Yes, this is a key question. Do you know for a fact that this experiment will fail? I was hoping wxPython was enough different that it'd sort of "stay out of the way" of the Tk event loop. Maybe not though. Is ActiveState's Pythonwin also incompatible with PyGeo then?

The issue is running VPython, not so much PyGeo. A VPython script runs fine when imported into either IDLE or PyCrust. BUT, in both cases - I just tried it - closing the VPython graphics window also closes the rest of the environment - whether it is IDLE or PyCrust. Frankly I'm not too hung about this at the moment. PyGeo is not really designed to be run interactively - though, as I mentioned, I might eventually get there. So its not a PyCrust thing in any case, as of now. And IDLE is one of maybe a hundred text editors that might be suitable for creating and running Python/VPython scripts. Its probably one of the few that would have a problem running VPython scripts, as far as I am aware. And that problem is scheduled to go way, quite soon. I guess I don't see it the issue too much since I never used IDLE for script writing in any case. Hopefully PyGeo will reach an audience where the need to have an environment spoon fed to them to that degree will not be a big issue. Though I do intend to point folks in need to the Scite site, at least until the IDLE issue is resolved. I have, BTW, what I think is a nice VPython distro for Windows done. Picked the brain of the IDLE setup.py, among other, to get there. The self installing executable is is under 400k. The source distro under 300k - tested with VC6. It includes all VPython related files - demos, docs, (and licenses :) ). It places itself politely under a site-packages directory. Happy to send it along to anyone. E-mail at my reply, if interestered. Art