On 12/03/2012 08:04 PM, Robert Dodier wrote:
> Without bringing signed zero into the picture, can you show that
> sqrt((-i)^2)/(-i) is something other than -1 ?
I agree with Robert.
Ray's proof that it gives +1 has a mistake because he initially writes
z=-i as 1*exp(-%i*%pi/2) but then, when he goes on to divide sqrt(z^2)
by z, he writes
z as -1*exp(-%i*%pi/2) which is i and not -i.
Jaime