>>5
It's actually pretty stupid because of the errors induced by inexact floats. Floats only have place in approximative calculations with what should be "real"-like numbers, but where errors can be tolerated. Normal fixnum/bignum and ratios/complex numbers should be used for everything else.