Re: P1788/M0024.03:RoundedOperations: NO
Ian,
On 7/7/2011 7:43 PM, Dan Zuras Intervals wrote:
To: stds-1788@xxxxxxxxxxxxxxxxx
.
.
.
I might vote yes, but I would wonder whether a requirement involving on=
ly
floating point operations on floating point values, not interval operat=
ions
on interval values, should be part of 1788 instead of a future 754
revision. It would be useful to floating point users and to interval
implementers, but shouldn't interval users avoid such usage?
That's an interesting point of view, and I've been wondering
about that myself. A number of researchers have done automatic
proofs using computer arithmetic, based on manually controlling
directed rounding, rather than using interval arithmetic throughout.
Doing so, they can sometimes achieve sharper bounds than with
naive use of interval arithmetic, although it can take much
more analysis and programming.
I view such endeavors as closely related to interval analysis,
and to be within the scope of our journal "Reliable Computing."
However, in which standard associated arithmetic operations
should appear (754, 1788, or both?) might be a different question.
Baker
--
---------------------------------------------------------------
Ralph Baker Kearfott, rbk@xxxxxxxxxxxxx (337) 482-5346 (fax)
(337) 482-5270 (work) (337) 993-1827 (home)
URL: http://interval.louisiana.edu/kearfott.html
Department of Mathematics, University of Louisiana at Lafayette
(Room 217 Maxim D. Doucet Hall, 1403 Johnston Street)
Box 4-1010, Lafayette, LA 70504-1010, USA
---------------------------------------------------------------