Python 2.0b1 is released!

[Tim Peters]

[posted & mailed]

[Tim]
> Bernhard (and everyone else as concered about GPL compatibility
> as we are!):  Please understand that we've (BeOpen.com) done all that
> we can now -- I'm afraid it's out of our hands.

[Bernhard Reiter]
> I know the situation, but this political question of course still is
> partly in our hands.

It's a "political" situation only if CNRI and the FSF fail to reach
agreement:  please help us make clear to them that resolution is what you
want (provided, of course, that you *would* rather see a peaceful resolution
than a political circus).

> Right now I cannot recommend the use of the new python if this
> situation prevails. I would have peferred not to release python 1.6
> or 2.0b1 and blame CNRI for the situation.  It looks a little bit
> like the people behind the release are not _that_ concerned about
> it anymore.

The situation was more complicated; there were also binding obligations to
produce a 1.6 release for CNRI, and to make 2.0b1 a derivative of 1.6.  We
could have delayed releasing 2.0b1 longer than we did, but once CNRI had
their 1.6 release it was no skin off their nose if BeOpen's 2.0b1 never got
released, and seemingly no skin off the FSF's nose either.

Rather than commit suicide in ineffective protest, we made a deal with CNRI
that they would continue negotiating in good faith with the FSF.  CNRI has
agreed that if they can reach resolution (nobody can promise they will, of
course), and *if* that requires a change in the CNRI license (despite your
apparent assumption to the contrary, that's still in dispute), CNRI will
release (with our help) a Python 1.6.1 under the modified license, and
BeOpen.com will be allowed to make 2.0final a derivative of that instead.

You're free to believe that we don't care, but we're doing the best we know
how.  It is true, however, that our first loyalties are to Python, and we
will not let Python die a martyr's death.

> A non-GPL compatible python is a disaster and this is grave for
> python and BeOpen.

I agree, but the question of GPL compatibility is still in dispute.  The FSF
says it's not compatible, CNRI's laywer says it is.  BeOpen.com hired
another lawyer, and they happened to agree with CNRI's.  I understand that
any issue remotely touching the GPL gets cast in the gravest terms, but in
this particular case it's still just a dispute about the legal
interpretation of words.  Stallman has said that he has no moral objection
to CNRI's choice-of-law clause, his representation has been that he believes
it's incompatible with the GPL in a strict legal sense.  His lawyer agrees
it's incompatible.  Two other laywers don't.  So this is much more a
question of law than of philosophy at this point.  We're relentlessly
encouraging both sides to get their lawyers together to discuss it directly.
It would be helpful if people *told* both sides that they *want* this
resolved.

> I admit, these are hard words. Please do not feel offended.
> My intention here is to stress again, that this is not just a minor
> glitch in the license.

I'm not offended.  I think you're rallying the Forces of Righteousness to
Mortal Combat <wink> prematurely, though.  I personally have no idea whether
the CNRI license is legally GPL-compatible or not (sorry, but Guido is the
only Benevolent Dictator for Life recognized on c.l.py <0.9 wink>), and
there appear to my layman's view of it good arguments on both sides.

You seem to have made up your mind already; that's fine, but I'm not sure
you've heard CNRI's side of it at all.  If you want to, ask them nicely, and
tell them it's *important* to you:  tell them in detail how failure to
resolve would hurt *your* projects.  CNRI doesn't appear to care at all for
debating philosophy, but they have shown real sensitivity to issues that
affect peoples' *work*.  So manifestos probably won't help, but something
like this will:  "I manage GNU project XYZ, and if the FSF says Python
1.6/2.0 is incompatible with the GPL, that will have the following specific
bad effects on XYZ: ________".

> From my (limited) knowledge of the story I give CNRI the blame.

Tell CNRI that too.  They *did* change the license before 1.6 in several
ways, when their lawyer agreed with the FSF's that some wording was
incompatible with the GPL.  On the last point remaining, the lawyers
disagreed.  CNRI pays a lot of money for their legal advice, so it would
have been rude to ignore it <wink>.

> The FSF stood for the same stable goals for a long time.
> They have to make sure that this is not a big testcase to bring in
> the US contract law into free software (which relyed on copyright law).

The FSF doesn't have the power to prevent that, though, short of settling
*now*.  If CNRI and the FSF can't agree now, CNRI will probably let the
matter stand where it is.  Then it's the FSF who will have to take legal
action against a Python user who takes CNRI's word on whether the new Python
license is GPL-compatibile.  So, watch out what you wish for.  Anything
short of a negotiated settlement now is going to hurt someone.  So far in
its 10-year history, the *only* people ever threatened with lawyers for
using Python have been us!  Go figure -- but we'd like to keep it that way.

and-from-my-pov-getting-hassled-by-the-fsf's-lawyers-is-no-better-
    than-getting-hassled-by-microsoft's-ly y'rs  - tim