At 22:31 23.05.2001 , Dan Ingalls wrote:

[snip]
>I can see how this may be powerful in some sense.  But if that becomes the 
>focus, then I can imagine energy flowing into StSq2.8 as a fork rather 
>than Sq3.2 (or whatever) as a join.  And also it would feel like more of 
>an alliance if there were an occasional message from the StSq folks along 
>the lines of "we just got the font swap (say) done, and it looks easy to 
>recreate in 3.1 whenever the time is right".

Ok, here's a start. I've got both the semantic model and the module stuff 
running pretty much the way I want it to (functional acceptance tests are 
at about 85%). The SUnit tests for this functionality are at 100%. I have 
loaded the code, and ran the tests, in 3.1 with no problems at all. In 
addition, I have loaded the code, and ran the tests at 100%, in:

- VisualWorks
- VisualAge
- Dolphin
- Smalltalk MT
- SmalltalkAgents
- VSE

I am able to instantiate a semantic definition in any of these dialects, 
open a socket, send the def to Squeak, and install it there. In fact, I can 
do it from any dialect to any dialect. Essentially, I have a firewall, and 
can completely separate my development environment from my deployment 
environment, so that when one crashes, it doesn't take the other with it.

This might give you a little insight into which direction I'm working.

All my opinions are my own - but I ain't really a Squeaker, just a lil' 
ole' Smalltalker...


--
  - Joseph Pelrine [ | ]
Daedalos Consulting
Email:  jpelrine at acm.org
Web:    www.daedalos.com/~j_pelrine

Smalltalk - scene and not herd!