maxima-bounces at math.utexas.edu wrote on 03/08/2007 01:53:37 PM:
> Until and unless we do that, I think we are forced to simplify 1^inf to
UND.
If we agree that 0 * inf = und and exp(und) = und, I think we
should have
1^inf = exp(log(1^inf)) = exp(inf * 0) = exp(und) = und.
Extended real arithmetic should *useful* to the limit function. I think
1^inf => und is more useful to limit than is 1^inf => 1:
limit((1+1/x)^x,x, inf) => 1^und => und /* Dude! Gotta try something
else. */
limit(exp(x),x,inf) => exp(inf) => inf /* OK ! */
Barton