Hi Stephane--

> Now my question what can we really do with the Interpreter. I
> thought that it was used to develop the Squeak VM (debugging it...)
> is it correct?

	Yes!

> what are the limits of the approach (can we access the VM internal
> structure...?)

	You can indeed access the VM internal structure. The main limit of the
approach is speed. :)  And because of that there are some obvious timing
problems when debugging certain things like media processing. But on
contemporary machines (>= 1GHz) a simulated interpreter session is quite
usable. :)

> Does somebody have a version running in 3.4...

	I do.

> because it seems to me that this is key asset of Squeak for exploring new ideas?

	I quite agree. I've made sure to keep it functional and up-to-date ever
since Bijan asked me to demo it last April, and I had to go back to 2.7
to do it. :)  I've fixed some more bugs in it since Anthony's changes
(e.g., snapshotting from the outside). I plan to release them along with
Squat.

	I'd be happy to help. If you haven't already, check out the Swiki page
I wrote about running the simulator:

	http://minnow.cc.gatech.edu/squeak/2465


-C

--
Craig Latta
http://netjam.org/resume
craig at netjam.org