[License-discuss] License incompatibility (was Re: Open source license chooser choosealicense.com) launched.
Bradley M. Kuhn
bkuhn at ebb.org
Thu Aug 29 13:51:06 UTC 2013
I've tried to reply at length below on the issue of license (in)compatibility.
The below is probably the most I've ever written on the subject, but it's in
some ways a summary of items that discussed regularly among various Free
Software licensing theorist for the past decade, particularly Richard Fontana.
Rick wrote yesterday:
> It's a fair, interesting, and relevant question, and I'd really like to
> know the answer.
Most things in policy, unlike science, aren't a technical problem where we
can provide a hypothesis and test the results. So, there probably isn't "an
answer". I've observed that many lawyers often build their careers on
*pretending* that there are answers to questions like this. Then, they
simply bluster their way through to convince everyone. Since IANAL (and,
BTW, TINLA), I don't tend to think that way. I won't pretend license
compatibility is a testable scientific fact like Einstein's theory of
relativity; it's a policy analysis and conclusion, based on what those doing
the analysis think is correct and likely to be permissible under copyright.
> I'd actually be interested in Bradley ... pointing to any caselaw that
> supports [his] view.
So, most importantly, I don't think there has been any litigation anywhere in
the world regarding "license compatibility". Specifically, Jacobsen
v. Katzer didn't consider it AFAICT. And, (speaking as someone who has
either advised the Plaintiff and/or actually been the 30(b)(6) witness for
Plaintiff in all of the GPL enforcement lawsuits in the USA), I'm not aware
of the license compatibility ever coming up in USA litigation. Also, note
that, no one else in this thread has put any "license compatibility" caselaw
forward; it just doesn't exist, AFAIK.
However, even if there were caselaw, I don't think looking at the caselaw
record is somehow the only way to consider these questions. My primary point
throughout this thread is that Free Software projects are regularly concerned
about compatibility (and license proliferation too), for good policy and
"project health" reasons; not because of what some Court said.
I'd suspect everyone to agree that you must meet the redistribution
requirements of all copyright licenses for a given work to have permission
to redistribute. Thus, license compatibility *exists* as a concept because
if you make a new work that combines two existing works under different
licenses, you have to make sure you've complied with the terms of both
licenses. Again, I'd be surprised if anyone disputes that as a necessary
task and requirement.
Where people disagree is about whether or not you actually need copyright
permission at all to create that new work. Some have a theory that it's
virtually impossible to ever need such permission. I'm pretty sure that
can't be right, and there is a lot of caselaw on *this* subject, but the
tests that courts have come up are *highly* dependent on the facts of a
specific set of circumstances for a specific work. (Dan Ravicher's article
the "Software Derivative Work: A Circuit Dependent Determination" is a
seminal source on that subject about the USA situation. Till Jaeger's talk
at FOSDEM, a recording of which I already linked to in this thread,
covered the European side quite well IMO.)
Of course, derivative works analysis has almost *nothing* to do with
license compatibility. It's just that folks love to fall into derivative
works debates because it's an interesting topic, and because those whose
primary goal is to circumvent copyleft as much as possible (and I've found
that's most people who work as "Open Source" legal experts) prefer to point
to the "derivative works issue" as some sort of insurmountable problem that
is therefore the "base case" of every discussion about Free Software
licensing. And, as you saw, this thread descended into that debate too. I
suspect that's what led Luis Villa to have his "oh no, not again" reaction.
Meanwhile, "license compatibility", as a concept, is a lex mercatoria.
(Fontana and others have talked at length about how much in Free Software
licensing are leges mercatorum; ironically, the derivative work question may
be the only central issue that *isn't* a lex mercatoria.) Specifically, no
court anywhere in the world, to my knowledge, has sat down and lined two Free
Software licenses up next to each other and tried to determine if, upon
creating a whole work based on two works under the two licenses, if the terms
of any license was violated and thus the distributor of that whole work
infringed copyright of one party or the other.
Thus, people argue about what a court might say. Some lawyers bluster and
claim they know the answer when they really don't. (This is why I love Till's
first FOSDEM slide: he admits, as a first principle, the Socratic Epiphany
inherent in this type of work.) In the meantime, though, we have to operate,
share code, and (hopefully) uphold software freedom -- with the tools we have.
That's what led me, back when I started working for the FSF in 2000 , to
publish on the FSF's site a list of licenses that rated them on two primary
criteria: (a) whether the license respects the four freedoms or not, and
(b) which are compatible with the GPL and which aren't. That page has been
improved regularly for 13 years: http://www.gnu.org/licenses/license-list.html
Of course, some people disagree with the FSF's conclusions. I tend to think
the license drafter's opinions on such subjects, particularly in a lex
mercatoria situation, deserve much more weight than random opinions from folks
with an axe to grind. For example, if the Eclipse Foundation publishes
compatibility guidance about the Eclipse License, I'm going to trust their
opinion more than my own, as should you. Similarly, I recommend that people
trust that the FSF knows what its own license says and how it's intended to
work. I think ignoring the license drafters view is a recipe for substantial
risk and potential copyright infringement. (The first person to trust, BTW,
even before the license drafter, is the copyright holder(s)! And, I note that
most people who tend to opine on this subject *aren't*, nor do they represent,
copyright holders in any copylefted Free Software works, and many don't
even *use* copylefted software all that regularly. Interesting, that, huh?)
Fontana and I, have, FWIW, spent an inordinate amount of time analyzing and
considering all possible license compatibility theories. (There are roughly
2-3 we've found that pass any sort of logical merit.) It's basically a
standing agenda item that he and I talk about every time we meet. I don't
know how much thinking everyone else in Free Software has done on this topic,
but I'd suspect Fontana and I have spent more time on this question than anyone
else. It's on the order of hundreds -- possibly more than a thousand -- hours
over the last six years. I take Rick Moen's hint that it'd be good if we
wrote more about the subject, but that's a tough thing to fund: we don't have
the huge resources of wealthy (often GPL-violator-defending) law firms and
academic institutions, which is where most writing on these topics comes from.
Rick Moen wrote further:
> And, like Larry, I'm not seeing any overwhelming flood of mailing list
I don't mind extra traffic either, as we can all easily delete threads that
aren't useful. My main concern was that the subthread might be truly
off-topic, but only Luis has indicated that it might be and others have said
they feel it's on-topic, so I'll go with majority there.
Thus, my only remaining objection to this thread is being told personally
that a large part of my life's work is "a fear-mongering infection upon our
community", but I'll just ignore the people on the thread who are claiming
that from now on. :) It certainly isn't the first time (or, even this first
time *this* year), that someone has resorted to name-calling against those
of us who enforce the GPL. At least Larry didn't compare me to Joe McCarthy
like Mark Shuttleworth did regarding the topic of GPL enforcement. :)
More information about the License-discuss