makes the float constructor behave in the same way as specified
by various other language standards, including C99, IEEE 754r,
and the IBM Decimal standard.