Licensing Status of Python 1.5

[Tim Peters]

[Henry Jones]
> IANAL, nor am I speaking for any party, but just look at the
> Python 1.6b1 license.  To my untrained eye, there ARE no "priveleges"
> being taken away relative to 1.5 and no new restrictions for
> someone merely using Python in commercial applications and/or
> embedding Python.

I mostly agree, but there is a new requirement, which you identified next:

> There is one passage:
>     "3. In the event Licensee prepares a derivative work that is
>      based on or incorporates Python 1.6b1or any part thereof,
>      and wants to make the derivative work available to the public
>      as provided herein, then Licensee hereby agrees to indicate
>      in any such work the nature of the modifications made to
>      Python 1.6b1."
>
> This seems very broad to me.

You're not alone!  *Nobody* seems to have a clue what that means, and I
personally haven't heard why CNRI wants this clause.  The only thing I can
pass on about it is that Guido said he was assured that his usual "What's
New in Python i.j.k" notes were sufficient to meet this clause.  So, e.g.,
if you believe me when I say that, and believe that Guido was correctly
passing on the true intent of whoever told him that, and believe that that
latter party was also reliable, ... then it certainly doesn't mean anything
like that you have to post a patch.  I desperately wish this clause could be
dropped, though, or failing that made much clearer as to what it's asking
for.

Especially people using-- or intending to use --Python in proprietary ways
are wary of this clause, to judge from the pvt email I've gotten about it.
I can only tell them to ask CNRI.  In the end they seem to be satisfied, but
not in the sense they understand what's required, but in the sense that
*their* legal counsel advises them that if the matter remains vague even
after asking, just about anything they write in response to this clause will
probably stand in court, precisely because the requirement wasn't made
specific despite asking.

If you can take comfort from that, more power to you <wink>.

Note too that, best I can tell, the users of your derivative work are *also*
licensees of CNRI, so that this requirement to "indicate the nature of the
modifications made to Python 1.6b1" also falls on them in *their* derivative
works of yours, and on and on, until, I guess, universal heat death <wink>.
Over time, I have no idea how the author of a derivative N links removed
will have any way to tell which part of the code is still from the original
1.6b1.  Since Python 2.0 will be a derivative of 1.6, this isn't just an
academic mind game:  your derivatives of 2.0 will face it at once.

Of course that's just the way *I* read the license (well, actually, lots of
people do, but ...), and perhaps CNRI intended no such thing.  That's why
we're trying to get a FAQ written, and part of why the license terms are
still under discussion.

> I can find nothing that reads of a restriction such as the GPL's for
> requiring availability of source code.

While I cannot speak for CNRI, I'm certain you're reading that correctly.
Whether the license is *compatible* with the GPL is a different issue, and
both CNRI's and BeOpen.com's legal counsel say that it is.  I'm not sure
whether Richard Stallman agrees, though.

> Of course for these generous terms perhaps we should thank the
> negotiations between CNRI and BeOpen and the input solicited
> from Open Source leaders.

CNRI's Dr. Bob Kahn and BeOpen.com's Bob Weiner have been at this for
months, and both do indeed deserve major thanks!  Eric S. Raymond, Bruce
Perens, and Richard Stallman have gone beyond the call of duty too.  And
there are many others, most of whom will probably never be named.  So, to be
safe, just thank everyone you meet <wink>.  Also the readers of
comp.lang.python!  You're the ones who have to live with the result, and
your many questions have been greatly appreciated (although, as I keep
saying, this is the wrong place to ask them <0.7 wink>).

> To my untrained eye, the additional sections that were added
> between the Python 1.5.2 documents and the Python 1.6b1
> license simply 1) explicitly acknowledge CNRI as the copyright
> holder and give other corroborating information and 2) strengthen
> the disclaimer of warranty or any other such guarantees.

Pretty much, and it's hard to fault CNRI for wanting to cover their own,
umm, bases this way.

> Now these are purely speculations, but one thing that has
> happened in between the release of 1.5.2 and today is the
> passage of UCITA in Virginia.  You can go to
> <URL: http://www.gnu.org/philosophy/ucita.html>
> to read RMS's misgivings about UCITA.  Be that as it may,
> individuals and organizations in states that have adopted
> UCITA have reasonable grounds for taking measures to protect
> themselves from liability.

UCITA is law in Virginia, but CNRI happens to have its headquarters in
Virginia too, and I believe that's purely coincidence.  Since the
interpretation of clauses can and does vary by jurisdiction, I believe CNRI
simply wants their license interpreted under the law they understand best
(which is the law in their home state).

> This is one reason I believe it is important to establish a
> free software community consensus about simply asking copyright
> holders when there are any questions and to not assume that
> "silence gives assent".

Indeed, yes.  Asking is often easier than getting an answer, though, and
getting an answer you're confident about is often impossible when your
question is about what courts will do.  But the law isn't everything, and
neither should it be.  If CNRI hasn't sued anyone yet, do the little the
license asks of you and it seems darned unlikely to me you'll ever get a bit
of trouble from it.

> Besides, every announcement of 1.6b1 I've seen has carried an
> opinion that 1.6 will be released "soon".

As soon as possible, yes, but this ends on a mildly ironic note.  Since
BeOpen agreed to release Python 2.0 as a derivative of 1.6, 1.6 probably
won't be released until both sides agree on *every* word in its license.
The discussions are still ongoing, but they appear to be nearing an end.  If
there's anything anyone can't live with in the 1.6b1 license, be sure to
tell CNRI what that is!  Also if anything is unclear.  I've been collecting
such questions for a FAQ, and want to hear them too, but only CNRI can
answer them.  If you want to know whether it's OK to fish in the King's
stream, don't ask the court jester <wink>.