Public Domain Python

[Tim Peters]

[Pat McCann]
> ...
> But there is always risk involved.  I've read of lawyers claiming that
> the GPL is revocable by some theory that I never really understood, but
> I think amounted to a version of the law that definitely exists in some
> places (in the US once, I'd like to know where now) that there is no
> binding contract unless the two parties have exchanged things of value.

The GPL is not a contract, and by deliberate intent:  it explicitly covers
only "copying, distributing and modifying", and explicitly leaves "the act
of running the Program" out of its scope (despite that its clause 12 appears
to be about nothing *but* runtime issues ...).

The ploy here was to try to rely solely on copyright law, which is much more
uniform across jurisdictions (& especially across national boundaries) than
contract law.  Whether the kinds of things the GPL is trying to do can
actually be done under copyright law alone is at least a matter of academic
debate, and without a court case will remain so.

Ironically, my non-lawerly reading of the GPL suggests it would be stronger
if it had a choice-of-law clause, to constrain the possible interpretations
of its fuzzy areas.  In particular, its disclaimers of liability
(constraining the interpretations of which is CNRI's stated reason for
wanting a choice-of-law clause in their license).

legal-schmegal-ly y'rs  - tim