Michael,

I know you are talking to Andy, but I would like to add a small thought here.

Since you appear to have an affinity for state machines, you probably have experienced the rapid increase in complexity when designing with concurrent state machines that need to interact with each other. I am not suggesting that the particular project you described is applicable, but in the remote chance that you have not run across them before, I'll take this opportunity to mention that Petri Nets offer a slightly higher level of abstraction than state machines and are generally recognized as being "better" (YMMV) at modelling and implementing concurrent state machines.

Please understand that I am not diminishing the attractiveness of state machines in any way -- I use them frequently. I merely want to make folks aware of "the next small step" to consider when concurrency adds too much complexity to a design based on state machines. I have not seen any reference to Petri Nets on 8052.com before, so I'll bring them up now as yet another tool to have in one's bag-O-tricks.

--
Dan Henry