Hacker News Reader

Comment by alfiedotwtf | Hacker News Reader

Since you’re in the know, do you have any opinion on Petri Nets?

I’m not the GP, but I think petri nets are awesome. At least they’re a great way to represent and visualize many systems. Statecharts I find tricky by comparison. For a simple example, imagine you’re making a game and the player has to talk to three people in any order before moving to the next stage. this is very easy to represent as a petri net. (you just need three places and have a token in each one represent having talked to the necessary person.) But I think it corresponds to 8 separate states in an FSM (or statechart? not sure). Which just seems overly complex for a simple situation like this

> But I think it corresponds to 8 separate states in an FSM (or statechart? not sure)

In a conventional FSM yes but not a statechart. Statecharts support parallel regions within a single state. You’d have 3 regions for your example: one waiting for each person. There would be 4 states total: one to wait on each person, each in its own region, plus the superstate. The superstate would exit when all 3 “waiting for person <x>” sub states exited, independent of their sequence.

I’m with you on Petri nets though, very helpful for modelling concurrent behaviour.

Thanks to the both of you. I have always just stuck to plain old Control Flow Diagrams, which can be quite tricky to get all the nuances of a FSM. Always been interested to commit to using Statecharts or PetriNets, but I think the usefulness of concurrency tips it over the edge… To Amazon I go

Reply on news.ycombinator.com