On Friday, 97/01/31, Jonathan de Boyne Pollard wrote to Murray Lesser
about "Floating point feature w" as follows:

JP> ML>
JP>   >                                         For best results, one should
JP>   > avoid floating-point calculations entirely, particularly if the result
JP>   > is a matter of life and death (or  money); something that is very
JP>   > difficult to do at times.
JP> ML>
JP> 
JP>   I'm sitting here wondering why anyone would want to take the sine of a
JP>   monetary amount ... (-:

Hi Jonathan,

"Creative accounting" perhaps. ... :-)

Actually, trigonometric functions can occur in financial applications,
but seldom applied to the monetary amount itself. They usually pop up
in Fourier analysis of time series.

Of course, for real amusement start contemplating monetary amounts
around 1E+18 in magnitude. You won't care about inaccurate sines then.

Regards

Dave


 * KWQ/2 1.2i * Your reality check just bounced.
--- Maximus/2 3.01