Re: [boost] Re: Finite state machine requirements

From: Scott Woods (scottw_at_[hidden])
Date: 2002-09-05 21:43:01

Hi Andreas,

> I guess the party is not going to happen here anyway, unless you'd call
two
> people (you and me) a party ;-).
>
> The lack of interest for my proposal here at boost made me think more
about
> your approach to FSMs. Maybe everyone thinks a boost FSM should be
> dynamically configurable? Or maybe I just overestimated the utility of
state
> machines in everyday programming (that's what boost libs are all about,
> aren't they)? I don't know. What also made me question my approach more is
> template code bloat. I guess I would carefully investigate the memory

Good to see something happening on your proposal. There are some technical
points I'm keen to respond to but that will have to wait for a following
(more
considered message). In the meantime I just had to get out this quick one so
that you know there is interest out there.

I think FSMs have huge application. I also think that in most commercial
development
circumstances there is not the incentive to apply formal methods such as
FSMs. In fact
a developer may be frowned on (by peers, management or customers) for trying
to
define an FSM for say a turnstile (convenient example but reasonable I
think). Its also
true that developers may "take the shortcut" and avoid the focused thinking
required to
complete a robust FSM, i.e. we can be lazy.

I think the lack of response is due to issues such as those mentioned above.
Sad too
because development of an FSM is often the forum that exposes all sorts of
operational
nasties.

Just about any software system that interacts with a user through a GUI,
interfaces to a
physical device or interoperates with other software components (esp. across
networks)
is inherently asynchronous. These systems *need* FSMs. I think some of the
scars
I have from working with MFC (for example) could have been avoided if
someone
had done the homework up front (wishful ;-). While your dealing with
small-scale,
desktop applications I think you can duck the issue. Once you start tangling
with larger
scale stuff that must meet reliability requirements your better off doing it
right. Its
funny though - once you've done a few FSMs for the larger systems it seems
silly
to do it any other way even for smaller ones!

Anyway, I will go over the recent mails and get back to you again soon.

Ta,
SW