GPL compliance

[Gary W. Swearingen]

Daniel Carrera <daniel.carrera at zmsl.com> writes:

> On Sat, 2006-01-07 at 15:00 -0700, Gary W. Swearingen wrote:
>> Reread clause 6:
>> 
>>    ...the recipient automatically receives a license from the original
>>    licensor to copy, distribute or modify the Program subject to these
>>    terms and conditions. ...
>> 
>> The original licensor, in one case, is Red Hat.  You receive a license
>> from Red Hat, subject to T & C.
>
> Let me see if I understand what you are saying:
>
> * A gives a work to B under the GPL.
> * B gives the work to C under the GPL.
> * Now C has a license *FROM A* (the GPL) without having to contact A.
>
> Is this what you are saying?

Yes.  And that's what clause 6 says, IMO, though it need not bother, IMO.

>> Nobody's talking about contacting RH; that's why they used a general
>> public license, in this case the GNU one.
>
> I thought that's what you were saying. If that's not what you are
> saying, then what *are* you saying?

Huh?  OK, yes, what I said there is what I'm saying.  :)

>> I believe it to be a sad fact that some developers share your
>> misunderstanding and claim a single-(re-)license (oh, say GPL) and its
>> coverage of an entire derivative when parts of it are owned by someone
>> else who uses a different license.
>
> If parts of the work were under a license that doesn't allow
> redistribution under the GPL then you can't apply the GPL to the whole.
> I don't think most people would be confused about that part. You can
> only use the GPL to cover the entire derivative when all the original
> components were GPL-compatible. I never said that you could grab GPL
> incompatible software and relicense it under the GPL, that's silly.

And I didn't say you said that.  I guess it would have helped to be
more specific.  The usual BSD, X, and most other open source licenses
_are_ GPL-compatible.  (At least they're widely considered so by the
bulk of GPL users including RMS & Linus.)

Note that the GPL claims to cover the entire derivative and says that
the GPL's "permissions for other licensees extend to the entire whole,
and thus to each and every part regardless of who wrote it."  The
first part is true, but you have to consider copyright and licensing
law to see that there's more (or less :) to "cover" than is obvious.
And the second part is only true because it's talking about
permissions, not terms and conditions.  The derived-from IP's license
must permit the same things as the GPL, which X-type licenses do, but
this doesn't mean that the derived-from IP is not still under it's
original ownership of exclusive rights and licensing.

A case (which I just thought of, I hope its OK) that makes the point
is where a GPL licensor uses the "version 2 or later" clause.  (Note
that "a GPL license" may involve more than the words of the GPL's
COPYING file; the license language usually starts near the copyright
notice, with COPYING included by reference.)  Then if version 3 says
something (which it won't) outrageous like "you may remove from the
derivative the text of all other licenses but this one", then the
derivative would still seem to be OK, but anyone deriving from THAT
software who did remove that text would be violating the T & C of the
X-like licenses. I haven't thought about this enough, but since the
X-like licenses wouldn't allow what the GPL allows (in this bad-v3
with v2-or-later-clause case), I guess then these licenses would be
non-compatible.  So legal could become illegal.  Again, very low risk,
but it confirms my long-standing distaste for "or later" clauses.