Having these methods in Numeric can get in the way of creating subclasses of
Numeric because they only support Fixnum and Float.