On 8/28/06, Robert Dodier <robert.dodier at gmail.com> wrote:
> ratsimp *might* be expensive but I am 100% certain
> that people will be surprised that is(equal(sqrt(n)/n, 1/sqrt(n)))
> => false for some positive integers n.
People should be equally surprised that the general simplifier doesn't
simplify sqrt(2)/2-1/sqrt(2) to 0.
This is simply a bug (reported a long time ago), which we should fix
regardless of whether 'is' calls ratsimp.
-s