IP type issues (was Re: [oc] Re: Merlin Hybrid System)

[John Dalton <john.dalton@bigfoot.com>]

> It is my understanding that OpenCores is just that. A place to submit ideas,
> concepts, and designs for work and release into the public conforming to the
> GPL.

I'll second this.

Of course each person participates in opencores for their own reasons.

Personally, I think people are entitled to keep their own ideas to
themselves, though I think it is an approach doomed to failure.
The idea of a bazaar is that you put the idea out there and everybody
runs with it, hopefully progressing much faster than an individual could
anyway.  Natural selection should kill off those who don't share.

Having said that, if you are using another person's work, you are bound
to follow their rules.  If their rules are the GPL, you must put any
extensions to their work under the GPL.

This is where opencores is different to the free software movement.  Ideas
are not covered by copyright, so the GPL does not apply.  Someone can extend
your idea and patent it.  Provided you have widely publicised your idea, they
can't (legally) patent the unmodified idea as it is prior art.  The expression
of your idea IS covered by copyright, so the GPL holds.  For example, if they
reuse your GPLed VHDL, schematic, PCB design, documentation and so on, their
modifications to these items must also be put under the GPL.  As I understand
it, there is no legal impediment to them starting with your idea, recreating
all your drawings, VHDL, and all other copyrighted items FROM SCRATCH, then 
not releasing their work.

As an aside, some might argue that they need to protect 'free' ideas from
commercial exploitation by using a patent.  I would argue against this.  If the
idea is free, anyone should be free to exploit it commercially.  The control against
profiteering is that anyone else is free to do the same.  Competition should
hopefully ensure that the designer gets to buy implementations of his design at the
cost of construction plus a reasonable profit for the person who went to the trouble
of putting it together.

The thing I would *really* object to is an idea being released, then someone
taking out a patent after the event.  Due to the crappy patent search system,
the patent gets granted and the original designer gets barred from using their
own design.  The original designer then needs to battle through the courts to
prove prior art and win back the right to use their own work.

This highlights a shortcoming with opencores: How do we prove prior art?

Does publication on a web site with a single mirror constitute prior art?
I suspect not, but I am not a lawyer.  We should be prepared to prove
prior art, otherwise one day we are going to get burnt.

Here are some suggestions which might help prove prior art:

1) At the very least have regular backups of the website archived permanantly.

2) Have a much wider mirroring system, which will be less prone to data loss
and hopefully more imune to tampering.

3) We should give serious thought to a 'prior artent' (for want of a better name).
For patentable ideas, the equivalent of a patent should be prepared.  If your design
is fully documented, the documentation might be okay.  This should
be printed on paper, without corrections or erasures.  Multiple copies should be
made and witnessed by independent witnesses.  Each copy should be kept in a safe place
in case it is need in court.  The document should be published as widely as possible
to maximise the number of witnesses who can testify that it exists.


What are people thoughts on these issues?


By the way, I must object to the use of the term 'free IP'.  It's an
oxymoron.  There are good arguments against use of the term 'intellectual
property' let alone putting the word free in front if it.  Opencores
will be doing a disservice to the world if it contributes is to bringing
the term 'Free IP' into common usage.  Let's get another term.
How about something like 'free design'?  (I considered 'free intellect',
but this is a tautology, as intellect is free by nature.)


End of rant

Best wishes
John



Sure, it's tedious
--
To unsubscribe from cores mailing list please visit http://www.opencores.org/mailinglists.shtml