Python 2.0b1 is released!

Tim Peters tim_one at
Sat Sep 9 05:56:45 CEST 2000

[Gareth McCaughan]
> As I understand it, the situation is this.
>   1 CNRI wants the applicable law specified, to make sure
>     they don't get sued in unpredictable ways.

Right -- or, at least, that's my understanding.  It is what CNRI said in the
license FAQ, specifically relating to the interpretation of disclaimers of
liability.  Which, BTW, the GPL also has (clauses 11 and 12 in GPL version

>   2 RMS wants the applicable law not specified, because
>     as it stands it would have to apply to the GPL as
>     well, and that's (allegedly) something that can't
>     be done.

I wouldn't be surprised if RMS would also *like* the applicable law to be
specified!  His reaction has been much more of a "hmmm" than a "NO! THAT'S
EVIL!" so far.  Perhaps something related will even show up in GPL 3.  But,
for now, GPL 2 is what exists, and he has to defend it for what it says.

I confess I don't see the problem.  GPL clause 6 says "You may not impose
any further restrictions on the recipients' exercise of the rights granted
herein", which is presumably the source of the conflict, and I just don't
see *which* right granted by the GPL is restricted by CNRI spelling out the
intended meaning of their license.  But IANAL, and these licenses are
written in anything but English <0.5 wink>.

> So what's wrong with these possibilities?

You'll have to ask CNRI and/or the FSF.  Nobody on can resolve this.
I'll give you my best guesses, though.

>   a Amend the licence so that it says "The disclaimer
>     of liability in section shall be interpreted
>     according to the law of the state of Virginia" or
>     whatever.

The FSF appears to believe that any choice-of-law provision whatsoever is in
conflict w/ GPL 2, no matter how hedged.

>   b Amend the licence so that it says "This license[1]
>     shall be interpreted according to the law of the
>     state of Virginia, except that this shall not apply
>     to the provisions of the GNU General Public Licence
>     unless it would do so in the absence of this paragraph."

How does this protect CNRI?  They're still the source of the software,
regardless of which license applies.  Liability is liability, and they're on
the hook whether it's their license or the GPL that applies.  Their lawyers
know how to defend the CNRI disclaimers given choice-of-law; they don't know
how to defend the GPL disclaimers without choice-of-law.

>   c Some other intermediate stipulation, like "Paragraphs
>     1 through 12 of this license shall be interpreted
>     according to the law of the state of Virginia" where
>     those are in fact all the paragraphs of the licence
>     (the point being to make it more explicit that there's
>     no attempt to have the GPL interpreted under Virginian
>     law).
> ...

CNRI apparently doesn't want to be in the business of defending itself under
the GPL, period; they didn't write it, and they think it lacks a protection
they need.

I've said in other forums lately that this is much more a question of law
than a conflict of philosophies at this point.  I was happy to see that
Richard Stallman appears to believe that too:

Note also that Guido did some Q&A there a little earlier:

More information about the Python-list mailing list