Re: GPL compatibility of DFCL
Steve Langasek Wrote:
> On Fri, Jun 14, 2002 at 01:50:07PM -0500, Branden Robinson wrote:
>> If someone aggregates my MITed work into a GPLed whole, and then
>> someone else takes my MITed work back out, my work *remains under the
>> MIT license*. *It is not licensed under the GNU GPL*. This is NOT
>> the same thing as *permitting distribution under the terms of the
>> GPL*. If it were, then all MITed and BSDed work would become
>> copylefted -- for everyone in the world -- as soon as someone combined
>> that work with a GPLed one. The BSDs, the XFree86 Project, and
>> (reportedly) Microsoft would thus suddenly become candidates for
>> lawsuits by the FSF for GPL violations -- all one has to do is combine
>> something with GNU Emacs, and bang.
>
> This, then, is the point that we disagree on. I understand
> "distributed under the terms of the GPL" to mean that ALL the terms of
> the GPL apply to the work as distributed, including section 2 granting
> third parties permission to modify the work and distribute modified
> versions of the work, without restriction on the KINDS of modifications
> I can make, so long as I comply with paragraphs a) through c) of
> section 2. (I
> maintain that the explanation at the end of section 2 has no power to
> remove a third party's permission to use portions of a work under the
> terms of the GPL if that work was distributed under the terms of the
> GPL.)
Yes, all the terms of the GPL apply to any subset of a GPL-licensed work.
However, you may have additional freedoms granted by the original license.
If the terms of the license don't allow this, then the license is not
compatible with the GPL. (hence RMS's comment that derivative works _can_
be licensed under the GPL -- em. added)
If, for example, I took a GPL-licensed function foo() , and a BSD-licensed
function bar(), and combined them in a meaningful way, the resulting program
would have to be licensed under the terms of the GPL (the more "restrictive"
license with regard to derivative works).
If someone else downloaded the combination, and worked backwards through the
CVS history to return to the original BSD bar() (call it bar-prime), they
have the right under the terms of the BSD license to incorporate bar() in a
proprietary product. In this case, bar-prime() is NOT a derivative work of
a GPL-licensed work, so is not subject to the requirements of the GPL.
In the case of the DFCL, the ENDORSEMENTS clause would have to be phrased as
an additional freedom (When you modify this document, you must remove the
ENDORESEMENTS section, but you may re-apply ENDORSEMENTS which are granted
by the endorser)
>> > If I'm using part of a GPL work, I shouldn't have to worry about its
>> > genealogy before being able to use it.
>
>> I am sorry, but I am afraid you do. Read all copyright notices.
>
> Though you may be correct that this interpretation would be upheld in
> court, I believe such an interpretation is contrary to the spirit of
> the GPL; and as such, if it /is/ correct, a bug report against the GPL
> would be in order. Even reading copyright notices is not guaranteed to
> give you a clear indication of the origins of each piece of the GPL
> composite, so people would be exposed to legal liability if they
> misjudge the lines between a DFCL work and the incorporating GPL work
> when creating derivative works.
>
> Then again, section 2 does say "If identifiable sections of that work
> are not derived from the Program..." Perhaps if I can't identify it as
> being entirely of DFCL origin, I'm free to distribute it under the GPL?
> Then, as with US patent law, my best defense might be to remain
> ignorant.
You're free to distribute it under the TERMS of the GPL. The licensee may
actually be more free than that, and if you (or someone you distribute to)
is willing to investigate the "genealogy" of the work, they may exercise
more rights than you are aware of. In fact, this may be equivalent to
creating a derived work from the original (non-GPL derived) source.
--Joe
--
To UNSUBSCRIBE, email to debian-legal-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: