp.ifi.uio.no:/pub/SGML/Newswire for the articles, one per file, named nnn.yyyy-mm-dd, for article #nnn, published on the dd'th day of the mm'th year, of the yyyy'th year. subscribers will find this number at the head of the Newswire article. if you use a reasonable FTP client, you should be able to give the command "ls -t" or "dir -t" to sort the file listing by date. (there may be "buttons" for this in your GUI "browser".) #\ -- NETSCAPISM /net-'sca-,pi-z*m/ n (1995): habitual diversion of the mind to purely imaginative activity or entertainment as an escape from the realization that the Internet was built by and for someone else. From: tad@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: SGML FAQ Date: 29 Jun 1995 12:50:43 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <3su7j3$rp@cliffy.lfwc.lockheed.com> References: <3so8o0$bc8$1@mhadf.production.compuserve.com> <3sstf6$nut@usenet.ucs.indiana.edu> X-Newsreader: TIN [version 1.2 PL1] Arijit Sengupta (asengupt@ucs.indiana.edu) wrote: : : 2.)What are the main differences between SGML and HTML, is the : second a : : subset of the first, and what does SGML get you that HTML doesn't. : Think of html as a program written in C, where SGML is C.... Or think of SGML as yacc and HTML as a grammar described in yacc syntax. Or think of HTML as a specific instance of an SGML application. -- --------------------------------------------------------------------- | Tad McClellan, Logistics Specialist (IETMs and SGML guy) | Lockheed Martin Tactical Aircraft Systems | voice: (817) 763-6314 | email: mcclellantj@lfwc.lockheed.com --------------------------------------------------------------------- | | Just be thankful we don't get all the government we pay for! | --------------------------------------------------------------------- From: hal@bnr.ca (Harold Floysvik) Newsgroups: comp.text.sgml Subject: Graphics Markup Language? Followup-To: comp.text.sgml Date: 29 Jun 1995 13:26:17 GMT Organization: Northern Telecom Message-ID: \ General question to everyone: Is there a graphical equivalent to SGML that allows one to define the structure and organization of graphics contained within a document? (And what about sound and animation - or is that HyTime?). If not, what comes close? Our objective is to deliver editable-form graphics which are device and application independent along with the SGML instance of the document. By editable-form, I mean that, for example, grouped objects stay grouped, polygons remain as discrete polygons. Transforming complex objects to dozens or hundreds of discrete vectors, though they are trivially 'editable', is missing the point. CGM and other so-called standards don't seem to have enough power to deal with modern graphics. Grouped objects, Bezier and spline curves, certain exotic polygons, and text within the graphic (particularly formatted text with tabs and returns) are not handled gracefully at all. Thanks in advance for any help or advice you can supply. ...Hal -- Harold Floysvik | My opinions are not necessarily the Northern Telecom Canada | same as those of Northern Telecom. Phone: 613-727-7979 | ---------------------------------------------- e-mail: hal@bnr.ca | "We are confronted with insurmountable | oportunities" ....Pogo From: sthompso@ix.netcom.com (Scott Thompson) Newsgroups: comp.text.sgml Subject: HTML/SGML Editors Date: 29 Jun 1995 13:35:44 GMT Organization: Netcom Message-ID: <3sua7g$am7@ixnews6.ix.netcom.com> I am looking for a good SGML editor. I realize that some companies have a demo of their software using the HTML DTD. Can anyone tell me of Web Pages/FTP sites or other ways I counld get hold of the demos or information on these editors. Please either email me or respond Scott P Thompson sthompson@metatec.com sthompso@ix.netcom.com From: quixote@Primenet.Com (Donald A. Hosek) Newsgroups: comp.text.sgml Subject: Which is broken: SGMLS or my brain? Date: 29 Jun 1995 15:46:43 GMT Organization: PrimeNet Message-ID: <3suht5$h9p@nnrp1.primenet.com> \ \ \ \ ]> \ \Foo \ If the content for test is foo? or foo*, I get a parse error. Change that to foo or foo+ and all is well. sgmls bug or have my brains pickled? -dh -- Don Hosek dhosek@quixote.com 909-621-1291 fax: 909-625-1342 Quixote Digital Typography http://www.quixote.com/ Publishers of _Serif: The Magazine of Type and Typography_ From: John Lehman \ Newsgroups: comp.text.sgml Subject: Re: SGML and double-byte character sets Date: 29 Jun 1995 18:31:48 GMT Organization: University of Alaska Message-ID: <3surik$3ua@news.alaska.edu> References: <3scjuk$grb@grok.provo.Novell.COM> <3sdti8$7is@news.Belgium.EU.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Macintosh; I; 68K) X-URL: news:3sdti8$7is@news.Belgium.EU.net Jacques Deseyne \ wrote: >... >The SGML Asia '95 Conference organized by GCA (the week of Oct. 25 ?) >could provide you a lot of information on these topics. > Does anyone have more details on this conference? --- John Lehman (907) 474-6275 (voice) (907) 474-5140 (FAX) University of Alaska Statewide Office of Network Services Snailmail: PO Box 755320, Fairbanks, AK 99775-5320 USA From: Mark_W_Lumsden@vnet.ibm.com Newsgroups: comp.text.sgml Subject: IBM Electronic Publishing Edition now available Date: 29 Jun 1995 18:38:57 GMT Organization: IBM Electronic Publishing Message-ID: <3sus01$1fd3@sernews.raleigh.ibm.com> Reply-To: Mark_W_Lumsden@vnet.ibm.com Summary: WWW publishing system for SGML source documents Keywords: IBM Electronic Publishing Edition X-Newsreader: IBM NewsReader/2 v1.02 IBM Electronic Publishing Edition, discussed previously in this newsgroup, is now generally available. A single server license is priced at $1999. EPE is a WWW publishing system for complex documents authored in SGML or word processors. Please see \ for more information and how to order. Newsgroups: comp.text.sgml From: chsaucie@qc.bell.ca (Christian Saucier) Subject: SGML and Arabic languages X-Nntp-Posting-Host: cm7c35.qc.bell.ca Message-ID: \ Sender: news@on.bell.ca Reply-To: chsaucie@qc.bell.ca Organization: Bell Sygma X-Newsreader: IBM NewsReader/2 v1.09 Date: Thu, 29 Jun 1995 18:41:11 GMT Big question here: Is there any SGML system that supports the arabic languages? I'm thinking about an Editor, a document formatter, online book reader. I'm just starting my investigation on Arabic support on today's computer systems and I'd really like to receive any input from anybody using SGML in an Arabic environment. Thanks, Christian P.S.: Jeez, I don't even know what kind of arabic support there is on Unix systems... -- Christian Saucier -- chsaucie@qc.bell.ca +1-514-870-8187 "disconnect me if you can" -SoM From: mskim@kiet.etri.re.kr (Kim Mansoo) Newsgroups: comp.text.sgml Subject: [Q] sgml parser for windows NT (or windows) Date: 30 Jun 1995 02:39:32 GMT Organization: ETRI Message-ID: <3svo54$b4d@nms.etri.re.kr> Reply-To: mskim@com.etri.re.kr X-Newsreader: WinVN 0.92.1 Dear, Would you like to inform me where sgml-parsers for windows or windows NT are ? (it is better that the parser is programed by C or Visual C++). Thanks in advance. From: mlpiotro@linguistik.uni-erlangen.de (Michael Piotrowski) Newsgroups: comp.text.sgml Subject: sp binary for HP-UX needed Date: 30 Jun 1995 05:13:11 GMT Organization: CLD, University of Erlangen, Germany Message-ID: <3t0157$fb4@cd4680fs.rrze.uni-erlangen.de> Hello, everybody! Is there someone out there who would be so kind to provide me with a HP9000 Series 700 binary of James Clark's sp? It doesn't compile here, and my sysop won't re-install gcc, as Nelson H. F. Beebe recommends on his great WWW-page ``Notes about SP'' (http://www.math.utah.edu/~beebe/sp-notes.html). I'd be very pleased if someone could help me. Thanks in advance Michael --- Eat your cereal with a fork | Michael Piotrowski and do your homework in the dark. | mlpiotro@linguistik.uni-erlangen.de ---Happy Harry Hardon | mlpiotro@cip.informatik.uni-erlangen.de From: timhnen@pori.tut.fi (Hamalainen) Newsgroups: comp.text.sgml Subject: Looking for converter - Ventura 4.2 to HTML Date: 30 Jun 1995 06:32:06 GMT Organization: Tampere University of Technology, Computing Centre Message-ID: <3t05p6$1n6@cc.tut.fi> X-Newsreader: TIN [version 1.2 PL2] I am looking for a program to convert documents written with WP5 (ventura 4.2 tags added) to HTML. Or is there a program where I can spesify tags to be changed to other. Thanks in advance, Timo Hamalainen timhnen@pori.tut.fi From: j_mcarthur@bix.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Semi-Regular Posting of the Unofficial Non-Sanctioned FAQ for c.t.s Date: 30 Jun 1995 06:52:37 GMT Organization: \ Message-ID: <3t06vl$raq@news1.delphi.com> X-Newsreader: Galahad 1.1c Un-Official, Non-Sanctioned Frequently Asked Questions List For comp.text.sgml Compiled by: Jeffrey McArthur j_mcarthur@bix.com 1. About This FAQ I wish to thank the many people who have sent me comments about my FAQ. I have tried to incorporate almost all of them. In particular I wish to thank Dave Peterson; I use much of his rewrite of the section on entities (and used some of his comments in other sections as well). 1.1. Q: Where is the FAQ for comp.text.sgml? Revised: Apr 29, 1995. A: You are reading the Un-Official, Non-Sanctioned version of the FAQ. The most frequently asked question on this newsgroup has been: "where is the FAQ?". Rather than start an rwar about who is right or wrong or if there should be a FAQ at all, I decided that it would be in my best interest to provide my own version of the FAQ. Quite a few people firmly believe that there should not be a FAQ for this group. Others believe that unless the FAQ is perfect in every way, it should never be published. I respect other peoples views, I just disagree with them. This FAQ is far from perfect. There are numerous flaws, inconsitancies, and things that are wrong. But I do this for fun and relaxation (I am a truely warped person). So if you love it or hate it, I would like to hear from you. Just no more death threats. (Just joking, no one has threatened my life, yet.) This version will be posted in ASCII format only. The FAQ is actually maintained in SGML. The DTD for this FAQ is based on work done by Jeffrey C. Ollie; but has been higly modified. If you would like either a copy of the DTD used for this FAQ or the SGML version of this FAQ just send me a message via email and I will be glad to send you a copy. Typeset versions using TeX DVI format or postscript are also available. Because this FAQ is still quite rough, it is not avaliable via ftp. If you would like a copy, the only way to get it is via email. One final word of warning, since I started and maintain this FAQ, it will reflect my opinions. YOU HAVE BEEN WARNED! 1.2. Q: Why is this not the officially sanctioned FAQ? Revised: May 2, 1995. A: Because as Marcy Thompson wrote: "This is *not* a official FAQ for comp.text.sgml. It is Jeffrey MacArthur's somewhat eccentric and iconoclastic view of what a FAQ should be.". My only real argument with that is I wish people would spell my name correctly when they flame me. Marcy continues "I have said before and now say again that this document contains error of fact as well as opinions with which many denizens of the newsgroup would disagree. These are not marked as "Jeffrey McArthur's opinion" and are instead propogated as fact." Well I have never been afraid to voice my opinion. As I pointed out in the very first section of this document, since I started and maintain this FAQ, it will reflect my opinions. So to avoid needless arguments, this is now, and evidently for all time, the Un-Official, Non-Sanctioned version. However, it is the ONLY attempt at an FAQ that is maintained for this newsgroup. 1.3. Q: Why is this document so contentious? Revised: May 2, 1995. A: I post my opinions. I am not one of the "breathern". I do not worship at the alter of SGML. Personally, I think SGML is very overrated. SGML was created by a "dedicated group of individuals", some who are lawyers. That says a lot about SGML. I work with SGML every day. But like the janitor, I may not like what goes into the pumbing, but it does get the job done. When SGML works, it works well. That does not make it pretty or an object to be worshipped. But it does do a dirty job better than anything else I have found. When SGML breaks things, well it can leave a "smell" that can stay around for days or weeks. I also post my opinions about some of the SGML tools I have used, or tried to use. Unfortunately I try and use tools for real work. So one of the first things I usually try is to load in one of the small ten to twenty meg pseudo-SGML files I have. (The files come from our keyboarders who take manuscript, because that is all we have, and type the document in. So the files have keystroke errors, spelling errors, and tagging errors, and are guaranteed not to parse.) Many SGML tools are not designed to work with documents much larger than a meg. These tests tend to color my opinions about SGML tools. But as the saying goes, "your milage may vary". 2. General Information 2.1. Q: What is Markup? Revised: Apr 29, 1995. A: Webster's does not define markup as a word. Webster's does define "mark up" as: 1 to cover with marks; 2 to mark for sale at an increased price; 3 to add overhead and profit to the cost in order to arrive at the selling price. Using a highlighter pen to emphasize passages in a book is "marking up" the book. That is, it add marks to an existing document. The highlights show passages that are important to the reader. Underlining is another form of markup. It is not possible to use a highlighter on an electronic document. To implement electronic markup a variety of ideas have been developed and used 2.2. Q: What is Tagging? Revised: Apr 29, 1995. A: ASCII has become the most commonly used form of information exchange. Almost every word processor has the ability to import and export an ASCII text file. The 95 printing characters of the 7-bit ASCII definition do not provide any information about the structure or the format of the document. Several methods have been developed to specify additional information in an ASCII text file. FORTRAN used a "C" in the first column of a punched card to specify a comment. This was one of the simplest forms of tagging. All the comment lines were tagged with a "C" in the first column. ALGOL allowed comments to be placed almost anywhere. This was done by introducing a start comment sequence, "(*", and an end comment sequence, "*)". The basic idea is to use a recognizable sequence of characters to define parts of a document. Each special sequence of characters is called a "tag". Below is a list of tags used in some computer languages: Language Start Tag End Tag FORTRAN C in first column column 72 ALGOL (* *) C /* */ Basic REM in first column end of line In an ASCII file, you can mark up parts of the text by enclosing them between tags. The process of doing this is called tagging. Over time, comments in computer languages have changed. One of the more interresting changes is the ability to nest comments. Most of the newer Algol family of languages allow comments to nest. This included Modula-2, Oberon, and Oberon-2. The ability to nest is important in tagging. This allows using the same notation over and over again. The meaning of a tag become context dependant. In Modula-2 the first "(*" starts the comment. Following "(*" not only continue the comment but add the requirement of an additional "*)" to end the comment. For each start comment tag, there is an end comment tag. 2.3. Q: What does SGML mean? Revised: Apr 29, 1995. A: SGML stands for Standard Generalized Markup Language 2.4. Q: What is SGML? Revised: Apr 29, 1995. A: SGML is ISO standard 8879. SGML is a method of defining tags and their relationship and use in data. 2.5. Q: What is SGML used for? Revised: Apr 29, 1995. A: One of the problems with tagging is determining what tags to use. SGML takes the concept of tagging one step further. It creates a method for defining a set of tags. This is why some people refer to SGML as a meta-language. SGML does not define a set of tags. It is a tool to define a set of tags. With SGML you can define a "tag grammar", and you check see whether a text conforms to that grammar. The process of doing this check is called "validating". 2.6. Q: What is the History of SGML? Revised: Apr 29, 1995. A: Erik Naggum \ wrote a short history in the official but non-maintained FAQ. Charles F. Goldfarb also wrote an article entitled "A Brief History of the Development of SGML" for the SGML Users' group. If I get his permission I will include it here. Or if anyone would care to write one I will post it here. 2.7. Q: Why go to all the trouble of using SGML? Revised: Apr 29, 1995. A: Word processor programs are very easy to use. You just type. Little thought is given to the structure of what is written. Style sheets provide some outline capabilities, but they do not force the document to match the style. Some SGML authoring tools try to allow a similar approach. You type. But SGML requires an attention to detail that is not required in a simple word processing document. Documents must be validated to ensure that they follow the structure defined in the DTD. This important step does not exist in word processors. SGML is tyrannical in its enforcement of structure. Documents must follow the defined structure, or they will not validate. But the tyranny is a benevolant one. Enforcement of structure guarantees some level of consistancy. The data must follow a predefined set of rules reguarding its structure. Another advantage is separating the "style" of the document from the "content" of the document. With a word processor, you are always aware of the form. You specify that a peice of text is to be printed in bold, but you do not specify why the text is to be bold. Unless the decision to make the text bold is a purely asthetic consideration, there is hidden information in the style change. SGML helps separate the style from the structure. The separation of style from structure is an important concept. It allows stylistic changes based on structure. For example, if all the part numbers are tagged then the way the part numbers are printed can be easilly changed. So it is possible with one simple change to make all the part numbers print out in a san-serif face instead of bold. Doing this in a word processing document would not be automatic since only the part numbers are changed to san-serif, not all the other bold text. One common misconception is that SGML tags can only define the structure of a document. It is possible to create a SGML document where the tags only describe the style. An example of this is in tables. Many table models only describe the formatting of the table. There is no attempt to represent any structure on the data other than the style. If the document is to be used only one time, for example a letter to a friend, there is no reason to use SGML. On the other hand, if the letter is to be placed into a system that is searched, queried, and/or printed many times in many different ways, then SGML is a major advantage. 2.8. Q: What is ISO 8879? Revised: Jan 15, 1995. A: SGML is an ISO standard. ISO is an acronym for International Organization of Standardization. ISO 8879 is the definition of SGML. The definitive document is: The SGML Handbook; Oxford University Press, 1990; ISBN 0-19-853737-9; by Charles F. Goldfarb. If you are serious about SGML, this book is a must. It is a very hard to read document. Also for a book that wants to show off the power of SGML the typesetting is awful. The indexes are almost useless because there is no distinction between a simple reference and a full description. Detail: ISO does not really an acronym for International Organization of Standardization, if it did it would IOS. The real name is in French. The rough english translation is International Organization of Standardization 2.9. Q: What is a DTD? Revised: Jan 15, 1995. A: A Document Type Definition (DTD) is an electronic document that defines a tagging structure. A DTD specifies where each tag is allowed. For example, a novel is made up of a set of chapters. Each chapter is made up of one or more sections. Each section is made up of one or more paragraphs. A DTD contains statements that define this relationship. DTD is the name for a tag grammar. 2.10. Q: What does it mean to validate an SGML document? Revised: Apr 29, 1995. A: This means that document has been tested against the DTD and matches the tagging grammar. This does not mean that the content of the document is correct. 2.11. Q: How do you validate an SGML document? Revised: Apr 29, 1995. A: You validate an SGML document by using a computer program called a parser. 2.12. Q: What is a parsing? Revised: Apr 29, 1995. A: Webster's defines parsing as: to break (a sentence) down into parts, explaining the grammatical form, function, and interrelation of each part. This is not exactly what we mean by parsing in SGML. Validating in SGML is done via a parser. A parser is a computer program that breaks down an electronic document into its parts and compares the form of the document based on the SGML tags to the form described in the DTD. Parsing is a check of conformance of a text to the grammar described in the DTD. Parsing is what separates SGML and other word processing formats. For example, in the case of a novel, this would mean that paragraphs only occur inside of sections, and sections only occur inside chapter. A word processor does not enforce those requirements. Detail: To be accurate, you do not "parse" an SGML document, you validate it. This is counter-intuative. The computer program that compiles is called a compiler; the process of using the compiler is called compiling. But in SGML the proper terminology is you use a parser to validate an SGML document and the process of using the parser called validating. If SGML were consistant, it would refer to the program as a validator. So parsing is an inexact term in SGML. It is often used as a synonym to validating. This use has some merit since to valid the parser must actually parse the data. But the ISO standard uses the term validate. 2.13. Q: What is legacy data? Revised: Jan 15, 1995. A: Legacy data is a term used by some to refer to data that has not been converted to SGML. The choice of terms is rather unfortunate. It gives the impression that nothing good could have been done prior to SGML. There are two issues options in converting legacy data. Change the existing data to match the DTD, or change the DTD to allow the structures in the existing data. The question is simple: what should define the DTD: the idealized model for new data, or the real-world existing data. As anyone who has done any work in physics realizes, working with real-world data can be a very difficult task. SGML enforces the structure defined by the DTD. But it is relatively easy to create a DTD that is totally unsuitable for a set of data. It is also possible to create a DTD that is so loose that no structure at all is enforced. Converting existing data generally requires a lot of compromise. If you have more than a couple of meg of unstructured data and want to convert it to SGML you will end up making massive changes to both the data and the DTD; unless you are very, very lucky. 3. SGML Language Features The syntax used to define a document tag definition. This section is used to provide a quick overview of of SGML and is not a complete description. Also the following is not exactly correct. There are predefined names for all the parts of each SGML statement. Although needed, the names build a wall to understanding for the novice. One aim of this FAQ is to make SGML easy to understand. So the following discussion will not use the proper names. 3.1. Q: What are Elements? Revised: Apr 29, 1995. A: Element are the basic building blocks of an SGML document. Each type of element is defined primarily by an element type declaration. One of the most common type of elements is to define a paragraph. Below is a simple element type declaration for a paragraph: \ There are 6 pieces to the declaration. The first two pieces are ". The end tags looks like this: \. The fourth piece of the declaration controls when the start and end tags are required. There are four values this piece can have. Below is a table showing what the values are and what they mean: - - Both tags required. - O Start tag required. End tag optional. O - Start tag optional. End tag required. O O Both tags are optional. The "O" can be either upper or lowercase. (You should be aware that "optional" here has a very specific meaning. SGML prescribes certain conditions under which a possibly missing tag can be inferred. If those conditions don't apply, even a tag specified "optional" here is still required.) The next piece defines the content of the element. In this case only PCDATA is allowed. PCDATA means parsable character data. The full technical meaning is somewhat complex. In general this is used to specify that a paragraph can have text (and a few other things) inside it. But no other tag can occur inside a paragraph. The content mode for an element type is actually a regular expression. A regular expression is a standard way of describing a pattern. Many editing programs support searching and replacing using regular expressions (if your editor of choice does not allow this, consider getting a more powerful editor, once you learn how to use them, they are very powerful). Does anyone care to write a good explanation of regular expressions? Below is a table showing the regular expression operators supported in SGML: Letter Meaning Notes ? Zero or one occurrence * Zero or more occurrences + One or more occurrences | or & and both in any order , and both in specified order The final piece of a element declaration is the closing ">". Another example is helpfull. The declaration for section can be defined as: \ A section's content must be one or more paragraphs. The "+" mean that there must be one or more (see the table above). Below is a pair of paragraphs showing how they are tagged as "para" elements: \Alex felt the melancholy stealing over him again. Nostalgia? For that germ-infested ball of mud? Not possible. He could barely remember it. Snapshots from childhood; a chaotic montage of memories. He had fallen down the cellar steps once in a childhood home he scarcely recalled. Tumbling, arms flailing, head thumping hard against the concrete floor. He hadn't been hurt; not really. He'd been too small to mass up enough kinetic energy. But he recalled the terror vividly. Now he was a lot bigger, and he would fall a lot farther.\ Detail: Actually almost everything in the above section is not true. It is a vast oversimplification. SGLM is not tied to any character set. So it is imprecise to refer to the actual characters. The actual grammar for the element section is shown in the table below: mdo, Normally As you can see the actual grammar is rather complex. In particular the meaning of what I refer to as SGML white space more complicated than a casual reader would expect. Element type, declared content and content model would require pages to fully explain. If you want to get into all the details of the actual SGML grammar you really need a copy of the ISO standard. For that, I recommend The SGML Handbook; Oxford University Press, 1990; ISBN 0-19-853737-9; by Charles F. Goldfarb. 3.2. Q: What are Attributes? Revised: Apr 29, 1995. A: Each element can have a set of attributes. Attributes allow additional information to be attached to the element. The paragraph example above works fine for simple paragraphs. But what about lists? Lists are little more that a sequence of special paragraphs. Defining a simple list is relatively easy: \ \ This works fine for simple lists. But there is no way to specify if the list is to be numbered, or bulletted, or whatever. Attributes provide the way to specify the type of list easily. \ There are seven parts to the attribute list. The first and second parts are the "" ends the attribute list. Attributes are specified as part of the start-tag. \ would define a bullet list because the type of the list is not specified and the default is bullet. \ would specify a numbered list. \ would specify a dashed list. There are a wide variety in the types of attributes. The example above is only to give an idea of some of the uses of attributes. A full description would be longer than the entire FAQ. Detail: Actually almost everything in the above section is not true. It is a vast oversimplification. SGLM is not tied to any character set. So it is imprecise to refer to the actual characters. The actual grammar for the attribute list section is shown in the table below: mdo, Normally An attribute definition list is one or more attribute definitions. The precise definition of an attribute definition list is shown below: attribute definition See below ( ps+ SGML white space required attribute definition)* Options attribute definitions An attribute definition is defined as: attribute name ps+ SGML white space required declared value ps+ SGML white space required default value Refer to the ISO standard for more information. 3.3. Q: What are Entities? Revised: Apr 29, 1995. A: Entities are one of the most complex topics in SGML. This is only a very brief overview. In general, entities are strings or streams of characters that are read by a parser. There are two general partitionings of entities: internal vs. external and general vs. parameter. Internal entities have their character streams specified directly in their declaration. External entities have their character streams kept elsewhere, and their declarations just point to that "elsewhere". General entities are referenced generally where data is permitted; parameter entities are referenced within markup where data is not permitted. SGML uses parameter entities to provide macro-like replacements in a DTD. For example, in the "list" example above the list of types is far from complete. The list of types can get quite long. Also the list may be different from document to document. A parameter entity would make it easier to change the list of types. Below is the example of the same list using a parameter entity. \ \ References to parameter entities are similar to tags in that they have an opening character string and a closing character string. "%" is used as the opening character. ";" is used as the closing character. In the attribute list for the list tag the list of possible types is a parameter entity reference--similar to a macro call. Notice the space between the "%" and "listtypes" in the entity declaration. This space is mandatory for parameter entity declarations. (Otherwise, the "%" and the name together would be interpreted as a reference to a presumably already declared parameter entity, and would probably produce an error.) Entity declarations are unusual in that is not an error to declare the same entity name several different times. But only the first definition is used. It is important to remember that the first definition is what counts. (Lots of clever uses can be made of this peculiarity of entity declarations; this FAQ is not the place to discuss them.) The names of general entities and of parameter entities are tracked separately. You can have one of each kind of entity with the same name. But this is a bad practice since it can become very confusing. The "listtypes" entity in the example above is an internal entity; its character stream ("bullet | number | dash") was directly given in its declaration. External entities allow a document to incorporate by reference an external file or other externally stored character stream. (SGML doesn't prescribe how these streams have to be stored.) The are handy as reusable parts of documents. DTDs, for example, are usually stored in separate external entities and each document usually incorporates its DTD by reference. This makes good sense because you often have lots of documents that use the same DTD. 4. Miscellaneous This section is dedicated to the obscure questions that sometimes come up on comp.text.sgml. They are included here just to appease my warped sence of humor. 4.1. Q: Could anyone tell me a way to set paths in parameter entities? Revised: Apr 29, 1995. A: Many SGML tools now implement an "entity catalog". This tells the parser, and any other SGML aware application, how to resolve references to external entities. Unfortunately many tools do not implement an "entity catalog". Also there is no standard as to what the calalog should look like. The following method gets will work. \ \ \ \ %getdoc; 4.2. Q: I'd like to know more about SGML, and what you can do with it. For example, can I write documents in it in a typesetter independent way, so that I can give the SGML to someone with nroff and someone with LaTeX (along with appropriate macros :) and they can both get the 'same' document? Revised: Apr 29, 1995. A: That depends on what you mean by "same" document. With different typesetting systems it highly unlikely that the end pages will be identical. But the content should be the same. Actually with a properly encoded SGML document you should not care what typesetting system they use. That should be transparent. This is one of the selling points of SGML. 4.3. Q: Is there Web page for SGML? Revised: May 2, 1995. A: Yes, http://www.sil.org/sgml/sgml.html. The web page is copyright by Robin Cover. 5. Parsers 5.1. Q: What parsers are available: Revised: Apr 29, 1995. A: There are numerous freeware and commercial parsers for SGML. This section is far from complete. If you know of any parsers that should be added to this list please send me email. 5.2. Q: ARC SGML Revised: Apr 29, 1995. A: This is the "father" of all the freeware parsers. It was originally written by Charles F. Goldfarb. It is now obsolete and has been replaced by sgmls. 5.3. Q: SGMLS Revised: Apr 29, 1995. A: sgmls is an SGML parser based on the arcsgml parser materials which were written by Charles F. Goldfarb. Executables, formatted documentation and source code are available for anonymous ftp from ftp.jclark.com in the directory /pub/sgmls. 5.4. Q: SP Revised: Apr 29, 1995. A: SP is an SGML parser based on the sgmls and ported to C++. Executables, formatted documentation and source code are available for anonymous ftp from ftp.jclark.com in the directory /pub/sp. 5.5. Q: YASP Revised: Apr 29, 1995. A: YASP stands for Yet Another SGML Parser. Other than that I have no information on hand about it. 5.6. Q: XGMLS Revised: Apr 29, 1995. A: XGMLS is a commercial parser available from Exoterica. Exoterica can be contacted in the US and Asia info@exoterica.com or Europe on exoterica.eur@utopia.fnet.fr. Ph: USA 1-613-722 1700. France 33 1 45 29 18 00 5.7. Q: Where can I get a Lex and Yacc grammar for SGML? Revised: Apr 29, 1995. A: The correct answer is you cannot. There are two reasons. First a fully conforming SGML parser requires a lexer that allows changes to the character set. This is beyond the capabilities of Lex (or Flex). Even with a custom lexer, the grammar for SGML is abiguous. (That alone says a lot about SGML.) Detail: Even though you cannot build a fully conforming SGML parser using Lex and Yacc you can do a lot. The Text Encoding Initiative has done some work with Lex and Yacc. You can find there work at: ftp://ftp-tei.uic.edu/pub/tei/bison/p2bnf.y and p2bnflex.l. C. M. Sperberg-McQueen has also done some work on a DTD parser which can be found at: ftp://ftp-tei.uic.edu/pub/tei/grammar/dpp/* 6. Converting Data to SGML 6.1. Q: If SGML is so Wonderful How Do I Convert My Data to SGML? Revised: May 2, 1995. A: There are many ways to convert your data to SGML. A lot depends on how much data you have, and how soon you need it converted. The best solution, is not to convert it at all. Instead, take the time to re-write the data using the structure imposed by SGML to improve your existing documents. Taking the time to rewrite the documents will improve their quality. Unfortunately few people take this option. If you have a small amount of data, it is probably easier to convert it by hand. But what do I mean by a small amount of data? Well if you can put all your data into a single file, and pull it up into your editor, it is a small amount of data. For me, that means 40 meg of data, is a small amount of data. I almost routinely edit 20-40 meg files. But the capabilities of your system and editor may vary. If you have more than a small amount of data you need a procedure to convert the data. The best procedures are ones that someone else has already developed (why continually re-invent the wheel). The Rainbow DTD and its various tools may be a good place to start. 6.2. Q: What Should I Do If I Cannot Find A Tool To Do The Conversion For Me? Revised: Apr 29, 1995. A: One option would be to give up and use something less painful than SGML. Another option would be to find a firm that will do the conversion for you. (Shameless plug: I work for such a firm, but I am not allowed to mentions its name.) The third option is to write the tools yourself. 6.3. Q: Are There SGML Aware Conversion Tools? Revised: Apr 29, 1995. A: There are numerous SGML aware conversion tools. The three most common are: OmniMark, TagWrite, InContext, and ArborText's PowerPaste. I have tried to use all three of them. OmniMark gets lots of praise from many SGML consultants. However, I have been unable to ever get OmniMark to run on any machine in our building. According to the Exoterica, the manufacturer of OmniMark, no one else has ever reported this problem. OmniMark causes every machine on our network to completely lock up, requiring a power-off (a hardware reset does not un-wedge the program). So caveat emptor, make sure this program runs in your environment before you buy a copy. TagWrite tries to give a WYSIWYG interface to conversion. It is slow, and cumbersom to use, and limited in power and flexability. The documentation needs work. InContext makes TagWrite look fast. You need to know a lot more about SGML to use InContext. If all your machines are Pentium 90's or better with lots of ram (more than 16 Meg), with big fast hard disks, you may actually be able to convert some data with this tool. I know nothing about ArborText's PowerPaste, except that "it provides a unique approach to legacy conversion that includes both interactive and batch processing." I know there are other SGML conversion tools. But so far, based on my own experience, you probably want to avoid them all. They are not worth the time and effort to get them to run. 6.4. Q: What Tools Should I Use? Revised: Apr 29, 1995. A: Converting data to SGML is not a unique problem. The conversion is just another standard problem in compiler theory. Anyone who has takes some college level courses in computer science should immediately recognize that the data conversion problem is the same problem that faces the writer of a compiler for a computer language. So why re-invent the wheel, just use one of the standard tools that have been invented for this very purpose. Lex or its cousin Flex are very good choices. They require a programmer, but then so do the SGML aware tools (they just try and say you don't need to know how to program to use them, but they lie). Lex/Flex have the advantage that they are well known tools that have stood the test of time. Since Lex/Flex produce programs that are actually compiled, the resulting conversion program are fast. There are other tools in the same family as Lex and Yacc that can do the conversion. Coco/R is a good choice is you prefer the ALGOL/Pascal/Modula-2/Oberon/Oberon-2 family of programming languages. Lisp and Scheme provide the higher level tools to do complex manipulations that may be required on complex data. There are Lisp-to-C translators. So it is possible to create hybrid systems that use Lisp for the top level control with Lex/Flex as the Lexer/Reader. This combination is definitely for the power user. If you have lots of data, or the conversion will be on-going, it is probably worth the effort to write a custom translator for the data. Using off-the-shelf techniques found in any undergraduate computer science course it is easy to write a recursive descent parser that should do the conversion quicky and efficiently. 6.5. Q: Does This Mean It Is Easy To Convert The Data? Revised: Apr 29, 1995. A: No. Just because the data has been converted, does not mean it will validate. It may take several interrations to get all the data converted and validated. Then just because the data validates, it does not mean the resulting SGML file is correct. It is very easy to create systematic errors in the data that will haunt your for months. That is why the best way to convert the data is to rewrite it from scratch. 7. Printing SGML This section is yet to be written. Possible sub sections should include: ArborText's Adept*Publisher, LaTeX, Plain TeX, Troff, Interleaf, Quark Express, Ventura Publisher, PageMaker. Do I have any volunteers? 8. Utilities This section is yet to be written. Possible sub sections should include: Author/Editor and ArborText Adept*Editor. Do I have any volunteers? 9. HTML This section is yet to be written. Possible sub sections should include: relationship of HTML to SGML and validating HTML. Do I have any volunteers? 10. Publications This section is yet to be written. Possible sub sections should include: TAG. Do I have any volunteers? 11. Vendor Information This section is yet to be written. Do I have any volunteers? 12. Consultants This section is yet to be written. Do I have any volunteers? 13. SGML Archive This section is yet to be written. Do I have any volunteers? 14. Bibliography This section should include a short list of books on SGML. "The SGML Handbook" by Charles F. Goldfarb published by Oxford University Press; 1990; ISBN 0-19-853737-9 "Practial SGML" 2nd edition by Eric van Herwijnen published by Kluwer Academic Publishers. ISBN 0-7923-9434-8. "SGML: An Author's Guide" by Martin Bryan published by Addison-Wesley. Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 30 Jun 1995 14:00:53 +0300 From: Katri Kristiina M{kinen \ Message-ID: <199506301100.AA29820@tukki.jyu.fi> Subject: constant in DTD Hello, I am writing a DTD for my documents and I would be grateful if someone could give me some advises about some problems that has occurred. My document always contain the same few sentences in certain places. So, is it possible to define them in DTD as constant text that doesn't need to be written while writing text of the document? They would instead be automatically put into the document. What is the best solution (perhaps ENTITY definition?) and how should it be carried out in DTD? Could you please give an example of the solution? Kristiina Mäkinen (kkm@tukki.jyu.fi) Newsgroups: comp.text.sgml From: ghouston@actrix.gen.nz (Gary Houston) Subject: Re: SGML and double-byte character sets Message-ID: \ Sender: news@actrix.gen.nz (News Administrator) Organization: None Date: Fri, 30 Jun 1995 12:11:23 GMT References: <3scjuk$grb@grok.provo.Novell.COM> <19950624T140535Z@naggum.no> <3smo0d$3u0@news.Belgium.EU.net> <3spuii$f3u@hopper.acm.org> X-Nntp-Posting-Host: ghouston.actrix.gen.nz In article <3spuii$f3u@hopper.acm.org>, \ wrote: > >A character (at least for ISO 8879) is an abstraction; a character set is >an association of bit-strings with characters, and we associate small >nonnegative integers with the bit-strings. There are two character sets >involved with SGML documents: the document character set, in which the >document is encoded, and the syntax character set, whose only function is >to associate small integers with characters so that the concrete syntax >can talk about them conveniently. The two are kept separate so that >one can, for example, use the reference concrete syntax even though the >document character set is EBCDIC--but the SGML declaration that references >that syntax is written in EBCDIC. Using ISO 646 as the syntax-reference >character set simply means you can still refer to a carriage return as 13 >and a line feed as 10, etc., *when describing the concrete syntax*. > >>Note 1 after production 171 in Clause 13 states that "The SGML declaration >>is intended for human consumption (in printed form!) as well as for machine >>processing, as it enables the recipient of a document to determine whether a >>system can process it "as is", whether character translation or other algorithmic >>conversion is needed ...". > >Precisely--you have to be told the character set before you can read the >SGML declaration. We can wish it were otherwise, but it ain't. Why wish it were otherwise? An SGML entity is a sequence of lines of text composed of characters, which is an abstract concept not unique to SGML. It only becomes a problem when the physical instantiation of this abstract concept is deficient. I guess with a database the problem would be easily solved. Maybe OpenDoc and OLE are the mandated solutions for the Apple and Microsoft worlds. So what should be used in a Unix filesystem? I'm starting to prefer an idea from Erik Naggum last year: use a magic header to indicate that a file contains text (although he seemed to be thinking of ISO-2022-like switching scheme?). For the sake of example, how about: *text*ISO_8859-2 -- encoded ASCII characters terminating with a LF character. The data following this header would be interpreted as encoded ISO_8859-2 characters. The character set name would need to be part of the file, not just in some out-of-band storage, otherwise it would be lost through a Unix pipe. If a programming language supported something like this, then opening a file in "text-mode" and reading a single character would produce the first character of data with a known character set. The file header would not be presented. Opening the file in "byte-mode" would allow the header to be read. Don't let this get out to any Unix newsgroups, I think it would be considered heresy. From: Physicians' Online <71154.760@CompuServe.COM> Newsgroups: comp.text.sgml Subject: Re: The Crux of the Problem (was Re: PDF versus HTML) Date: 30 Jun 1995 13:52:17 GMT Organization: via CompuServe Information Service Message-ID: <3t0vih$o8v$1@mhadg.production.compuserve.com> References: <19950625T105426Z@naggum.no> >>now, here's an idea: the document as stored on either end will be in an uncompressed, general format. on the wire, it will be compressed by the protocol engine if requested by the receiver. of course, the compressed image may be cached by the server so as not to waste CPU-time if at a premium. the purpose of this separation is that decompression also takes time, and on a very fast network, having to decompress a document for all viewers will have severe impact on the speed of the browser, especially displaying on X terminals and running on a loaded server system. << Actually, on the PC's, I discussed with Phil Katz (of PKZIP fame) doing just that; creating a set of protocols by which a set of files can be zipped, sent across the modem WHILE being compressed, and reconstructed at the other end. He was favorable towards the idea, but did not have the time to develop it himself. Bart Lidofsky From: L.P.H.Bergstrom@telub.se (Peter Bergstrom) Newsgroups: comp.text.sgml Subject: FMV Grund-DTD / STEP and SGML Date: 30 Jun 1995 14:03:14 GMT Organization: Telub Inforum AB, Sweden Sender: -Not-Authenticated-[7516] Message-ID: <3t1072$4pm@terminus.inforum.telub.se> X-Posted-From: InterNews 1.0.1@pebe.inforum.telub.se Xdisclaimer: No attempt was made to authenticate the sender's name. On behalf of Jarl S Magnusson, Director at the CALS Office at the Swedish Defence Material Administration (FMV), I have the pleasure to announce the release of FMV Grund-DTD version 1.10. It has been published on the net, together with it's documentation (in HTML and Postscript format) and some presentation slides (in Powerpoint format), under the Swedish SGML Users Group home page. At the same time, the SWEDCALS White Paper on the interoperability between the two international standards STEP (ISO 10303) and SGML (ISO 8879) has been published in HTML-format. Here's the addresses: Swedish SGML Users Group home page: http://www.admin.kth.se/SGML/ Location of FMV Grund-DTD version 1.10: http://www.admin.kth.se/SGML/Bibliotek/DTDer/FMVGrund-DTD/ Location of STEP/SGML White Paper: http://www.admin.kth.se/SGML/Bibliotek/Litteratur/whitep/wp.html FMV CALS Office encourage the use of the DTD, and look forward to receive feedback from all parties that read, analyse or use the FMV Grund-DTD. Experiences from use of other content-based DTD's will also be welcomed. Have a nice summer! Peter Bergstrom -- Peter Bergstr\öm Chairman, Swedish SGML Users' Group Celsius Inforum AB internet: L.P.H.Bergstrom@telub.se 351 80 V\ÄXJ\Ö phone: +46 - 470 425 97 SWEDEN fax: +46 - 470 397 96 From: John Czarnecki \ Newsgroups: comp.text.sgml Subject: DC Area SGML User Groups Date: Fri, 30 Jun 1995 10:48:57 -0400 Organization: Express Access Online Communications, USA Message-ID: \ Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Does anyone know if there are any DC area SGML groups? Interested especially in Northern VA. Thanks much, John Czarnecki AERA, Inc. jczarnec@aera.com From: Jacques Deseyne \ Newsgroups: comp.text.sgml Subject: SGML Declaration (Was Re: SGML and double-byte character sets) Date: 30 Jun 1995 14:55:25 GMT Organization: SEMA Group Belgium Message-ID: <3t138t$mt@news.Belgium.EU.net> References: <3scjuk$grb@grok.provo.Novell.COM> <3sdti8$7is@news.Belgium.EU.net> <19950624T140535Z@naggum.no>,<3smo0d$3u0@news.Belgium.EU.net> <3spuii$f3u@hopper.acm.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 1.1N (Windows; I; 16bit) davep@ACM.ORG posted a reaction to my answer to Eric Naggum's remarks to my previous posting, etc. As Dave uses the "class, pay attention" metaphor, I am tempted to say that a first prerequisite for being a teacher is to be able to listen carefully to the remarks from the class. [Jacques Deseyne] >>Indeed, from a formal point of view, the term "ASCII" should not be used, >>but rather "ISO 646 IRV". > >That's a red herring. Neither ASCII nor ISO 646 are mandated as the >document character set. What is mandated is that the entire document, >from the initial "\all in the same character set. I never said that they are mandated as the document character set ! Also, it is _UNTRUE_ that from the initial \... >A character (at least for ISO 8879) is an abstraction; a character set is >an association of bit-strings with characters, and we associate small >nonnegative integers with the bit-strings. There are two character sets >involved with SGML documents: the document character set, in which the >document is encoded, and the syntax character set, whose only function is >to associate small integers with characters so that the concrete syntax >can talk about them conveniently... Of course ! The SGML declaration should start with a mdo in the Reference Concrete Syntax, which means that the first data you encounter are the "small nonnegative integers" 60 and 33. They can be encoded as 7-bit, 8-bit, 12-bit, 13-bit, 16-bit, 32-bit or whatever machine size data. If you read my original posting, you will see that I was saying that you can use a "non-ASCII" markup, except in the SGML declaration. This literally means that in the SGML declaration, "markup" (mdo etc.) has to use the small nonnegative numbers with the meaning they have in ISO 646. That's all what I actually wanted to say. Period. I think you are adding to the confusion, as you give the impression that, for instance, the EBCDIC encoding of the " Sema Group Belgium Stallestraat 96 B-1180 Brussels Belgium info-markit@sema.be info-sgml@sema.be ------------------------------------------------------ From: Kai Grossjohann \ Newsgroups: Lee.Althen@daytonoh.attgis.com,comp.text.sgml Subject: Re: MS SGML Author Application Building Experience Date: 30 Jun 1995 17:20:34 +0200 Organization: Universitaet Dortmund, CS Dept, Chair 6 Message-ID: \ References: <1813@wtcd.DaytonOH.ATTGIS.COM> <3sqjvc$g2f@sally.dma.org> Reply-To: Kai Grossjohann \ In-reply-to: melhaven@mail.rcinet.com's message of 28 Jun 1995 03:57:32 GMT To: melhaven@mail.rcinet.com X-Newsreader: (ding) Gnus v0.89 >>>>> "Ken" == melhaven \ writes: Ken> SGML stand alone products are not cheap. I use Emacs together with the psgml package which is a very great SGML editor, and it cost me exactly zero Dollars. How much cheaper can it get? I admit, though, that not everyone wants to use Emacs, but I feel compelled to say this... Also, isn't there a product from MicroStar (Near\&Far) which isn't any more expensive than, say MS Word? \\kai{} -- Life is hard and then you die. Newsgroups: comp.text.sgml From: entrpoop@netcom.com (Larry Larson) Subject: Re: DC Area SGML User Groups Message-ID: \ Organization: NETCOM On-line Communication Services (408 261-4700 guest) References: \ Date: Fri, 30 Jun 1995 16:34:11 GMT Sender: entrpoop@netcom4.netcom.com John Czarnecki \ writes: >Does anyone know if there are any DC area SGML groups? >Interested especially in Northern VA. There's a SIG-SGML in SIGCAT; that's about it. -- Larry From: mjamesb@bnr.ca (M. James Bartley) Newsgroups: comp.text.sgml Subject: Is there a general-purpose telecommunications DTD? Followup-To: comp.text.sgml Date: 30 Jun 1995 18:13:40 GMT Organization: Bell-Northern Research Ltd., Ottawa, Canada Message-ID: <3t1esk$hfu@bmerhc5e.bnr.ca> Reply-To: mjamesb@bnr.ca Originator: mjamesb@bmerh987 We're looking ahead to putting FrameMaker+SGML through its paces and are wondering if there is a telecommunications industry DTD for general product description or operations manuals and the like. I have poked around the FTP directories at Bellcore and there are a number of DTDs there, or at least there are a number of versions of a couple of DTDs, I'm not quite sure which is the case. I am asking around internally but I felt that this was probably also a good place to ask. -- M. James Bartley, Bell-Northern Research Ltd. \\ Internet: mjamesb@bnr.ca P.O. Box 3511, Station C, Ottawa, Canada, K1Y 4H7 \\ Phone: 613-763-3556 ** Opinions expressed are mine, not BNR's. - MJB ** \\ Fax: 613-765-5750 "Stupid pigeons!" From: tad@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: constant in DTD Date: 30 Jun 1995 18:14:07 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <3t1etf$9ai@cliffy.lfwc.lockheed.com> References: <199506301100.AA29820@tukki.jyu.fi> X-Newsreader: TIN [version 1.2 PL1] Katri Kristiina M{kinen (kkm@tukki.jyu.fi) wrote: : My document always contain the same few sentences in certain places. So, : is it possible to define them in DTD as constant text that doesn't need to : be written while writing text of the document? They would instead be : automatically put into the document. What is the best solution (perhaps : ENTITY definition?) and how should it be carried out in DTD? Could you ^^^^^^^^^^^^^^^^^ that's it exactly... \ . . . \\&sentence;\ parser sees: \this sentence appears in certain places\ BTW: my great-grandfather emigrated from Finland. Sisu! -- --------------------------------------------------------------------- | Tad McClellan, Logistics Specialist (IETMs and SGML guy) | Lockheed Martin Tactical Aircraft Systems | voice: (817) 763-6314 | email: mcclellantj@lfwc.lockheed.com --------------------------------------------------------------------- | | All I want, is a little more than I'll ever get. | --------------------------------------------------------------------- From: stevmorgan@aol.com (StevMorgan) Newsgroups: comp.text.sgml Subject: Re: AmiPro Date: 30 Jun 1995 14:54:23 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3t1h8v$rd2@newsbf02.news.aol.com> References: <3strkn$stm@hemuli.tte.vtt.fi> Reply-To: stevmorgan@aol.com (StevMorgan) In a recent review magazine of the new version of Lotus AmiPro, it is noted that the new version will be renamed WordPro and that it will posess full SGML capabilities. The products retail pricing is to be $99.00 U.S. This could be an attractive package and it is worth noting that IBM is in the process of purchasing Lotus. From: jsc@tabitha.atype.com (Jin S. Choi) Newsgroups: comp.text.sgml Subject: HTML and the \

tag Date: 30 Jun 1995 15:03:48 -0400 Organization: Massachvsetts Institvte of Technology Message-ID: \ Reply-To: jin@atype.com X-Newsreader: (ding) Gnus v0.88 I have a question regarding the \

tag in HTML. say I have \ text a \

text b \ My handy dandy SGML normalizer (SP) converts this to \ text a \

text b \

\ I would all like text data tagged as leaf nodes; the above parse leaves "text a" untagged except as part of \. In other words, I'd LIKE to see \ \

text a\

\

text b\

\ since HTML uses \

as a separator (argh! who came up with that?), but that's not what SP parses it to. Is there any elegant, easy way to get what I want or do I have to hack it by hand? (Note that the last example is actually disparaged by the entry for \

in the Annotated HTML DTD!) Newsgroups: comp.text.sgml Approved: erik@naggum.no Date: 30 Jun 1995 16:30:52 -0400 From: Matt Timmermans \ Message-ID: <9506302331.AA0180@notes.microstar.com> Subject: Re: constant in DTD [Katri Kristiina Mäkinen] | My document always contain the same few sentences in certain places. | So, is it possible to define them in DTD as constant text that doesn't | need to be written while writing text of the document? They would | instead be automatically put into the document. What is the best | solution (perhaps ENTITY definition?) and how should it be carried out | in DTD? Could you please give an example of the solution? I assume you mean the your documents are always presented with the same few sentences in certain places. I suggest that if these sentences are part of _every_ document of a particular type, then they are redundant (i.e., they don't count as information) and shouldn't appear in you SGML documents at all. Have your output system generate them automatically. If they always (or even often) appear, but not in the same place, add an empty element to mark there existence -- but not their text. If they often appear with slight modifications, then add attributes or content to that element to contain the modifications only. If, for example, you're a publisher and you send out a lot of rejection notices which are all pretty much the same, your generic "letter" DTD can treat these as a special case: \ \ \me \some aspiring author \ \ Note that in this instance, 'type' is a CONREF attribute. In this case, the output system would produce: me my organization current date Dear some aspiring author, We've read and rejected your book, A Dark and Stormy Night. Please do not call us again. Sincerely, me \ -- Matt Timmermans | Phone: +1 613 727-5696 Microstar Software Ltd. | Fax: +1 613 727-9491 34 Colonnade Rd. North | BBS: +1 613 727-5272 Nepean Ontario CANADA K2E-7J6 | E-mail: mtimmerm@microstar.com From: steveshannon@highnoon.com (Steve Shannon) Newsgroups: comp.text.sgml Subject: SGML Validation Date: Fri, 30 Jun 1995 22:19:29 GMT Organization: InterNex*Tiara ISDN Inter Net Service Message-ID: <3t1tb0$rj9@voyager> X-Newsreader: Forte Free Agent v0.55 Does anybody know of any good SGML validation tools? email is best. Thanks From: davidbern@aol.com (DavidBern) Newsgroups: Lee.Althen@daytonoh.attgis.com,comp.text.sgml Subject: Re: MS SGML Author Application Building Experience Date: 30 Jun 1995 18:28:39 -0400 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <3t1tqn$1ld@newsbf02.news.aol.com> References: \ Reply-To: davidbern@aol.com (DavidBern) >>..."I use Emacs together with the psgml package which is a very great SGML editor, and it cost me exactly zero Dollars. How much cheaper can it get? " I'm a little confused about psgml. Is it written in Emacs Elisp? Or does it run standalone in unix or dos/windows? I'm trying to get started in SGML on the Mac for zero cost, and have downloaded the gnu Emacs port for Macintosh. Now I'm trying to find sgml modules to run with or concurrent to that environment. Any help would be greatly appreciated. Thanks. David Bernhardt DavidBern@aol.com