On Wed, 11 Apr 2007, Jay Belanger wrote:
> If you do calculations in which every step can be carried out
> in the current precision exactly, and yet the answer is off, then I
> think something worth noting is going on.
Exactly! At last I completely agree with you.
"The current precision" means in most cases IEEE double precision. Some
calculations, like 0.5 + 0.25 for example, can be carried out in the
current precision exactly. And they are. Some other calculations, like
1.4^2, cannot be carried out exactly. And they are not.
Andrey