Patents and revocation clauses
Bruce Sass <email@example.com> writes:
> Is a patent a one way ticket into non-free?
Let's look at it from IBM's side. If we assume that the license author
at IBM is a nice guy, what he is fearing is clearly a scenario like
- someday company B steps forward with an obscure algorithm patent
that IBM unknowingly violated in their original source code.
- if IBM cannot then revoke the wold-be-free license it could end
up being stuck with giving people permission do distribute of
- company B could set their price for re-licensing rights, and IBM
would have no option but to pay.
At least this (given my rather weak knowledge about American IP law)
seems like a plausible scenario, and the wording of the license that
was posted here some time ago seems to indicate that this is indeed
the possibility IBM worries about.
However, this kind of revocation clause is not the only way around
the unknown-patent problem. I like the way the GPL handles the
situation better, and certainly GPL haven't had any DFSG problems
in that respect.
Clause 7 of the GPL simply states that the right to make copies
of the covered program does not apply if there's a patent which
applies to the code and whose owner contests the free distribution
of the code.
Say that IBM released their code under GPL in the above scenario.
Company B might still go to court, and they might succesfully argue
that their patent is valid and applies to the program. However, in the
same moment the judge decides this, >poof< the GPL effectively
vanishes from all copies of the program (or at least all copies in the
same jurisdiction). IBM wouldn't be stuck with having licenced out
infringing code, so company B wouldn't probably get more out of their
suit than if there had been a revocation clause which IBM had