One way to a new license: Complete re-engineering
Last updated at 3:35 pm UTC on 14 January 2006
Licensing Discussion in 2004
Has Squat eliminated most of the Apple owned code?
From: Craig Latta Sent: Sunday, January 04, 2004 5:41 PM
... can anyone enumerate those portions [owned by Apple]? I suspect I've removed or rewritten just about all of them as part of the Squat work. I think I'm pretty much left with Xerox stuff currently, and that could get redone too (whether or not that's necessary has already been discussed, I won't rehash it here... yet :). (general Squat info at
Re-engineering is the way to go
From: Cees de Groot Sent: Wednesday, January 07, 2004 4:26 AM
As a license for a software package where the idea is to share common code between licensees, enable commercial use, and not exclude anyone from the group of licensees, I think [the current license] is OK. It has some warts, and these warts are the cause that SqueakL cannot qualify for e.g. OSI approval, but it's not too bad (certainly not in the light of the fact that this license was written well before the whole open source thing
really took off). It's certainly better than the legal jungle that will be caused by relicensing.
.. short of a full re-engineering effort (which I personally think has a much higher chance of succeeding than a clean re-Licensing effort), there's not much one can do about [issues with the current license].
Goran summarizes the arguments and proposes a muddle through strategy
From: goran krampe Sent: Wednesday, January 07, 2004 6:33 AM
.. The only real practical problem AFAIK with the current license is that it isn't OSI certified. IIRC the export clause was the problem. That wasn't a problem for the Debian people, they instead got scared by the indemnification stuff. Anyway, these two clauses in Squeak-L thus prevents Squeak from being an option/included in many different contexts.
- Know who all the contributors are. Is Disney for example included?
- "Changing license": The big problem with changing the license is that we need to
- Get them to all agree on a revised license.
My guess is that this includes Apple, Disney, and a whole bunch of private persons that might be trackable using the signatures in the image.
The other venues that have been lurking in the back of my head are:
- "Sublicensing" given the fact that Squeak-L is a bit... loose. IANAL etc so I have no idea about the validity of this, but IIRC Craig had also noted some "possibilities" in that direction. This route is of course shaky, because who knows if the Big Two Players "wake up" in the future.
- "Benefactor": The possibility of some big player stepping in and helping out by putting pressure on The Two Big Players.
- "Jumping ship": We all simply move over to Slate eventually. :)
- "Rebuilding a free Squeak" from the ground up using the work of Craig (squat). I think Craig himself is thinking in those directions.
Ok, please feel free to add to the list, but given these I think "Changing license" will not happen because I can't see Disney cooperating. It would cost them money to evaluate their options and I don't think they are interested in doing even that.
"Sublicensing" might work in theory, but I think it wouldn't result in a clear situation. We would always be looking over the shoulder anyway. The "Benefactor" possibility is probably not going to happen AFAIK. "Jumping ship" is too bold a move, Slate is still just in its infancy.
The only option that looks interesting to me is to actually rebuild Squeak from the ground up - and we (Craig) are already doing that "as we speak".
So finally, my suggestion is to continue with defining Squeak in terms of packages and their proper licenses, continue Squat, meet halfway and then be able to produce images by combining packages and then, simply by choosing packages under MIT or other OSI licenses gradually getting a Squeak that gives us all the possibilities we want.
3 Re-engineering Efforts
From: goran krampe Sent: Thursday, January 08, 2004 8:12 AM
[Responding to Andrews comment that there were no efforts to “start from scratch and rebuild a new Smalltalk”]... There are at least 3 efforts ongoing that I know of:
- Slate (though not motivated by license issues - it is a new "Smalltalk" implementation under MIT)
- Squat (Craig’s from-the-ground-up project with the license issue as a base ingredients AFAICT)
- TFNR/partitioning-the-image (The community's effort of defining the image in clear packages)