Jonathan de Boyne Pollard wrote in a message to Murray Lesser:

ML>
  >                                         For best results, one should
  > avoid floating-point calculations entirely, particularly if the result
  > is a matter of life and death (or  money); something that is very
  > difficult to do at times.
ML>

 JdBP>   I'm sitting here wondering why anyone would want to take
 JdBP> the sine of a   monetary amount ... (-:

In all seriousness, it does occur often in financial modeling.  One popular
technique involves using what would look like FIR and IIR filters to an
engineer, but applied to monetary events.  Certain kinds of financial and
actuarial calculations also involve trigonometric and exponential
mathematical operations, and the result could very well be something as
tangible as your annual insurance premium.
 
-- Mike


---