The State of Python

[Paul Duffin]

Tim Peters wrote:
> 
> [posted & mailed]
> 
> [Greg Ewing]
> > There's one thing I don't get about this whole
> > license fiasco. What the heck gives CNRI the right
> > to dictate what sort of license BeOpen is allowed
> > to attach to its next version of Python? Surely
> > they can use any license they like as long as it's
> > compatible with the existing Python license.
> 
> CNRI claims that the existing (CWI) Python license isn't a valid license,
> and while that claim makes little sense to me I'm not a lawyer.  At least
> one enthusiastic lawyer on each side of that question is known to exist,
> though, so that would make it a protracted court battle.  We just want to
> get on with our Python work!  It's bad enough that it's taking so long to
> resolve without sucking our ponderous legal system into it.
> 
> Beyond that, there are all sorts of complications, and we *want* this to end
> on friendly terms.  For example, Guido did promise CNRI to produce a final
> 1.6 release for them, and BeOpen agreed to honor that commitment.  As the
> copyright holder of record, CNRI can certainly put any license they want on
> *that* release (the CWI license sure won't stop that, "valid" or not).  But
> so long as we can get them to agree on a *reasonable* (as defined by us --
> and if CNRI isn't asking you, we're the best hope you've got <0.1 wink>)
> license, we'll be happy to ship Python 2.0 too using CNRI's 1.6 license.
> That does look like the most likely outcome right now.  But if a reasonable
> license can't be gotten, at some not-distant point we'll have to just cut
> this off and risk Python's future on the validity of the CWI license and the
> vagaries of the law.
> 
> But nobody on either side *wants* that.  JPython is also cringing in the
> background here, and JPython is wholly & indisputably owned by CNRI (unlike
> CPython, JPython was both born and raised at CNRI).  It would be of real
> value to the JPython community to get a friendlier license for JPython too,
> and CNRI has said they would consider re-releasing JPython under the CNRI
> 1.6 Python license.  That's worth some pain to achieve.
> 
> CNRI also owns the python.org domain and has copyright on that website (btw,
> that's why it's barely been updated this month -- the ex-CNRI folks at
> PythonLabs lost admin access to it), ditto jpython.org, hosts the Python
> Consortium, and runs the PSA as well.  So it's not just the source code for
> Python that's at risk here.
> 
> Of course, CNRI may have a completely different tale to tell about all this,
> buy I can't speak for them.
> 

That sounds absolutely horrendous. Tcl is also going through some major 
changes in its development process to make it more Open Source but it is 
nowhere near as painful as what seems to be happening to Python. I hope
that you get it resolved as quickly as possible.

Who will hold the license on Python 2.0 ? 

Is it going to be BeOpen ?

Could the license be held by the community in some way to prevent this
sort of legal wrangling in future ?

How about having a concept of stewardship, i.e. BeOpen is the current
steward of Python on behalf of the community. 

By Python I mean all Python related resources that BeOpen manage, so 
this would include all information on the web site, the domain name, etc. 

If BeOpen could no longer support Python for whatever reason then those 
resources would be turned over to the community and they could then look 
for a new steward for Python.

The community would also have the option to choose a new steward for 
Python if they felt that BeOpen was not supporting Python properly or
someone else could support it better.

This would have to be a legal contract so for instance if BeOpen was
taken over, or went bankrupt Python could not be sold off to the highest
bidder.

The biggest problem with this is determining what the will of the 
community is.