Urgent Question about Python licensing
thomas at xs4all.net
Sat Mar 10 11:02:28 CET 2001
On Sat, Mar 10, 2001 at 08:43:26AM +0000, Dave LeBlanc wrote:
> While reading up on the recent Python conference on O'Reilly's site
> (http://python.oreilly.com/news/pythonday1_0301.html - actually on the
> 'day 2' page), I cam across the following:
[ The python licence hassles.]
> My questions are these:
> 1. What prevents Python and GPL code from being distributed together?
> If I write some Python code that drives a GPL'd app (but which does
> not depend on that app), what prevents me from offering both my python
> code and the GPL'd app on the same medium (disk/CD/web)?
'derives' is a bit of a general and weird term. If your code does not depend
on a GPL library or binary, you can just distribute it. You can also
distribute both a GPL application and a Python-licenced application on the
same medium -- there is no restriction to that, as long as you make clear
what licence applies to what part. What you cannot do is link a
Python-licenced piece of software to a GPL-licenced part, *and* distribute
that resulting binary. Not even using dynamic linking, for some reason.
(Don't ask me, I'm not a lawyer.) I am not sure if you can attempt to import
a module, which might be a C module that is covered, possibly indirectly, by
the GPL. My gut says that that should be just fine, but my gut usually tries
to empty itself when it comes to lawyerese.
> 2. Is it currently so, or will Python end up with, some license that
> requires that python scripts be covered by the GPL?
No. Python is used in too many non-GPL-compatible environments for it to
ever become GPL. It might fork into a GPL and a non-GPL fork, but since the
non-GPL fork can actually be used in all places a GPL fork could be, that
would never actually work. The one exception is currently the
GPL-incompatibility, but it's a rare case and it is being solved.
Most Python developers, if not all, consider the BSD-ish licence a good
thing. And more importantly, Guido does. And absolutely most importantly, at
least since last monday, the Python Software Foundation does. The PSF holds
the rights of all future Python enhancements, and hopefully a good chunk of
the previous ones. And the PSF is a mix of corporations and individual
> 3. Is Python 2.0 covered by that evil CNRI 1.6 license? I avoided 1.6
> mostly because of the license, and thought that 2.0 had "skipped" 1.6
> and used a far more equitable license then the 1.6 one. (1.6's license
> permits the copyright holder to suddenly decide that anyone using the
> covered software must pay a fee (or any other thing they so desire -
> this is only an example), and by virtue of your even so much as having
> read the license are implicitly bound to their unilateral declaration
> - lovely Virginia laws).
Because 2.0 is a work derived from 1.6, it *has* to be covered by the 1.6
licence. The 1.6 licence clearly states that it applies to all derived works
as well. I'm not a lawyer, so I cannot answer your doubts about the fee
clause, but the intent of the core Python development team is to give Python
as free a licence as technically possible. If Python was not already licence
burdened (by the CWI, the CNRI and BeOpen), the Python Software Foundation
would own the copyright to all of it, and would probably opt for a free and
uncluttered a licence as possible.
One of the goals of the PSF is to *fix* this damned licencing issue once and
for all. That means holding the IP of as much of Python as we can possibly
get, and safeguard it against this kind of confusion forever more. That
might mean changing the licence again (once the PSF owns all the copyright,
which might take a long time), it might even mean changing it a lot, but it
means the PSF will do everything it can to keep Python free to use in any
circumstance, without people having to worry about the licence at all.
I'd like to note, though, that the PSF can only succeed in all that if it
has the community's support. It seems to have the support of the attendees
of the Python Conference last week, but I'd like to hope we have the support
of most of the rest of the python community as well :) We (actually the
officers of the board we assigned) are working on the infrastructure for the
PSF where you can actually send questions and suggestions, but until then,
feel free to bring issues up here, or address one of the board members or
executive officers directly. Like me, for example ;)
For more info on the PSF: http://www.python.org/psf.html
Thomas Wouters <thomas at xs4all.net>
Hi! I'm a .signature virus! copy me into your .signature file to help me spread!
More information about the Python-list