The big win here is that we can have special int vs double constructors for Fn, which means we can use Real(1) semantics instead of Real(1.0) for data.
Doing this exposed a weakness in FractionDecimal where you can overflow Long too easily; we switched to Double for the numerator and denominator so that you just lose precision vs overflow when this happens.
The big win here is that we can have special
intvsdoubleconstructors forFn, which means we can useReal(1)semantics instead ofReal(1.0)for data.Doing this exposed a weakness in
FractionDecimalwhere you can overflowLongtoo easily; we switched toDoublefor the numerator and denominator so that you just lose precision vs overflow when this happens.