(no subject)

[Laurent Guerby]

Joerg Schilling wrote:


> I don't know what you did understand from PGL, but GPL definitely states
> that if you modify source _and_ make it available in binary form you need
> to undisclose the complete source too (which includes your changes).


My point was the "and make it available in binary", we're in agreement then.
No binary distributed => no legal obligation to disclose the source of 
modifications.
I assume now (I may have missed it from your previous emails, sorry 
about that)
that *a binary was indeed distributed* in your case.

If people have taken your GPL code, modified it and *distributed a binary*
then they will probably be stopped from distributing the binary. Once
you get the court to stop their distribution of the binary,
they might have to pay you some damage and/or set an agreement
with you, like publishing their modification :).

I doubt someone having received the binary can get automatic access
to the source of modifications: imagine that the infringing
company X merged some sources from author A
with your GPL source. A might not have granted X the right
to disclose the source, just the right to use it in a binary
so it would be logical not to disclose A sources because
it would cause undue damage to A on behalf of
to X misbehaviour with your sources.

If X = A, I don't know if a judge has the power to force
the author to disclose the modification under the GPL,
the judge power might be limited to monetary damage payments,
that is definitely something you have to ask to copyright lawyers.

There's also probably nothing GPL specific there, it's just
a generic copyright on software legal problem since software
has two "copyrightable" forms: source and binary, one of those
being not really interesting to authors and readers.
Very interesting case then :).

But that's just pushing "common sense justice" logic, and IANAL :).

If it goes to court and at this point with FSFE and/or EFF and
that you need funding,  I'll be happy to contribute a few euros.

Sincerely,

-- 
Laurent Guerby <guerby at acm.org>