Standardized distributed robotics

Actually, don't you think there are slightly more options than simply "right" and "wrong?" There are multiple options that do not clearly qualify as "wrong." Just as there are multiple options that do not clearly qualify as "right."

Obviously, I don't think I am wrong, but I can see how some people wouldn't choose the approach I am taking.

I looked at player/stage and I like the amount of support it has, but I just saw too much abstraction and not enough modularity.

Much of the design work that I need to do is the sort of stuff I do professionally, so it isn't a huge leap for me.

Oh yes, but that spoils a good argument.

Have you looked at TIPC:

which is now in the linux kernels, and so *might* be useful for providing distributed robot data structures...

cheers, Rich.