Hi Marcus, I agree that compatibility reduce possibility to evolve/improve the system. But on the other hand being not compatible you risk to stay alone. I think that everything lyes in clever judgment: how much to sty compatible and how much go into new, fresh and deeper waters. I'm sure it is good not to touch the Smalltalk core things like syntax and this is even not needed, Smalltalk fathers did a really good job to make its syntax eternally useful and feature complete. Libraries are something to evolve, but again very carefully the fundamental parts (collections,...) while other, like GUI, well, here are a lot of opportunities to evolve. Best regards Janko P.S.: Yes, that's how I evolve Aida/Web. Some parts are fixed, while other evolve like crazy. Without much compatibility hassles, so far. Marcus Denker pravi:
One huge problem with compatibility is always that it reduces any possibility in evolving/improving the system. If the goal is to be compatible to e.g. all of Smalltalk X, Visualworks, Squeak, Gemstone.... than, in the end, this means we can not do anything anymore, and, most importantly: we can not do any *fun* things anymore.
Beeing compatible means reducing what you do to the subset of all the dialects, and than stop doing anything.
If I you should choose between a) "inventing the future" and b) "be compatible to VisualWorks", what would you take?
And I personally have already choosen for the "inventing the future" route, I guess. It makes no sense to be in Research (and beeing payed those wonderful tiny salaries) and than do boring stuff. That makes no sense.
I personally think that the possibility and duty of working on interesting things is part of the overall compensation package of people in Research.
Marcus
-- Janko Mivšek AIDA/Web Smalltalk Web Application Server http://www.aidaweb.si