Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: Siegfried's recent paper, and other matters



Ok, I like this idea.
We will focus on writing a motion to address the issues (1)-(4) you list below. I will also add another item: 5. What are the digital roundings of the Kaucher intervals for arithmetic operations and library functions I will still supply a reference implementation to accompany the motion, so it can be studied in the context of real-world applications to computer science.
Nate

----- Original Message ----- From: "Alexandre Goldsztejn" <alexandre.goldsztejn@xxxxxxxxx>
To: "Nate Hayes" <nh@xxxxxxxxxxxxxxxxx>
Cc: "John Pryce" <j.d.pryce@xxxxxxxxxxxx>; "stds-1788" <stds-1788@xxxxxxxxxxxxxxxxx>
Sent: Tuesday, January 03, 2012 12:45 PM
Subject: Re: Siegfried's recent paper, and other matters


Nate,

Writing an appendix to the standard seems to me too early: In order to
vote for an appendix, we certainly need to see how it will be
integrated within the standard, which is too early since the standard
is not yet completely settled. This is the reason why we vote on
motions instead, which are independent and self-contained and will be
used to guide writing the standard.

I therefore recommend instead writing a motion, detailing definitions
related to Kaucher intervals and their related operations proposed be
included in this appendix. This will help focussing on the content
instead of the form...

I think furthermore several issues should be addressed before any
detailed appendix or implementation is discussed. For example:

1- Will the operations over Kaucher interval be restricted to bounded
Kaucher intervals?
2- What correspondance will be provided between the classical empty
set and Kaucher intervals?
3- How to compute operation with Kaucher interval that are not
included in the domain definition of the functions?
4- How to compute reverse operation on Kaucher intervals?

Regards,

Alexandre

On Tue, Jan 3, 2012 at 2:52 PM, Nate Hayes <nh@xxxxxxxxxxxxxxxxx> wrote:
It seems to me a more scientific approach would be to vote on a document
like the one described below, as well as its accompanying C++ reference
implementation. Otherwise IMHO its a bit of a pig-in-the-poke.
Nate

----- Original Message ----- From: "John Pryce" <j.d.pryce@xxxxxxxxxxxx>
To: "Alexandre Goldsztejn" <alexandre.goldsztejn@xxxxxxxxx>
Cc: "stds-1788" <stds-1788@xxxxxxxxxxxxxxxxx>
Sent: Tuesday, January 03, 2012 1:06 AM
Subject: Re: Siegfried's recent paper, and other matters



Alexandre

This seems a good idea to me. Will you propose this motion? I was only
expressing my personal opinion in my points 1 to 4.

John Pryce

On 3 Jan 2012, at 06:11, Alexandre Goldsztejn wrote:

Dear all,

Ok. I will write the text for this appendix. Svetoslav has agreed to
help.
If anyone else wishes to contribute, please contact one of us offline.

I will also make a commitment to providing a C++ modal interval
reference
implementation to accompany the appendix.


Before Nate and others spend so much time writing these documents and
implementations, shouldn't someone propose a motion to include Kaucher
intervals in an appendix of the standard?

Personally, I don't agree with points 3 (the standard should be as
much as possible compatible with Kaucher intervals) and 4 (an appendix
of the standard should be included in the standard) of John about
Kaucher intervals, a motion could be the opportunity to discuss them.






--
Dr. Alexandre Goldsztejn

CNRS - Laboratoire d'Informatique de Nantes Atlantique
Office : +33 2 51 12 58 37 Mobile : +33 6 78 04 94 87
Web: www.goldsztejn.com
Email: alexandre.goldsztejn@xxxxxxxxxxxxxx