Still no new license -- but draft text available

[Tim Peters]

[Henry Jones]
> [Huge amounts snipped.]

[Tim]
>> Whether or not it made legal sense, he was very arguably speaking for
>> the copyright holder at the time, and that's enough assurance for
>> most.  Pity he can't do that anymore, because the bottom line is
>> that people didn't particularly trust the CWI license, they
>> trusted Guido.

> What do you mean by "he was very arguably speaking for the
> copyright holder at the time"?  I presume you are talking about
> CNRI.

Yes, and I think that was clear from the "huge amounts snipped" context that
got clipped.

> This is the crux of my biggest question about Python 1.5.2, and I
> hope you can add it to the FAQ.  Did Guido van Rossum explicitly
> receive permission from CNRI upper level management to release
> Python 1.5.2 under the license/copyright documents included
> with the source code?

Trust me on this:  they'll never answer that, unless it ends up in court.
There's a reasonable chance they'll answer questions about what they
*intend* to do going forward, but retrospective statements can't do anything
for *them* except potentially open more cans of legal worms.  The past is
past; focus on what's *going* to happen.

> This is not an unreasonable standard.  The Free Software Foundation
> suggests for copyright assignment an employee should obtain the
> permission of someone at least at the vice president level or a
> "general manager", although it does say that in a pinch it can be
> "anyone else authorized to license software":  <URL:
> http://www.gnu.org/software/gcc/fsf-forms/assignment-instructions.html>
> It probably would be wise for the rest of the free software community
> to adopt similar standards.

It's certainly not Guido's style to ask for stuff like that, and, indeed,
since moving Python to SourceForge, the requirement for submitting blobs of
boilerplate legalese with patches has been dropped.  Whether that's wise is
open to debate, but it's not a debate Guido will take part in <0.9 wink>.

> Can Guido van Rossum at least say whether he was so authorized?

You've asked that before, right?  If he hasn't answered by now, he either
doesn't want to, doesn't have time to, doesn't think it's *worth* his time
to, doesn't think his answer would help anything, or hasn't seen the
question <wink>.

> If he can't say, can anyone tell me what is the evidence that CNRI
> authorized that license?

As above, since nobody has been saying one way or the other, I suggest it
would be wiser to press for a stmt about what *will* happen.  If you get
assurances you won't get sued (or whatever it is you're worred about), what
difference do the answers to these questions make?  Contrarily, if you get
answers to these questions, and even if they're the answers you wanted to
hear, what assurance does that give you won't get sued (or whatever it is
you're worried about)?  Focus on the result you want in the end; CNRI
clearly isn't interested in debating the issues, or in explaining any of
their past actions.

> In my uneducated mind the biggest problem I have reading the
> 1.5.2 license/copyright document is that the real copyright
> holder CNRI simply isn't listed as the copyright holder.  Instead
> an earlier copyright holder is listed from the years 1991-1995.
> Obviously I am not a lawyer, but this absence makes this license
> bear no resemblance to any other such document that I have ever seen.

It is curious, but since they *are* the copyright holder I don't see that it
matters.

> As far as the Python 1.6b1 license goes, I argue the only real
> question left to answer is whether the license is GPL compatible.
> Has rms made his decision yet, and if so, is there a link where
> he announced it?

CNRI's and BeOpen.com's legal counsel believe the 1.6b1 license is GPL
compatible.  At this point, I don't know whether RMS agrees.

> If Python 1.6b1's license is GPL compatible, there are simply no
> other real problems left.  The license does not make restrictions
> such as requiring source code release of derivative works, it
> does not prohibit commercial use, what other impact could it have?

I spelled out the concerns of some (current and would-be) proprietary users
at length in a reply to one of your other msgs, within the last day.  Python
has a very diverse user base, and we want to keep *all* of them happy with
the license.

> On the other hand, if the license is not GPL compatible, then
> rms may indirectly be drawn in even if he were not issuing his
> opinion.  My impression is that some distributors of Python link
> it with the readline library, a library which is GPLed not LPGLed.

So let's suppose:  As I said above, the lawyers CNRI and BeOpen had look at
this believe the CNRI Open Source License is GPL compatible.  What if the
FSF disagrees?  What then?  Would you stop using Python?  (That's the
"everyone you", not specifically *you*, Henry <wink>.)  There are likely
hundreds of projects using Python released under the GPL now.  Anyone who
would drop Python in this hypothetical case should let CNRI know that!
There's only so much we can do to protect everyone's interests on our own --
people have to look out for themselves too.

>> I'm not sure a readable license is legal <wink>.

> I totally disagree.  Look at the Python 1.6b1 license ...

Like Pat said in reply, I meant "understandable".  After all, if *you*
understood what it meant in all cases, you wouldn't have questions <wink>.
A "good" license to lawyers seems to be one bursting with subtleties beyond
the plain-English meaning of its words.

> ...
> As I have been repeatedly saying, I think the questions simply
> boil down to asking directly two relevant copyright holders,
> CNRI and the Free Software Foundation, and to not blithely assume
> that "silence gives assent".

AFAIK, FSF holds no copyrights on Python source excepting python-mode.el
(Barry Warsaw and I completed the paperwork to assign copyright on that to
the FSF last year, and it did go through).  The copyright holders are CNRI
and CWI, although by all indications CWI couldn't care less about any of
this provided you keep them out of it <wink>.  Certainly agree that people
shouldn't assume silence gives assent here -- indeed, that's why I keep
pressing for people to ask CNRI and express their concerns directly to them.
Until BeOpen slaps on its own copyright and license too, what PythonLabs
(incl. both me & Guido!) says is simply irrelevant from the legal POV.

not-to-mention-legally-incompetent-too-ly y'rs  - tim