I concur with the general point that Forth stacks don't need to be big.
We've done a few studies on large applications, and the stacks rarely
get deeper than 4-5 items. It's ironic that Forth's stacks are explicit
and obvious, but remain consistently way smaller than C's implicit
stacks. The stacks on the SEAForth chips seem small, but then the whole
chip is small, and one will have to be sensitive to this when
I do like being able to detect stack underflows, though. That's very
helpful in catching programming errors. Ideally, of course, one
shouldn't make errors, but none of us is perfect.
Elizabeth D. Rather (US & Canada) 800-55-FORTH
FORTH Inc. +1 310-491-3356
5155 W. Rosecrans Ave. #1018 Fax: +1 310-978-9454
Hawthorne, CA 90250
"Forth-based products and Services for real-time
applications since 1973."