Re: Non-gcc releases

[Anthony Green <green at redhat dot com>]

On Tue, 2004-04-06 at 21:08, Etienne Gagnon wrote:
> An example situation: as a Canadian [;-)], I do not wish giving away
> my copyright to a non-Canadian institution, regardless of the nobility of
> its causes.  In Canada, we have Copyright laws that I know, and a court
> system in which I have much confidence.  I can't say that I know enough
> about foreign laws and court systems to give away my code to a foreign
> entity.  In addition, if I ever needed to enforce my copyright on some work,
> I have a University to back me. 

Well, it's unlikely I will be able to convince you - even as a fellow
Canadian :-)

>  I personally have no compelling reason
> to give my code away, and I do not ask my students to give me their code
> either.  So, I can understand the reluctance of any contributor to give
> away his code.

>  From your message, I deduce that you are facing one old contributor that
> does not want to assign his code to FSF. 

Well, I can see how it looks that way on the surface, but it's a little
more complicated than that.  I don't want to go into the details until
we're able to make more progress on sorting it out.

>  This is my biggest problem with
> copyright assignment: it often drives away potential contributors, and
> ends up slowing a project.

I agree that this is the biggest problem with copyright assigments, but
I don't feel it's too big a problem for libffi.  Most of the libffi
contributions from the last few years have come from the GCC community,
who are familiar with this process.  And, perhaps I am suffering from a
lack of imagination, but I'm having trouble seeing where we need legions
of new libffi contributors.  I mean, the core work is basically done -
modulo a few clean ups, more ports as they happen and perhaps an
optimization or two.  It's not really an ambitious project where
development resources are desperately required.

> I would reverse the question and answer: Given the permissiveness of the
> current license, I see no compelling reason to require copyright assignment
> on contributions to libffi.

I don't understand how the license has anything to do with this.

AG

-- 
Anthony Green <green@redhat.com>
Red Hat, Inc.