Re: [boost] Re: Re: Re: Finite state machine requirements

From: Joel de Guzman (djowel_at_[hidden])
Date: 2002-09-10 09:50:22

----- Original Message ----- 
From: "David Cattarin" <ditto_at_[hidden]>

> David Abrahams wrote:

> >It's worth noting that compile-time and runtime techniques can be used
> >together in the same system with a reasonably smooth interaction between
> >them. Spirit lets you choose at which level you are going to have
> >runtime-dispatched transitions. I don't know whether that applies to your
> >problem domain or not. Using the BGL (or similar) to represent FSMs leaves
> >you entirely in the runtime domain, where transition dispathcing is
> >concerned.
> >  
> >
> 
> I've only had a cursory glance at Spirit, but I have a hard time seeing 
> how I could adapt it for non-parsing uses, such as message handling 
> where messages are filtered. In fact, I think there is a strong division 
> between event generation and state transitioning. In other words, the 
> state machine framework should be independent of how events are generated.

FWIW, there are some people who are interested and are
working on FSMs using Spirit expression template techniques
and in particular techniques used by the the subrule. I can hook 
you up with them if you are interested.

--Joel