(j3.2006) Interpretation F08/0055

[Robert Corbett]

I shall vote no on F08/0055.  I agree with the
interpretation in general, but I disagree with
some details.  The answer given to question Q6
("No.") is at best misleading.  Also, the edits
need some work.

Question Q6 includes an example that shows a
sequence of increasing values being written
and a sequence of values in mixed order being
written.  The question is is that result the
intended result, and the answer is no.  That
is technically correct, but misleading.  The
example given depends on the scale factor
being nonzero.  In general, when the scale
factor is nonzero, the number of significant
digits in values written in the form of an
F edit descriptor are different from the number
of significant digits in values written in
the form of an E edit descriptor.  There was
a recent e-mail exchange that dealt with this
issue.  If the number of significant digits
produced is different, the value can change
in either direction.  Consider the value 1.48.
It turns into "1.", "1.5", or "1.48" if
rounded to one, two, or three digits.

The edits speak of the "decimal exponent value,"
without giving a definition of the term.
Searching 10-007r1 turns up uses of the term
"decimal exponent," but they are context
dependent and two of them are clearly not what
is meant in the edits.  A definition of the
term that works is that the decimal exponent
value of a number N is FLOOR(LOG10(ABS(N)))+1.
While this looks bad in ascii text, it looks
better when expressed using Iverson brackets
for FLOOR and vertical bars for ABS.

Robert Corbett