Python Cannot be Killed

Terry Reedy tjreedy at
Wed Jun 18 17:53:07 CEST 2003

"Tim Rowe" <tim at> wrote in message
news:5g5vevsu4c5rr5d5scks4tfjqe5mpn1h4j at
> On Tue, 17 Jun 2003 14:13:25 -0400, "Terry Reedy" <tjreedy at>
> wrote:
> >Every few months, someone asks for tips on convincing 'management'
> >that Python is as safe a business bet as closed-source, proprietary
> >products.  Here's a point that is sometimes overlooked: unlike the
> >latter,
> >
> >Python cannot be killed.
> If Python were perceived as a threat to Microsoft I would expect
> to produce a "Visual Python" with a snazzy interface, some fairly
> proprietary web tools and subtle incompatibilities with the
> GPL-compatible one we know and love, and integrated with MSIE.
> Unfortunately Sun is struggling to stop them doing this with Java,
> the Python community would, AFAICS, have real trouble finding the
> money to mount an effective opposition.
> I think a false sense of security can be very dangerous.

Your scenario is quite different from Microsoft (or anyone else)
'buying' the CPython codebase and then prohibiting us from using it
after current 1-year licenses expire, which is the analog of what they
reportedly plan to do with RAVAntiVirus.  That they cannot do, (unless
they spent billions to buy off every Python programmer in the world).

I think your scenario would worst effect Pythoneers with managers who
insisted on switching from real Python to Microsoft's version.

I believe Sun's problem is that they specifically authorized Microsoft
to produce and distribute a Windows Java version -- which Sun wanted
MS to do to 'validate' Java as a standard run-anywhere language.  They
got some of the benefit they wanted, but having stepped into the ring
with a bulldog, they have to deal with the downside.  So now they have
to argue in court that Microsoft stepped over a somewhat blurry
technical line in the contract.

If it has not done so, the PSF should probably trademark 'Python' as
computer language name.  With trademark rights, PSF could avoid Sun's
'problem' by refusing to license Python to Microsoft -- even if
asked -- and forgo it 'impremature'.  We are arguably doing well
enough without it now.  Or it could try to write a better contract.
Perhaps sufficient would be specifying that anything bearing the
Python name, with some version number, must pass the corresponding
test suite, just like CPython.  If a licensed version had avoidable
and unwanted (by PSF) incompatibilities, add more unit tests for the
next version to catch them.

Terry J. Reedy

More information about the Python-list mailing list