If what you want is something that scales up, then you're attacking the Wrong Problem. Rather than focus on getting your thread overhead as small as possible in order to support as much real-time concurrency as you can, you should (as has been suggested) try to get simulation-time concurrency without worrying about what happens in real-time. Then you can have arbitrarily many things happening at once.
This introduction to discrete event simulation explains the idea: http://www.dmem.strath.ac.uk/~pball/simulation/simulate.html mike _______________________________________________ Tutor maillist - Tutor@python.org http://mail.python.org/mailman/listinfo/tutor