Public Domain Python

[Tim Peters]

[Alex Martelli]
> I don't think anybody is claiming proprietary rights on Python
> proper, i.e., the *language*?

Right, nobody is.

> Copyright on the implementation,

Definitely.

> and maybe trademark on the name in certain contexts.

CNRI has applied for trade and service marks at the U.S. federal level on
the names "Python", "JPython" and "Python Consortium".  They may or may not
assign one or more to Guido (they've said they intend to, but have not said
they *will*, and so far have not).

> But if you reimplemented the language independently (e.g. the way
> Mr Skaller did with Vyper), deliberately "cloning" it, would you
> be running any legal risk...?

AFAIK, none at all, provided (of course) that you didn't "borrow" any of the
copyrighted implementation files from CPython.  Even then, all CPython
licenses to date *allow* you to borrow those files and do what you please
with them; but the licenses tag along with those portions, and then the
legal status gets real messy real fast.

> I thought only Java and ADA were popular languages undergoing that
> level of 'protection'.

I believe that's based on trademark law rather than copyright law.  That is,
nobody can stop you from reimplementing Java, but Sun can stop you from
*calling* it "Java" (or any "similar" name).

> It's easy to get confused between a language and its implementation
> when there's only one of the latter, but, with JPython, CPython, and
> Vyper, not to mention forthcoming developments such as Python.NET,
> it doesn't seem to me that this is the case.  Or are all of these
> versions derived works of CNRI's-claimed stuff?

CPython, yes.  Vyper, I don't believe so.  Python.NET, no (but note that
there's a potential dispute about that *name*, owing to Christian Tismer's
prior registration of the python.net domain name).  JPython is a 100% CNRI
production, and is not related to CPython in any legal sense.