Hi dan and others,

I just want to say two words about forking.

I think that the big bang approach we all dreamed a year ago 
(especially me) failed for various reasons
(mainly not enough resources allocated, a complex model and having an 
existing image).
Apparently incremental process is the way to go in such a situation. We 
learned it.

Now what is the most important is not to have model x or y but that 
people realized that we could really have
a mini-image and a build process. That images are good for developing 
and feeling like home but
that clever changeset are needed. I think that as soon as we will have 
a good working system
it will improve gradually, I'm sure that once the basic issues of 
packaging modules
will be resolved (there are already), dependency analysis and 
management, new architecture (such as the already working
file list in 3.3 or the dynamic menu proposed by daniel) will arrive as 
well as versioning and
eventually namespaces.

I still dream about ScrSqueak, SqEtoy and all the other ones we could 
have build on a
mini-core and a building system. I think that the effort of daniel, Avi 
and goran are driven by
concrete problems and this is good. They should continue.

Stef

stop talking and latexing again and again.....