Sublicensing

Daniel Vainsencher danielv at netvision.net.il
Sat Aug 16 10:02:26 UTC 2003


Unfortunately, cleanups and refactorings of existing code don't change
its license. So cleaning up SystemDictionary, for example, by moving out
functionality, doesn't help this aspect (though it sure is nice to be
able to browse SystemDictionary without taking a break for lunch...).

The way to change is to start doing more stuff like the Notifications
work by Roel, that is completely new code. That stuff is categorized
Licenses/SqueakL + MIT on SM, and that's exactly what we need. The ideal
package is a new, from scratch implementation of existing functionality,
so that we can take some SqueakL licensed stuff out of the image,
without losing power. For example, I now notice that ClosureCompiler is 
licensed SqueakL. If we can relicense it as dual under SqueakL+MIT, and 
use it to replace the existing compiler, that would be significant progress.
Anthony?

BTW, anything licensed MIT is compatible with any other license, if only
because it can be freely sublicensed, so that can be the last licensing
decision you're forced to make...

Daniel

Stephane Ducasse <ducasse at iam.unibe.ch> wrote:
> Hi daniel
> 
> I would really like that all the cleaning we are doing, is done under 
> MIT or dual licensing
> I'm really sad about this Squeak-L license. so how do we proceed for 
> that. Should we include in all the changesets we produce a MIT license?
> 
> Stef
> 
> 
> 
> On Friday, August 15, 2003, at 10:12 PM, Daniel Vainsencher wrote:
> 
> > I gave the VM as an example. To end up with a completely free squeak,
> > we'd need to eventually replace all the code we cannot get relicensed 
> > as
> > free software.
> >
> > This will take a lot of time, and somewhat change how we do things, and
> > people need to decide if its what they want. One positive aspect of
> > having to rewrite everything is that it makes it easier to mini-"burn
> > the disc packs". We'd be replacing whole subsystems anyway, so putting
> > some new ideas in there would be easier.
> >
> > So if we look at it really positively, it does have an exciting aspect
> > ;-)
> >
> > Daniel
> >
> > Joshua 'Schwa' Gargus <schwa at cc.gatech.edu> wrote:
> >> On Fri, Aug 15, 2003 at 09:27:14PM +0200, Stephane Ducasse wrote:
> >>>>> Now, this lawyer is not a copyright/opensource expert, and it is
> >>>>> possible we'll find someone more confident about getting smart. 
> >>>>> But I
> >>>>> think we need to start planning on solving this problem by 
> >>>>> organizing
> >>>>> *and by coding*, rather than by PR, lawyers or licensing. I say 
> >>>>> this
> >>>>> quite sadly, because this will not be easy to do :-( OTOH, it 
> >>>>> could be
> >>>>> fun. Anyone care to design a new VM? :-)
> >>>>>
> >>>
> >>> Just  a naive question. If we would come up with a new VM, this would
> >>> solve the problem?
> >>
> >> No.  We'd have to replace all the Collection classes, Magnitude
> >> classes, etc.  Eg: pretty much anything in Squeak up to the point
> >> that Squeak Central left Apple.  And maybe more.
> >>
> >>> What means a new VM exactly how different would it have to be?
> >>> Stef
> >>
> >> Ask a lawyer!  That's a very good question.  Probably not that
> >> different, since this VM is based on the blue-book spec.  Just
> >> as long as code is not copied directly.
> >>
> >> Joshua (IANAL)
> >



More information about the Squeak-dev mailing list