RE: raw March 17 Minutes
[Amended for attendee list]
IEEE 754r Meeting minutes March 17, 2004
Sun, Santa Clara
Attendees (partial list): Affiliation
Dan Zuras Self
David Hough Sun
Eric Schwartz IBM (phone)
Alex Liu Sun
Joe Darcy Sun (host)
Jim Thomas HP
Mike Cowlishaw IBM (phone)
Ivan Godard OOTBC
Jeff Kidder Intel (minute taker)
Fred Tydeman self
Peter Markstein HP
Leonard Tsai self (phone)
Dick Delp self
Note taker: Jeff Kidder
Called to order 1:15
Agenda for two days:
Draft review [today]
Would like to extend schedule through August.
Would like to confirm we have come to a
consensus on NaNs.
What are we going to do about extended.
Fixed mode declarations (ahead of Alt exception)
* Draft review
3.0 Page 18 "The level table" Table 0
We should represent SNaN vs QNaN first at level 3
We should represent NaN fraction content first at level 4
Added QNaN U SNaN at level 3.
Page 19 at the top
3.1 Page 20 Second bullet
The fractional description.
The integer description.
Change "of the form" to "of the form of a digit string"
Also in previsions section.
The suggested was made to represent multiplication
with an explicit product symbol.
Tabled as a global discussion.
Incorporate Jim's input on the reduced precision of
3.2 Page 23
Another digit string....
Note entity is not in the definition list.
Show both level 3 and level 2.
3.3 Page 25
Section removed because moved to decimal
we changed from red to purple since it
should be made available somewhere.
Bullet 4: Remove x: 3xJ -> 3J
4.1.1 Page 38
Add "nearest" back in to make the "bracketing" clear.
Same with top of page 39.
Add some purple prose to remind us of the subtle
problem with the phrase "the two nearest numbers".
Moved to appendix N.
5.3 Page 43
Add "in the same radix" to "... or "
5.4 Page 44
Changed to just "... in section 7".
5.5 Page 44
Change "as specified in Section 4" to
"analogously to Section 4".
5.11 Page 57
Need text for quantize(ZERO, INF)
5.12 Page 58
Add "Except for quantize..." to "Numerical values of results..."
Removed "The exception to this rule ...are not in the same cohort".
Replaced text Jim recommended to the last paragraph of page 58.
Someone needs to go through and look at each operation
and figure out which operations.
Split into two tables. The first which apply only for exact results
(exact add sub mul div fma sqrt d2d-converts(?) and a third column for
the inexact case. The second table applies to all results.
Ivan raised a desire for a Prior(x,y) predicate which is a total
We need a proposal.
Strict "The sign of a NaN is not determined by this standard and"
"The sign of the result of the round" added " or convert..."
We will ask Professor Kahan for some proposed wording about Invalid.
Item 1 Need to do something about predicates with SNaN
Page 65 Item 8
Change "signaled" to "indicated"
Appendix 1 Page 69
Changed last two sentences of paragraph to "They are described
generically as in section 5"
Item 15 Page 71
Accepted pending additions.