Squeak and LGPL

>>>>> "Paolo" == Paolo Bonzini <[hidden email]> writes:

Paolo> This is not a random person volunteering to answer [hidden email]
Paolo> e-mails.  It is the FSF "licensing clerk" that answered me.

And that's still not legally binding in a way that makes me (or apparently
others) feel comfortable.

Here's the only two ways out of this tangle:

1) FSF could *formally* issue a *legal* document that amends the LGPL
specifically for GNU Smalltalk, although I'm not sure that would actually make
any difference, or if it would even be possible for existing code.  This can't
just be "the FSF licensing clerk".  It needs to be officially issued by the
board.

2) The owners of the GST code could *dual* license their source (as I
suggested a while back) so that it would have both the LGPL and a
Squeak-core-compatible license.

Until then, I think we've now clearly demonstrated that FSF is still
hardlining the LPGL on this code, and LGPL would be infectious if attached to
the Squeak core.  This is not acceptable for most Squeak developers.

I wouldn't be so adamant about this if there wasn't any interesting in GST.
But I *can't* look at GST *and* develop for Squeak core (and I do have a goal
of contributing to the 3.11 release somehow).  There's no legal way to do that
right now.  *Any* derived work (arguably including looking at it, and coming
up with something similar) could give the FSF the grounds for an ownership
lawsuit, and that would be very bad to Squeak.  So either the FSF needs to
explicitly waive those rights in perpetuity, or the GST distro needs to be
relicensed.

--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<[hidden email]> <URL:http://www.stonehenge.com/merlyn/>
Perl/Unix/security consulting, Technical writing, Comedy, etc. etc.
See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!

On 3-Feb-08, at 4:07 AM, Randal L. Schwartz wrote:

> Until then, *DO NOT LOOK AT GST*

Gimme a break. Copyright does not protect ideas, only their expression.

Have you ever seen a Hollywood remake of a foreign film? They can do  
that because the ideas in a film are not protected by copyright, only  
the actual footage. Closer to home, SCO's lawsuit against IBM is dead  
in the water, in large part because they couldn't identify actual  
source code that had been copied from Unix to Linux and were just  
handwaving about "methods and concepts."

The famous "clean room" technique that Phoenix used to reverse  
engineer IBM's PC BIOS is a pretty rare case. Smalltalk has a lot more  
design latitude than a BIOS, and there aren't billions of dollars at  
stake.

I'm all for having Squeak on a solid legal and licensing foundation.  
I'm really happy to see the relicensing effort proceeding well, and I  
think it will be great for Squeak to have stronger ties to the larger  
open source and free software movements. But refusing to examine the  
source code of other Smalltalks is just silly.

Colin

OK, so all have been said now.

Very easy, if you want to contribute to both squeak and gnu, publish in
squeak first (MIT for core image please), then port to gnu (relicence or
not it does not matter).

If you want to backport a Gnu library in Squeak (not for core, just an
add-on loadable package), then publish as LGPL explicitely and no
problem at all.

If you want to port some core gnu method to core squeak (But this is
highly hypothetical, and i am not aware of any such example), then take
care. Don't over nor under estimate the problem, exercize your wisdom.

IMO squeak kernel modifications should rarely exceed a few methods, and
most should have good roots in ST-80 easy to exhibit at lawyers face.

So now, we have wasted enough time for an hypothetical problem (never
happened) but potential (who knows), everybody is warned, we have a
clear policy for collaboration.

I vote for this debate to be closed.

Nicolas


Randal L. Schwartz a écrit :