In learning about Timed Automata, Coloured Petri Nets, and Process Calculi, I am wondering what the benefit is of having the Petri Net transition be a separate type of node in Petri Nets.
It seems that you can model the same stuff (eventually) using Timed Automata or Process Calculi, which both only have states and transitions (as edges, not objects). Instead of transition nodes, you could just have two types of "state", one be the Petri net place, the other be the Petri net transition. Then this second node does all it's queuing/guarding/checking of the tokens and whatnot. So I'm not sure the benefit is of having transition be a separate object in Petri Nets, when it could just be a second kind of "place" or "state".
That makes me wonder if there is a generalization of Petri nets to arbitrary number of node types (where instead of just place/transition, you have x, y, z, ...). In this sense it seems that process calculi are a generalization of Petri Nets.