OT: Software implementation of Finite State Machine

[Michael Welch]

My last experience with finite state machines was over a decade ago in my
electrical engineering courses. And as far as I can remember I've never implemented
one in software (just on paper using models of flip flops and the like).

It seems very simple, but I keep running into snags. I thought about using
the State pattern presented by GOF, but my initial searches in the newsgroups
and Google seem to suggest that the State pattern is not really about finite
state machines: (http://groups.google.com/groups?q=im...tt.net&rnum=12)

And anyway for my needs it does seem overly complex.

Does anyone have some recommendations or good links on how to approach the
problem. I know I'll be able to implement it, but I'd like to read up a
little bit on the theory. Primarily because I want to make the code easy
to maintain as the requirements change and events and states are added to
the state machine.

I'm also concerned that even for my simple requirements, my code is already
getting messy and unclear.

Thanks,
Michael

[Michael Welch]

>It seems very simple, but I keep running into snags. I thought about using
>the State pattern presented by GOF, but my initial searches in the newsgroups
>and Google seem to suggest that the State pattern is not really about finite
>state machines: (http://groups.google.com/groups?q=im...tt.net&rnum=12)
>
>And anyway for my needs it does seem overly complex.
>

Well, MSDN (online) has an article that recommends using State pattern for
finite state machines. In addition, one of the pros of this pattern is that
adding new states later on is very simple. http://msdn.microsoft.com/library/de...l/PATTERNS.asp

What do you all think?

Michael