acos(1)

[Michel Hack <hack@xxxxxxxxxxxxxx>]

Fred Tydeman observed that the standard used "preferred quantum" in
section 9, but "preferred exponent" elsewhere.  That's probably an
editing slip due to differing terminology at IBM, where we decided
to use "preferred quantum".  The quantum is the value of the least
significant digit, or 10**exponent in what IBM calls "right units
view", and IEEE calls "significand viewed as an integer" (I think).

We decided to use "quantum" because that's independent of the view,
whereas "exponent" is not.  The standard had to be very careful
about this, which is why the letters "q" and "e" are used to denote
the exponent ("q" is NOT a quantum here) in the right units view
(typically used with DFP) and the left units view (typically used
with BFP).  I don't recall this issue actually being discussed (I
do remember the discussions that led to the use of "e" and "q"),
but I suspect it was simply too late to change the standard from
"preferred exponent" to "preferred quantum", given the time pressure
near the end of the working group's life.

Michel.
Sent: 2008-08-29 06:27:56 UTC