[License-discuss] License incompatibility (was Re: Open source license chooser choosealicense.com launched.)

[Bradley M. Kuhn]

Lawrence Rosen wrote at 17:00 (EDT) on Tuesday:
> I asked for practical examples. You cited none. In the world of
> copyrights or most logical pursuits, absence of evidence isn't
> evidence.

License compatibility issues come up regularly on lots of bug tickets
and threads about licensing on lots of projects.  I don't have
a saved file of evidence to hand you, mostly because it's such an
unremarkable occurrence that I don't note it down when it happens.  I
see it at least once a month somewhere.  I suspect anyone who follows
Free Software development regularly sees it just as much.  I can tell
you that I dealt with two issues of license incompatibility in my
day job recently, but I can't disclose the specifics since it
was confidential advice.

Meanwhile, if you need evidence to satisfy your curiosity right away,
I'd suggest debian-legal archives would be a good place to start your
research on this, but...

> Awaiting your evidence to the contrary....

... I can't spare the time to do this basic research for you.  If
anyone else here does agree with you that license incompatibility
isn't a problem in the real world, then perhaps it's worth continuing
this thread, but I suspect you may be alone on this one. :)

> Most FOSS licenses (including the GPL!)  don't actually define the
> term "derivative work" with enough precision to make it meaningful for
> court interpretation. ...  The court will therefore use the statutory
> and case law to determine that situation.

That's as it should be.  It's not the job (nor is it really appropriate)
for a copyright license to define statutory terms, so the GPL doesn't do
that.  The GPL has always tried to go as far as copyright would allow to
mandate software freedom.  That's what Michael Meeks (and/or Jeremy
Allison -- I heard them both use this phrase within a few weeks of each
other and not sure who deserves credit) call "copyleft's judo move on
copyright".  It's the essence of strong copyleft.

> all major FOSS licenses that I'm aware of *except the GPL* make it
> abundantly clear that the mere combination of that licensed software
> with other software (FOSS or non-FOSS) does not affect ("infect"?) the
> other software.

Indeed, weaker copylefts give guidelines for certain derivative works
that can be proprietarized, and the rest are left under copyleft.

BTW, if you are interested in how the European lawyers view this question,
I refer you to an excellent talk by Till Jaeger at FOSDEM 2013:
   http://www.faif.us/cast/2013/mar/26/0x39/

> So what's the worry about license incompatibility all about?  Is it
> merely that so many licenses are deemed incompatible with the GPL,

Many licenses are incompatible with the Eclipse Public License, too,
since it's a stronger copyleft than, say, the MPL or LGPL.  There aren't
very many strong copylefts around.

> with other software (FOSS or non-FOSS) does not affect ("infect"?) the
> other software.

Finally, I'm unlikely to respond to this thread further as I think the
use of slurs like "infect" (notwithstanding the quotes, and '?') to
refer to copyleft are just unnecessarily inflammatory.  I've asked you
not to talk about copyleft using slur words so many times before in the
thirteen years we've known each other, it's really hard for me to believe
you aren't saying "infect" intentionally just to spread needless discord.
-- 
   -- bkuhn