>1. A native code compiler, possibly building on the GNU compiler >backend.

Keith & Co. are working on something in that direction. Using the GNU-backend would be the way to go, I think. (Portability!)

>2. A more sophisticated make facility, so that information about >modules could be stored into a repository (i.e. precompiled modules >resp. classes).

Real precompilation would not easily allow full optimization, but preparsing and analyzing could probably already speed things up quite a bit.

>3. A set of libraries, including graphics, openGL, scientific (FFT, >linear algebra, ...), CORBA, MPI, PVM and other network, so that the >Sather system could be an interesting alternative to languages like >Java or C++.

Especially for C++, we should be able to offer a very good alternative! And with the Java-Bytecode backend that Erik Schnetter said he could contribute (At least I think it was him) We could offer a great alternative here, too!

>4. A GUI library, possibly based on GTK or QT (maybe Motif, Tk does >exist as far as I know.) A matching GUI builder would be nice too.

GTK is on the works, Tk is included already. I still hope we can have a platform independant GUI-API once in the future.

>5. I like the class browser that is included in the Sather >distribution. It could be extended or reimplemented to a complete >IDE. Thanks to the features of Sather this could be far more advanced >than existing Java or C++ tools. >
>6. A CASE or design tool (possibly based on UML) specific to Sather. >
>7. Integrate concurrent programming (i.e. merge the development of >pSather).

Yea, that's the way to go, I think. We could just have Sather include all the concurrent stuff and only offer a compiler switch to enforce serial programs. But this would be a large change to the language specs and will have to wait for some time. Anyway, it will have to be considered thoroughly.

>8. Tools like SatherYacc and SatherLex. > >9. A full-featured Sather debugger, possibly using a frontend like >DDD. > >10. A documentation generator. It might be possible to slightly change >the language so that documentation is more intimately integrate than >in Java or C++.

There is such a thing included already, but it needs some work.

>These are a lot of suggestions, but I think it would interesting to >discuss each one, maybe adding some more, and clearify the existing >ones.

Does anybody bave an idea how we could keep track of ideas like these? If we only keep them in a newsgroup archive, they'll be lost in no time. There should be a possibility on the Sather home-site to browse all the ideas and comments in a comfortable way, so that anybody can see, what is planned and what ideas were perhaps rejected for some reason or put on hold for the moment. Just keeping a set of HTML pages would be an idea, but it would be tremendous work to keep them up to date. Perhaps a system that anybody can contribute directly with the possibility to edit previous entries?