>>>>> "willisb" == willisb <willisb@unk.edu> writes:
willisb> 1. Look for %e ^ x and evaluate it as (exp x) instead of
willisb> (expt 2.718... x). Similarly, we could look for x ^ 0.5 and
Ok, it's exptb in simp.lisp. I think there's a bug there in that it
does (exp x) but forgets to (float x) so we get a single-precision
result on clisp/cmucl instead of a double-precision.
Also, some tracing of functions seems to show that for the %e^%e
problem, simplifya is returning it unevaluated. But %e^2 does
something different wherein simplifya actually evaluates this (but to
single-precision!).
Don't know how this all works.
Ray