MPL 2 section 11

[Tzeng, Nigel H.]

Heh...here I attempt to summarize a summarization ...

On 11/26/10 6:43 PM, "Joel Rees" <joel_rees at sannet.ne.jp> wrote:

> On 平成 22/11/25, at 1:54, John Cowan wrote:
> 
>> [...]
>> 4) Therefore anyone can make such a work as long as they have the
>> legal
>> right to distribute the individual sources, which all OSS licenses
>> guarantee.  The GPL's attempt to prevent it as a breach of
>> copyright is
>> therefore unenforceable as a matter of law.
> 
> I'm lost here. Please unpack "he GPL's attempt to prevent it as a
> breach of copyright".

I believe he means FSF rather than GPL.

> Somehow, what you're saying here sounds (to me?) as if you're arguing
> that the GPL is an interpretation of copyright law, rather than a
> license, unilaterally offered in the framework of copyright law, in
> the absence of which the only recourse is fair use (unless other
> license is arranged with the copyright owners, separately, which
> would invoke a separate web of rights entanglements).

My summarization of his summarization in slightly different order is:

1) You have been granted rights to distribute the source code because it's
open source.
2) Putting these source files in a tarball isn't a derived work but a
collection of works that you have rights to distribute either via #1 or
because you wrote it.
3) The compiled version of that tarball collection isn't a derived work
either.
4) Therefore anyone can create such a work because of the above chain even
though the FSF says they believe you cannot. So it is their interpretation
of what is or isn't a derived work is questioned.

Presumably you still need to provide a copy of the GPL'd source to retain #1
but not of the other parts of the collection if they do not require
it...which would result in GPL being a weak copyleft.

Whether 4 is true obviously depends on whether you believe 2 and 3 are true
in the US.  

Whether successfully defending 2 & 3 in court is a pyrrhic victory is
somewhat debatable.  Certainly the FSF would not like that outcome since it
would invalidate the concept of strong copyleft.  Few, if anyone, in the
open source community is likely to actually want this.