Re: Strange result with exponents...

>>> It does seem to be a disclaimer that we
>>> shouldn't expect float to be terribly precise,
>>> which is a different attribute from repeatable.
>
> And thus from a backwards compatibility standpoint,
> it's pretty clear that changing exponentiation to use
> a floating point value (if that's indeed what happened)
> actually reduced precision.

I am SO sorry for the miscue!  Exponentiation always used float.  The V5R1
manual has the exact same wording as the V5R2 manual.  My V4R4 manual has
the same wording and there are no change bars there, so I would bet that
this behaviour has been documented for a long time.  I understand loss of
precision, but the lack of repeatability between releases has me a bit
puzzled.  I suspect that something in the runtime has changed, and since
FLOAT8 doesn't guarantee precision beyond about 15 digits, whatever change
occurred passed validation testing.

  --buck