Aryeh Gregor wrote:
This is C-oriented, but the application to MediaWiki
is fairly clear.
Extensions will invariably make function calls back and forth to core
code, and share data structures (= objects). This conventional
understanding is reflected in MediaWiki's README file, which has
stated that extensions must be GPL (citing the above FAQ) for over
four years
<http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/README?r1=11468&r2=11770>:
It's fuzzier for skins, though. It can be almost a php interface, very
similar to the mentioned borderline case.
However, we don't actually enforce this against
extensions, even ones
hosted on
mediawiki.org. (I'd be in favor, but most developers seem
to be against.)
I think that's because they are more liberal licenses, like MIT*. Do we
have any code that isn't GPL-compatible?
*To which extent they can do that is debatable. Maybe they can only
license under that license *some* pieces of their code.
Another question is JavaScript: stuff that people put
in Common.js or whatnot is probably a derivative work by the FSF's
standard, but on Wikipedia it's licensed as CC-BY-SA 3.0, which is not
GPL-compatible
We should probably change on wmf the editnotice on .js pages to require
it to be colicensed as GPL. Specially since user space javascript
sometimes goes into mediawiki.
Andrew Fitzgerald wrote:
> So strictly speaking by their own definition,
shouldn't WordPress be
> licensed under the PHP license?
I consider that completely unrelated. PHP is a platform, similarly as
how you can use a non-GPL program on a GPL kernel. Or write a document
on a GPL text editor without it being automatically open source.
Aryeh wrote:
Similarly, most Linux kernel
developers maintain that drivers written for Linux must be released
under the GPL -- but this presumably doesn't apply if BSD clones the
driver interface, and the driver is written for BSD, and thus happens
to work on Linux too. IANAL, and the FSF FAQ doesn't cover this case,
so I dunno how this works.
It wouldn't need to be under GPL. AFAIK, the case for that is that
kernel drivers usually copy code from the GPL ones. Also note that since
some version, they have apis not available for closed drivers.
I think there are some interesting discussions about this on lkml archives.