Use of GPLed Python extension modules

Patrick Maupin pmaupin at
Sun Nov 23 17:04:55 CET 2003

Daniel Berlin wrote
> On Nov 22, 2003, at 5:33 PM, Rainer Deyke wrote:
> ...
> > Either you are misinterpreting the FSF, or the FSF is wrong.
> Sigh.
> Can't we simply leave legal questions to lawyers and judges?

Sure.  That's what democracy is all about :)

Or the less flippant answer is:  there are several areas of the
law in which laymen can have a legitimate interest and something
to say.  This is true about any subject.

> The FSF is entitled to their view, and one could argue it and expect 
> some chance of success.

Yes, where "some chance" is an extremely small number.

> One could argue the other way, and expect some chance of sucess.
> This is because there simply is no settled law, precedent, etc,
> on the subject.

Repeating this mantra, as you and several others on this thread
have done, simply does not make it true.  Like most canards, this
has a kernel of truth, namely that the FSF itself has not been
in litigation over this issue.  But if you think that this issue
has never been litigated, you have not boiled it down to its
most basic elements and then looked for case law.

As I understand it, the abstract version of the case would boil
down to this:

- A company notices that program A is a useful tool for other
  programs to use, and produces program B, which contains no
  code from program A, but which invokes and uses program A
  at runtime.
- When this company distributes program B, the makers of program
  A sue the company for copyright violation.  (They do not sue
  the users who actually combine the two programs together, because
  a) those users probably have a license to do this, and b)
  that would be bad for business.)

This issue has come up repeatedly and has been litigated extensively,
by companies with big budgets and good lawyers.  I do not personally
know of any cases where the producer of program A has prevailed (absent
literal copying of program A), but I _do_ know of more than one case
where the maker of program B has prevailed (see Sega vs. Accolade for
a start).  Note that in these cases, the producer of program B
has usually copied program A in the process of reverse-engineering
it (which would not even need to happen to use a GPLed program),
and this copying has been ruled to be fair use.

> So saying they are wrong makes you wrong, because they are correct
> in stating that it would have to be decided by a judge.

Judges and lawyers are so expensive that, instead of hiring
a full set of them to decide a particular issue, people and
companies routinely look at similar situations which have come
up before, and examine the outcome of those situations before
a judge, and then decide to modify their behavior (or not) based on
that outcome.  If one did that on this issue, one could reasonably
come to the conclusion that ON THIS ISSUE the FSF is wrong,
and then another could unreasonably come to the conclusion
that the one is wrong, because nothing is ever settled until
it is brought before a judge (as if judges are never wrong and
never overturned).

> >  The FSF has no legal right to restrict the distribution of
> > any software unless that software contains code which is
> > copyrighted by the FSF.
> This is, of course, incorrect, unfortunately.
> For example, the FSF could own the exclusive right to license some 
> piece of code.  Not saying that they do, but they could, and thus, even 
> without being the copyright owner, would have the right to enforce it's 
> license.
> Any of the rights granted by copyright can be licensed without having 
> to transfer the copyright itself.
> I passed copyright law in law school, so i'm at least sure of this much.

I didn't even take copyright law, and I'm sure that either a) you are
deliberately being unreasonably pedantic, or b) you are exceedingly
dense.  If you were to read the OP in the full context of the thread,
you would find that what he was really saying was that if party X
generates a pile of new code all by himself, the fact that the FSF
has some OTHER copyrighted code (and in the context, I'm sure the OP
means either owns or somehow controls this other code) does not
give the FSF any right to restrict what party X does with his own
code, so party X is free to give his own code to whomever he cares
to, under whatever terms he chooses (but under some circumstances,
party X can certainly be enjoined from distributing the FSF's own code
alongside party X's own code).  This is true EVEN IF party X's code
has the ability to make use of the FSF's code.   In my opinion, the
OP made this point clearly and succinctly, and I'm sure that everybody
except you understood him perfectly.

> In addition, DMCA grants them the right to prevent distribution of 
> certain other types of code (code that circumvents effective access 
> controls).

Umm, yeah.  I'm thinking real hard here to try to come up with a
license that is _less_ compatible with DMCA-style "access controls"
than the GPL.  I'm drawing a blank here.  Can anyone help me out?

(Hint:  The GPL is the "anti-DMCA".  Only use it on code you want
to insure is available to anybody, at any time, in any place.)

> > Whether the code may be linked to GPL code at runtime
> > or not is irrelevant.
> This may or may not be true.
> Stating your opinion as fact is not helping.

Yes, I find your deliberate misunderstanding of the OP _so_ much
more helpful.  Throwing in that DMCA red herring was a nice touch
as well.


More information about the Python-list mailing list