Arbitrary precision decimal floating point arithmetic
