I really like in-shell floating point operations.
But this:
> / 1 0
+Inf
Is not the correct answer - even as a limit - since the limit of 1 / x (when x -> 0) can be +Inf or -Inf.