Re: GPL and nib files

[PUYDT Julien]

On jeu, 2003-12-18 at 19:55, M. Grabert wrote:
> IMHO .nib files would qualify as data. I wouldn't have any doubts
> if the file format/encoding was documented and could be used by
> different OpenStep implementations (without modification).
> 
> Data by itself is not a derivative work, and is therefore not
> covered by the GPL.

Well, as I said, it depends of the case at hand...

> If the .nib was created for a non-GPL product, and you use it for a
> GPL-programme, it is obviously no derivative work of the GPL-programme.
> Therefore it doesn't have to be under the GPL-license and doesn't have
> to be included.

I would say that if the file is a must-have to run the program, it
should be GPL-covered too (the file, not the non-GPL product that uses
it too)(there is also the fact that the copyright holder can do whatever
(s)he wants with the sources...).

> > Would your program still compile without the nib? If not, then it
> > probably is part of it.
> >
> > Is the nib something independant from your program? If not, then it's
> > probably part of it.
> 
> To both points: probably, but not neccessarily.
> 
> Let's consider implementing a game that uses Doom3-maps and sound-files.
> It's useless without the maps and sound-files, maybe it even won't
> compile without it. But that doesn't suddenly requires the maps to
> be GPL'd.

Notice that sound&maps are generally data files: ie they're their own
sources!

And the 'game' is in several parts:
* the engine
* the maps
* the sound
==> it makes sense to release only the first part (isn't it what id
software did?)

> So if the .nib were developed without being specifically used ONLY by
> GNUstep, and/or if they were there before the GNUstep-application was
> being developed, you might consider them a non-derivative work.

Yes.

> > On a similar note, I seem to remember that D.Colnet was a little
> > hesitant to release the now-called smarteiffel compiler under the gpl,
> > and releases the sources as C sources. RMS then asked him if those where
> > really the sources he worked with, the answer was no: he was using the
> > eiffel sources, and the se compiler turned them into C, that gcc turned
> > into an executable. We now have those eiffel sources, which are the real
> > sources.
> 
> Most people misunderstand the GPL. The GPL doesn't require/force you
> to re-release your proprietary code into GPL if you violate it.

Well, if I understood well, he wasn't really sure what to consider
'sources', and was still a little hesitant about the GPL. Seems he was
finally convinced ;-)

The copyright holder isn't bound by the GPL, since it is based on the
copyright.

> It just says: if you are distributing GPL'd code, you have to make it's
> sources accessible aswell. And if you want to distribute the code with
> any modifications to it (derivative work), the modification must be
> released under the terms of the GPL aswell, otherwise you will violate
> copyright law, since you don't obey the GPL agreement.

Yes.

> By not publishing the source code of you modifications you
> violate the original GPL licence (agreement), and hence you now have no
> longer the right to modify/distribute the original GPL code.

Yes.

> You may be sued by the copyright holders of the GPL-programme, and
> then you might have to pay money (if you don't choose to solve the issue
> by releasing your modifications under the terms of the GPL or by any
> other means, ie. making arrangements with the copyright holder).

Yes.

> However you won't be forced to release your modifications under the terms
> of the GPL by court order. It's entirely up to you how you resolve the
> copyright violation.

Yes.

> So in your example D. Colnet resolved the issue by releasing the
> real smart-eiffel sources under the terms of the GPL. I assume
> he couldn't make any other arrangements that would have suited
> the copyright holder (in this case probably the FSF/RMS amongs others).
> He also could have done nothing and being sued, and then he
> would have ended up paying (lots of) money.

He was the copyright holder, he released the full source, and the once
"smalleiffel" compiler is now "smarteiffel", and the first betas of the
version 2.0 should appear quite soon.

> Max

Snark on #gnustep

> PS: IANAL

PS: I'm not either ;-)