Public Domain Python

Huaiyu Zhu hzhu at
Fri Sep 29 12:50:12 CEST 2000

On Wed, 27 Sep 2000 22:16:53 -0400, Tim Peters <tim_one at> wrote:

>Of course the copyright holder has the right to change licenses.  GPL, CWI,
>X/MIT, BSD, ... doesn't matter, that's always the copyright holder's
>perogative.  Unsure what "retrospectively" means.

As I said before, the issue is what does "change licence" really mean. Does
it affect copies already distributed and their (possibly future) derivative
works and their (possibly future) distributions?  I'd say that for GPL'd
work, a change of licence could only be applied to the new stuff that comes
with the new licence.  Is this so for every open source licence?  Even under

I'm not concerned if someone changes licence, unless that applies to copies
already distributed.

Here's an analogy.  Suppose I buy books from a vendor.  If he increases
price by ten fold afterwards, that's fine.  But he shouldn't ask me to pay
nine times more for the one I've already bought and read.  So what if he's
so nice and only asks me to return the book and get a full refund if I don't
pay the extra?  What if he's even nicer and says I can keep the book and
read it, but I have to pay extra only when I let my friends read it?  No
way.  His new price should only apply to the copies he hasn't sold, and I
should be able to do anything I could do with that book when I bought it
under the old price.

[Discussion on whether checking into CVS counts as publishing]

OK, whenever lawyers are involved things always get muddier.  IANAL ANWTBO
(I'm not a lawyer and never want to be one).  I'm only viewing this from a
common practice point of view.

To me, if a text file is on the the Internet (for an extended period if you
will) and contains a prominent statement that it is "copyrighted and
distributed under GPL", then this is as good as any statement to the effect
that I could take it and use it the same way I could with something from
FSF.  Unless there is a statement like "this file should not be distributed
alone", in which case I need to check the copyright and licence files on the
whole distribution, I have no obligation to check what is the "whole
distribution" of which it belongs (which is pointless on the Web anyway).

>> With GPL, as long as it is authorized, no one can take it back.  That's
>> sufficiently clear from my pov.
>Sure.  What isn't clear is why you think the GPL is special in that respect.

I'm not sure if GPL is unique in this way, other than that it explicitly
forbids adding any additional conditions not in the time it was copied. Once
you get a copy of a file by any means under GPL, your only obligation is
limited to that specified in the licence which accompanies the original copy.

Well, the problem with Python licence may not be whether it was GPL'd, but
rather who is the copyright holder and what the licence is for each
individual file.  Suppose each file in CVS contains a statement that it is
either put in public domain, or else it is copyrighted and distributed under
CWI/CNRI/BeOS licence, this might be as clear as GPL, I suppose.

If that is the case, I'd advocate every author to mark their new files this
way. One or two comment lines wouldn't affect performance very much, I
suppose, but it would guarantee that everything in Python has an Open Source
licence up to the minute. These should be also be more trustworthy in the
long run than the papers you guys sign with BeOS, although the latter should
solves the authorization problem.

>> Anyway, all this just add to the argument: if the earlier releases were
>> authorized and under GPL, it would not matter what a later new licence is.
>Surprise:  I still see nothing special about the GPL in this respect --
>unless copyright had also been assigned to the FSF.  To the contrary, due to
>some of the points above, I think it would have been weaker than the CWI
>license wrt the status of stuff in the CVS tree.

OK, maybe GPL is not special in this respect.  But from practical point of
view it is quite clear that no one who put some thing under GPL in publicly
accessable CVS tree like SourceForge would ever be able to reclaim it back
under another licence (instead of just releasing another copy with a
different licence).  I take it that it is not clear that the previous CWI
licence has this property, or the way it was distributed has this property.

licence-really-makes-my-head-ache-ly yr's


More information about the Python-list mailing list