On Tue, 07 Mar 2006 18:41:32 -0800
Jeff Newmiller <[EMAIL PROTECTED]> wrote:

> "finite state machine"

Indeed, I love state machines. I've built entire projects around them.

Had to do some bit-twiddling radio communications (including generating
-- and detecting, of course -- Manchester coding) on a PIC micro once.
Boy, that was fun, let me tell you! Naturally I used a state machine,
as I was generating (and receiving) the bits on-the-fly. And, well, a
bunch of stuff. It was complex, and the most complex part was in the
battery-powered handheld remote, which had the least resources
available. (For speed I stored the next state directly as a pointer
rather than as an index for a case statement.) And naturally, after I
left they designed in a more sophisticated and faster radio system,
which means they took all that carefully designed and tuned code and
tossed it right out the window. *sigh*


State machines are particularly useful where you've got an
interrupt-driven system that needs to remember what it was doing last,
and what it needs to do next, based on current conditions and/or input,
because you're not sitting in a polling loop waiting for the next
event. Or anything where the events happen asynchronously. 

Like, debouncing and tracking the mute button.

Oh right, they changed that too. O_o
_______________________________________________
vox-tech mailing list
[email protected]
http://lists.lugod.org/mailman/listinfo/vox-tech

Reply via email to