[cfp-interest 3878] Re: Comments re Annex F Introduction
Damian McGuckin
damianm at esi.com.au
Wed Apr 15 22:04:54 PDT 2026
Hi all,
I would like to raise one of the discussion items for which we ran out of
time earlier today. These are editorial issues and seek to improve clarity
by eliminating repitition.
In C2Y, the two wildly separate paragraphs F.3#20 and F.10.1#16 exist, the
later under a heading of Recommended Practice.
The C functions in Table F.3 correspond to mathematical operations
recommended by ISO/IEC 60559. However, correct rounding, which
ISO/IEC 60559 specifies for its operations, is not required for
the C functions in the table. 7.35.9 (potentially) reserves cr_
prefixed names for functions fully matching the ISO/IEC 60559
mathematical operations. In the table, the C functions are
represented by the function name without a type suffix
ISO/IEC 60559 specifies correct rounding for the operations in
Table F.3 recommended by ISO/IEC 60559, and thereby preserves
useful mathematical properties such as symmetry, mono- tonicity,
and periodicity. The corresponding functions with (potentially)
reserved cr_-prefixed names (7.35.9) do the same. The C functions
in the table, however, are not required to be correctly rounded,
but implementations should still preserve as many of these useful
mathematical properties as possible.
They talk about the same thing, i.e. the functions in Table F.3. They also
both reference 7.35.9, the reservation of cr_-prefixed names for correctly
rounded recommended operations of mathematical functions.
I propose to being them together. I suggest four (4) paragraphs and one
(1) footnote, the first of paragraph appearing before the table when Table
F.3 is first mentioned, the last three paragraphs appearing after the
table appears. Those last 2 appear under a heading of Recommended Practice:
The C functions in Table F.3 correspond to mathematical operations
recommended by ISO/IEC 60559.
The C functions in Table F.3 are not required to be correctly
rounded. [this should be very blunt and obvious]
Recommended Practice
Since ISO/IEC 60559 preserves useful mathematical properties such
as symmetry, monotonicity, and periodicity for its operations in
Table F.3 by specifying their correct rounding, C implementations
of those operations should still preserve as many of these
properties as possible.
For any function f in Table F.3, an implementation may provide a
corresponding function cr_f with the same underlying functionality
as f but with a correctly rounded result guaranteed.
There is a footnote against cr_f which says ...:
Those cr_-prefixed names are (potentially) reserved (7.35.9).
Should this cross reference appear within the paragraph like it does
currently?
Please comment before the next meeting to save time during the meeting.
Thanks - Damian
More information about the cfp-interest
mailing list