Originally posted by Jim Yingst:
NetRexx: dunno, but I'm going to guess they just use BigDecimal.
IEEE 754: wouldn't that lead to roundoff errors? I think you're going to have a hard time getting answers exactly like Jignesh originally specified. The alternative of using a binary integer combined with a base-10 exponent (i.e. what M^2 did, and what I had planned to do, and what BigDecimal does internally) seems like the most memory-efficient choice that also gives exact answers for all decimal numbers of finite length. Methinks there must be an IEEE standard somewhere which describes this sort of format, as it seems a natural choice for exact decimal calculations.