Re: discussion period begins for Motion P1788.1/M002.01: Required operations
All,
I am for including cancelMinus and cancelPlus, for practical
reasons: (1) I have found them useful when "updating"
sums by removing previously added terms, such as in the
interval Gauss--Seidel method; (2) It doesn't seem to add
complexity to the "simplified" standard.
Of course, there may be counter arguments associated with them
not being a part of a closed system, etc. I'm "all ears."
Baker
On 10/12/2015 09:30 AM, Nathalie Revol wrote:
Dear Colleagues,
as Dmitry has moved this motion and Evgenija has seconded it,
(thank you to both of them), this is Motion P1788.1/M002.01:
the discussion period now begins and will end November 2nd.
Best regards
Nathalie
On 12 Oct 2015, at 16:17, Evgenija Popova <epopova@xxxxxxxxxx> wrote:
I second the motion "Required Operations".
Evgenija Popova
Institute of Mathematics and Informatics
Bulgarian Academy of Sciences
P-1788:
Does someone second this motion?
Baker (acting as chair)
On 10/12/2015 07:02 AM, Dmitry Nadezhin wrote:
Motion "Required Operations".
Operations required in all flavors of the standard IEEE 1788-2015
are required also in the simplified standard P1788.1 .
Specifically, 4 operations missed in the initial draft are required
in P1788.1:
- arithmetic operations
roundTiesToEven(x)
roundTiesToAway(x)
- cancellative addition and subtraction
cancelPlus(x,y)
cancelMinus(x,y) .
-----------------
Rationale.
This (together with better conformance of interval literals) permits
more formal relation between IEEE 1788-2015 and P1788.1 .
It makes P1788.1 become a flavor of IEEE 1788-2015
though text of P1788.1 doesn't need to mention flavors.
If P1788.1 satisfies requirements of IEEE 1788-2015 flavor without
any exemptions
it starts a good tradition of adding new flavors to IEEE 1788-2015.
The specification of roundTiesToXXX(x) and specification
cancellative operations
doesn't add too much text to the P1788.1 .
There are simple implementations of these functions.
A possible implementation of roundTiesToXXX(x) operations can be
found here:
https://java.net/projects/jinterval/sources/svn/content/trunk/jinterval/p1788-launcher-java/src/main/resources/net/java/jinterval/p1788/integerFuns.c?rev=379
.
And Michel Hack wrote how to do it simpler:
http://grouper.ieee.org/groups/1788/email/msg08423.html .
There are subtleties in implementation of cancellative operations
when the answer is nearly a one-point interval (or nearly undefined).
See notes in subsection 12.12.5 of IEEE 1788-2015.
Nevertheless, Marco Nehmeier suggested a solution:
http://grouper.ieee.org/groups/1788/email/msg07858.html .
--------------
--
---------------------------------------------------------------
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
---------------------------------------------------------------
------------------------------------------------------------------------------------------------------------
Nathalie Revol INRIA Grenoble - Rhone Alpes
LIP - projet AriC tel: (33) 4 72 72 84 36
Ecole Normale Superieure de Lyon fax: (33) 4 72 72 80 80
69364 Lyon Cedex 07, France Nathalie.Revol@xxxxxxxxxxx
http://perso.ens-lyon.fr/nathalie.revol/
------------------------------------------------------------------------------------------------------------
--
---------------------------------------------------------------
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
---------------------------------------------------------------