Hacker News new | past | comments | ask | show | jobs | submit login

The bigint you describe is just a poor man's rational, given that no computer architecture or mainstream language support them natively.

There is a slight difference where it forces decision-making about rounding at every step. Most importantly, it makes errors that print money (or burn money) impossible states. I'm aware of BigRational libraries or more clever currency abstractions, but this is the least magical way of doing it, kind of like representing datetimes as 64-bit UNIX epochs.

Well, you could look at it as x/100 rationals representing a dollar value. But you could also look at it as an integer amount of a smaller unit (cents). The difference is insignificant; computers support it natively.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact