Re: Eclipse 3.0 Running ILLEGALY on Kaffe
On Thu, 13 Jan 2005 17:02:52 -0500, Brian Thomas Sniffen
> Why are copies OK, and derivative works not? I see GPL 2b talking
> about any work that in whole or in part contains the Program.
> Eclipse+Kaffe contains Kaffe, GPL 2 then exempts mere aggregation --
> which this is not. It also exempts separate sections *when
> distributed separately* -- and explicitly covers them when the Program
> is distributed as a whole.
GPL section 2 is all about "works based on the Program", read in the
context of copyright law. Later in that section, it says that "the
intent is to exercise the right to control the distribution of
derivative or collective works based on the Program." Eclipse+Kaffe
isn't a "work based on the Program" because it isn't copyrightable as
a derivative or collective work. In other words, you can't combine
Eclipse and Kaffe, by running Eclipse on Kaffe, and then sue me for
copyright infringement for doing the same.
The "mere aggregation ... on a volume of storage or distribution
medium" bit may clarify the contract writer's intent to some degree,
but it doesn't automatically make other ways of combining two works
copyrightable. To get at the cases the FSF is shooting for, they
would have to use terms of art instead of "derivative or collective
works", and would have to insert far more draconian provisions to
create an action for breach of contract when GPL and non-GPL works are
> That feels like it's written to address this particular case -- or at
> least a tarball containing Kaffe and Eclipse. This is similar enough
> to count, it's just a funny compression scheme.
It's written to address that case, all right; but I think that it's
irrelevant and that the FSF is trying (unsuccessfully) to have it both
ways. By inserting the "mere aggregation is OK" clause, and by
publishing an aggressive stance on other forms of combination in their
FAQ, the FSF tries to control interpretation and distract from the
fact that courts will interpret "derivative or collective works"
according to their meaning in the copyright law context that the
preamble proclaims. Meanwhile, they claim the existence of a "law of
license" outside contract law (with no basis that I can find), trying
to avoid contract law standards such as "balance of harms" and take
advantage of the relative ease of obtaining a preliminary injunction
under copyright law.
As I have detailed previously, the Progress Software v. MySQL case
makes it pretty clear that they don't succeed in avoiding contract law
standards once they actually get into court. Also, the judge in that
case didn't rule on the factual issue of whether Progress Software had
created a derivative work, but did make it quite clear that the FSF
has no more authority on that point than any other expert witness that
files an affidavit.
The question "does linking create a derivative work" won't really be
settled in the US unless and until a new case arises, is adjudicated
in a court of fact, is appealed, and is properly analyzed with
reference to precedent by a circuit court. I think (IANAL) that the
outcome is quite predictable in light of precedents I have cited from
several circuits. Ultimately, linking against the public interface to
a published library doesn't detract from the rights in authorship that
copyright law was created to protect, and US appellate courts are
surprisingly good at sticking to the purpose of copyright in the face
of legal and technical contortions that attempt to leverage a
copyright monopoly beyond its legitimate bounds.