Martin DeMello wrote:

> Why classes? They're simply constants of type Float (defined in the IEEE
> standard), and ruby already does provide access to them, complete with
> pretty printing. It's just that it's a bit roundabout - e.g. you have to
> say a = 1.0/0 rather than a = Infinity.

Having them implemented in a custom way would let you use them portably. 
(Negative) infinity is quite useful when working with Ranges even if it 
does not consider itself equal to 1.0 / 0.0.