>To me, Traits is a much more experimental technology than Anthony's 
>VI4 work (full closures, context stack unwinding, etc.), so I'm a 
>bit wary.

This is exactly right, in my view.  Traits is our research, which we 
happen to be conducting on Squeak.  Full closures was research 20 
years ago, but is pretty well understood by now.

It isn't my intention, nor that of Nathanael or Stéphane, to try to 
get Traits into core Squeak.  We are happy to have early adopters use 
it, and give us feedback, but it is way to early for "prime time".

As for documentation: we are working on it.  There will be a paper at 
ECOOP, and we plan to submit another to OOPSLA, and another to ESUG. 
But we all have other things to do as well as work on this.  In my 
view, the documentation is even more important than the code, because 
if we have published a good description of what the correct behaviour 
is, there are a lot of folks who will help us track down and fix the 
bugs!  The danger is not having a clear view of what "correct 
behaviour" is.

	Andrew