"Richard Fateman" wrote on 10/14/2008 09:01:39 AM:
> isn't exp(-inf) = 0)?
Yes, I think it is. But (ideally) simpexpt would simplify exp(-inf) to 0;
it would be
redundant for mnqp to trap this case. I suppose my proposed meqp-by-csign
should look
for things like inf ^ x. Currently, we get an asksign:
(%i1) is(equal(inf^x,0));
Is x positive, negative, or zero? pos;
(%o1) false
Barton