Python no longer GPL-compatible (Squeak implications?)

Paul Fernhout pdfernhout at kurtz-fernhout.com
Sat Sep 9 13:34:37 UTC 2000


Helge Horch wrote:
> 
> At 18:08 08.09.2000 -0400, Paul Fernhout wrote:
> >One of my concerns [...] is that Disney might decide to do the
> >same thing. Since Disney claims a copyright interest in the current
> >Squeak version, but has not to my knowledge explicitly ever given out a
> >specific license to use their changes, they might insist at a later date
> >that their changes might not be used, or might only be useable under
> >some specific undesirable license.
> 
> Waitaminit, I'm confused.  I thought we were *all* -- Disney or not --
> still operating under and were bound by the Apple Squeak License.  My
> understanding (IANAL) was that, according to Squeak-L's section 2,
> 
> "If the Modified Software contains modifications, overwrites, replacements,
> deletions, additions, or ports to new platforms of: (1) the methods of
> existing class objects or their existing relationships, or (2) any part of
> the virtual machine, then for so long as the Modified Software is
> distributed or sublicensed to others, such modified, overwritten, replaced,
> deleted, added and ported portions of the Modified Software must be made
> publicly available, preferably by means of download from a website, at no
> charge under the terms set forth in Exhibit A below."
> 
> and I always experienced a "recursion flashback" when I read said Exhibit A.
> 
> So yes, the wizards may claim copyright on their respective changes, but
> since they are publishing them according to section 2 under the terms of
> Exhibit A, why do we need to worry?  (Assuming no third-party patents are
> violated etc.)
> 
> [Upon rereading, and unwinding my work image to the base Project:
>    Oh wait, I see -- there it is, right in the MVC startup window:
>    "(c) 1997-1999 Walt Disney".  I never noticed that, and why doesn't it
>    say "portions..."?  Hmm, what's the comment from Disney (and Apple?)
>    on this?  I'm even more confused now, but will let the rest stand as-is:
> ]

I have no problem with anyone claiming copyright to their work which
they add to Squeak. That is their right under the law and is a sensible
thing if authors want to help preserve their rights under copyright law.

One issue hinges on the meaning of "distribute" which is the only thing
that invokes the related clause of the Squeak license.

The other is separating additions from the original. The Squeak license
only applies to changes to the core of Squeak as it was released by
Apple. It is my understanding that for example Morphic is almost or
completely new addition written primarily by the folks at Disney. Thus,
as the Python case demonstrates, its use could be considered by Disney
to simply not be covered by the Apple license. The same for other
additions, especially ones made under Morphic.

Note there are some specific additions which do have a different license
typically in their class comment. One of the web servers has this I
believe. That is a case where the end user knows the licensing status of
what they are using and thus can use it without as many concerns. 

> >[...] In addition, and admittedly stretching things a bit,
> >if Disney argued it never actually officially "distributed" Squeak, in a
> >way like CNRI claims it never formally licensed Python for Internet
> >Download, even the core changes might not be under a specific license.
> 
> I thought so, because they were distributed per section 2.  They'd still
> fall under the original Squeak-L, no?

Again, this hinges on what it means to "distribute" or "sublicense".
However, if such code is considered to be "distributed" or "modified",
then it would fall under the licensing requirements of exhibit A -- but
only at that point.

Again, my point is that if CNRI can claim basically that CNRI never
licensed Python for internet download, why can't Disney do the same?

This gets back to the issue of being explicit in granting permissions
and licenses to use copyrighted works being made open source. The
further issue is that the people granting the licenses have the legal
right to do so on their own (e.g. they aren't for example employees of a
corporation that legally owns all their work and which has not given its
permission for that work to be redistributed.) In the Python example,
CNRI must have owned all his work on Python, so he personally did not
have the right to license it, even though everyone presumed he was the
one to do so. (And as a reminder, beware of such blanket clauses in
employment agreements, even when going to work for a not-for-profit such
as CNRI. Try to at least get an exclusion in advance for open source /
free software work.) 

I think even the GPL has this problem. One could always claim an
addition to a GPL module was never meant for distribution and so can not
be properly used by the receiver. Think what a company would do for
example if an employee working on their own illicitly released source
code developed only for internal use. One ideally needs a specific
statement by the author that this work or extension is explicitly put
under the GPL to address this concern. (How often that happens in
practice is a different story.) 

It is also the case that if a contributor outside of Disney has made
additions to Squeak and said they are under the a Squeak-like License
and then Disney modified theses additions, then the Disney modifications
may fall under a Squeak compatible license -- although again, only when
they are considered to be "distributed" or "sublicensed" by Disney.
 
Open source and free software are still sort of new things to the legal
system. Obviously their are analogues (like creating an encyclopedia or
an anthology), but in general this kind of collaborative intellectual
activity made easy and speedy on a huge scale by the internet is a new
thing. I'm sure in time there will be definitive legal rulings on what
constitutes appropriate notice of permission to use all or part of such
collaboratively developed software and what the legal boundaries of use
are of such software, especially when there is an "implicit" license
based on the behavior of the developers (such as might be construed in
the Python/CNRI case). But, I don't think we are there yet. I am not a
lawyer, and I'm sure one could point out existing cases that attempt to
resolve these issues, but at the very least my point is that there isn't
a clear community understanding of these issues, and to an extent the
intepretation of each situation is probably somewhat subjective and
based on lots of specifics. In the case of CNRI, one might expect years
of development on Python by an employee of a not-for-profit done in a
public fashion with everyone using the result on the basis of a license
from a previous not-for-profit employer would be a clear indication of
an implicit license to use the improvements, but CNRI evidently disputes
this in creating the new Python license.

> >For example, has anyone outside of Disney ever actually downloaded a
> >Squeak release from a Disney server in a way that would be
> >unquestionably considered "distribution" of Squeak by Disney?
> 
> I think not, because they don't?  My understanding is that the Disney folks
> are (like the rest of us) still abiding the original Apple Squeak-L.  I'd
> be surprised to learn that there is such a thing as a "Disney Squeak License."

Again, I don't think the Squeak license covers things like Morphic,
which is mostly or entirely an addition not a change. I can't recall a
post by someone like Dan Ingalls or Alan Kay ever for example saying
"Disney legal has given us explicit permission to put all (or some
specific part) of our work with Squeak under a license like the Squeak
license."

To some extent the ambiguity has worked well. And as long as people just
mainly use Squeak for private use, or restrict the use to within a
company, this is not a big issue. If one wants to make a major
investment into Squeak such as building and ship a commercial product
with Squeak, this is more of an issue. I'm not saying one can't do it,
just that there is a risk which could be reduced by an official
statement from Disney legal. Obviously, many people find the new Python
license acceptable and open enough to keep using Python, despite things
like the GPL-incompatability of specifying a jurisdiction for the
license.
 
> >[...] The fact as I see is that Disney has never formally
> >licensed its changes to Squeak.
> 
> I thought they implicitly had; Squeak Central's changes were under the
> terms of Squeak-L.  Exhibit A again, no?

Again, only if they are considered "distributed" and only to the extent
they modify base classes or the VM of the Squeak distribution Disney
started working from (around 1.13 (based on dates) or 1.16 (based on
lack of copyright)).

> >[...] I am instead hoping this motivates people at Disney to formally
> >clarify the licensing status of distributed modifications to Squeak made
> >by Disney personnel such as at SqueakC. [...] If anyone can point me to an
> >existing clarification of the licensing status of SqueakC contributions
> >from an official representative of Disney, please do so.
> 
> (I of course can neither provide the former nor the latter.  It'd be nice
> to have this settled once and for all though, I agree.)
> 
> Cheers,
> Helge

Thanks for the comments. 

I'm sorry to have to bring licensing issues up again. Last time there
was a licensing discussion on this list some objected and asked instead
a vetted summary be posted to avoid confusion -- which is probably a
very good idea, assuming any IP lawyers out there want to review this
issue and post an official pronouncement. I'm happy to let this wind
down on the list -- but I still think it is an issue an should be
addressed.

-Paul Fernhout
Kurtz-Fernhout Software 
=========================================================
Developers of custom software and educational simulations
Creators of the Garden with Insight(TM) garden simulator
http://www.kurtz-fernhout.com





More information about the Squeak-dev mailing list