The intention of this trait is to indicate whether a given class behaves like a floating-point
type, and thus allows division of one value by another with acceptable loss of precision.
will be treated as
if it behaved like an integral type for the purpose of these conversions.
— end note