The purpose of the EXTNUM class implementation is to provide the user with
a floating point number type that extends the dynamic range
(i.e., the allowable range of the exponent) for a double type floating point
number, while preserving the approximate 15 digit precision of the double as
well. The class presumes that the internal format of the double type on the
user's machine is compliant with the IEEE 754 standard. The class is intended to
mimic the numerical behavior of the built-in double floating point type
while EXTNUM-type values lie within the range of the double.