Subject:
|
Re: events or infinite loops... which is better?
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Mon, 14 Oct 2002 11:32:26 GMT
|
Original-From:
|
PeterBalch <PeterBalch@[spamcake]compuserve.com>
|
Viewed:
|
902 times
|
| |
|
|
Rob
Something else you should consider is reliability.
There used to be a saying: "There are no interrupts on the Space Shuttle".
Maybe it's still true.
(By "Events", I assume you mean interrupts. Really, using Events is just a
design philoshopy and can be implemented using either interrupts or
infinite-loop polling.)
It is very, VERY difficult to write a reliable program that uses
interrupts. I write a lot of code for PIC chips (broadly similar to the
chip used in the RCX) any always avoid interrupts if I possibly can.
Typically, the chips have to keep an eye on half a dozen tasks at once - a
couple of serial I/O ports, a coin mechanism, a swipe-card reader a cancel
button, a relay or two, a text display, etc. etc. And I always do it all
with single master loop that polls each peipheral. It's a real pain keep
track of what each task is doing and never allow any task to hog more than
a millisecond of processing time but the programs must be 100% reliable and
have a MTBF of months or years.
Very occasionally I need to respond in microseconds and am forced to use an
interrupt. But usually a millisecond response time is perfectly acceptable.
Programing with interrupts is guaranteed to expose you to the full force of
Finagle's Law: "Anything that can go wrong, will".
Peter
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
