Re: GPL and Copyright Law (Was: Eclipse 3.0 Running ILLEGALY on Kaffe)
On Wed, 19 Jan 2005 15:32:13 -0500, Raul Miller <moth@debian.org> wrote:
> On Wed, Jan 19, 2005 at 12:01:48PM -0800, Michael K. Edwards wrote:
> > The end being achieved is a major factor in finding a "functional
> > interface" for legal purposes.
>
> We're in violent agreement, here.
I agree, violently. :)
> > The GPL is indeed an offer of contract, but it ties standards of breach
> > so closely to copyright infringement that there isn't much room to
> > argue that non-infringing use still breaches the GPL.
>
> True, but the distinction between what's infringing and what's not is
> ... sometimes subject to debate.
Agreed. But the procedures for evaluating the question are becoming
more and more standardized, both within the US and among countries
with quite different legal histories. Litigating a software contract
isn't the crapshoot it was even a decade ago. I don't think it's that
hard to read the tea leaves now with regard to the GPL in the US. Its
ambiguities aren't in the areas where state laws diverge (such as
trade secret protection and the right to terminate in the complete
absence of a term and termination clause). (In a UCITA state, though,
all bets are off.)
> Anyways, as I understand it, the GPL was inspired by a case where the
> author of a program (who just happened to be RMS) was not allowed to
> read modified copies of his own work unless he signed a legal agreement
> which limited his right to further distribute his work.
Bravo to RMS for having the courage to balk. That seems to have been
rare then, and if you ask me, it's rarer now, at least among people
who code for a living. That regular paycheck is seductive and makes
people do and accept strange things.
> As a mechanism for releasing software for public use without exposing
> authors to these kinds of risks (and, thus, encouraging increased computer
> literacy), I think the GPL does a pretty good job.
I think so too. I think it does so largely by enforcing conditions on
derivative works which amount to publication, and thereby removing
trade secret law from play. Once the delusion of value through
obscurity evaporates, most people behave pretty reasonably with
knowledge that belongs in the commons anyway.
> > Canadian case law seems to be similar, and Canadian courts make
> > careful use of US precedents in this area (such as the "abstraction -
> > filtration - comparison" test of Computer Associates v. Altai). I'd
> > be surprised if US and Canadian appeals courts were to reach seriously
> > divergent conclusions on facts similar to, say, MySQL v. Progress
> > Software.
>
> The parties settled out of court.
But not before the judge issued an opinion on the requests for
preliminary injunctions.
> In essence, the only thing the judge decided was that the issues worth
> taking to trial. Given the lack of precident, that's hardly a surprising
> decision.
On the contrary. Judge Saris may not have come into the court
sophisticated about software development practices, but she wasn't
ignorant about what the law required her to decide at that stage of
the game. Much subtler questions of copyrightability and fair use
have been decided in preliminary injunction proceedings. Copious
precedent is available to her in this area, and she knows where to
find it; but she didn't need it to reach the conclusion that a
preliminary injunction on GPL issues was unwarranted.
Remember, Judge Saris's opinion doesn't even include the word
"copyright", nor does it comment on the nature of the GPL as a
contract, because she considered it completely unremarkable. She
evaluated the parties' conduct with regard to their obligations under
the GPL, and concluded that MySQL was neither likely to succeed on the
facts, nor at risk of irreparable harm, nor likely to be more harmed
by lack of injunction that Progress would be by its issuance. She
also implicitly concluded that there was no public interest overriding
these factors (she listed the public interest as the fourth relevant
factor in her preamble).
Had she thought that the GPL should be evaluated under copyright law
standards, or had she thought it remotely possible that Progress had
breached it in such a way as to justify rescission, she would have
been legally obligated to proceed completely differently. She would
have started with an evaluation strictly of the facts of
copyrightability and copying, which are indisputably on MySQL's side
in the absence of license. Then MySQL would be entitled to an
automatic presumption of irreparable harm, which Progress would have
an opportunity to rebut. Then Progress could proceed to affirmative
defenses such as the existence of a valid license for their conduct
under the GPL, which MySQL could attempt to rebut with evidence of
grounds for rescission.
I am not a lawyer. I have not spent enough time in a courtroom of any
kind to know how often a district judge's procedure and opinion
completely fail to conform to the sequence prescribed by a century of
appellate precedent. But I think there are only two probable
explanations: either counsel for MySQL didn't make any reference to
copyright in their complaint, leading the judge to evaluate the
situation exclusively in terms of breach of contract, or the judge
assessed the evidence to be overwhelming that the GPL was an ordinary
contract and that any breach had been cured beyond the possibility of
rescission by MySQL. Either way, the FSF's position on the GPL as a
non-contract license seems to have made no impression on her
whatsoever.
> So, yeah, a Canadian judge would likely decide the same way... but the
> significant thing this says about the GPL is that it's something new.
No, I think it says that the GPL is the same old same old, just with
less money. :)
Cheers,
- Michael
Reply to: