Re: FSM vs. Petri nets for component modelling...

>I'm really enjoying this discussion.

me too!

>How much effort do you want to spend on making these calls POSIX
>compatible? (I, for one, would be willing to go _very_ far in this respect.)

Our runtime libraries for linux and solaris already use pthread and ipc 
libraries calls. Our current problem is that the application's behaviour is 
still sensitive to implementation details in each system, and we must re-design
the low level transition mechanism to fix that. 
Cannot be done for to morrow :-<

> Mmmm, does this reflect a fundamentally different design, or just
>inappropriate syntax?

Fundamental. Some problems are easily expressed with a data-flow style and 
others more easily with a control flow style. And, up to now, we are not able 
to merge the two styles in a same source file.

	Daniel