banks

[Roger Schlafly]

>> The main conclusion is that banks should *not* use floating-point (contrary
>> to popular belief) because it is in a sense unpredictable.

The most popular financial software (by far) is Lotus 1-2-3.  It uses
IEEE 8-byte binary floating point.  They could add BCD arithmetic if
there were market demand for it, but there isn't.  Floating point
works fine.  Hence I see no empirical evidence that bankers are actually
dissatisfied with binary floating point.

>> Banks are not interested in error estimates, they want exact figures.

An error estimate is fine if it is used to monitor rediculously
implausible events, such as someone depositing the entire GNP of Japan
and withdrawing it all the next day.  With floating point arithmetic,
such an operation might be off by a fraction of a yen. <g>

--------------------------------------------------------------------------
| Roger Schlafly * voice phone or fax: 408-476-3550 * BBS: 408-476-1055  |
| US Mail: Real Software, PO Box 1680, Soquel, CA 95073 USA              |
| MCI: RSchlafly * CIS: 76646,323 * GEnie: R.Schlafly * Prodigy: TTDX08A |
| Internet: 76646.323aCompuServe.com or 0002956602aMciMail.com           |
--------------------------------------------------------------------------