The new motion

[Ulrich Kulisch <ulrich.kulisch@xxxxxxx>]

Dear colleagues,

this motion makes me really sad! We are going to make a big mistake with this motion. For interval arithmetic we have the unique chance defining a calculus that is free of exceptions. Instead the motion tries to benefit from binding it more and more to IEEE 754. Let me comment a litlle on it:

In 2009 a motion requiring an EDP within P1788 passed by the given rules with shall. Shortly after that (April 2010) the EDP was listed in Draft 2.2 of the standard on page 3, under
0.3. Inclusions. The standard specifies
-  Data type and operations for the calculation of exact sums and dot products.
This was  level 1, where it indeed should be installed to give it appropriate weight.

Later it was argued it does not really fit to this location and will be listed later in the text. After a very long time it reappeared under 11.11.11. This was level 2. The text, however, did not quite reflect the intention of the motion. So Van Snyder and myself requested a few changes. But nothing happened. After some time we repeated our request for changes twice. But nothing happened until today.

Then it was argued there is some confusion around Motion 9 and clarification was requested. I answered with several detailed mails with attachments to the diverse requests. Please read these mails and the attachments (mail of May 18 with 4 attachmnents, 2 mails of May 20 with one attachment, mail of May 22, mail of May 31 with one attachment, mail of June 6 with one attachment).  I got no serious response. I felt that I was talking to a wall.

Then motion 45 was launched. It lowers the requirement for the EDP from shall to recommended and lifts its status from level 2  to level 3. This is momentarily the result of a long tactical journey. It has nothing to do with science. The next step in this chain could be lifting it to level 6? Under Rationale motion 45 shows the following text:

(1) I do not think Prof Kulisch has made the case that there are many problems that can only (or even best) be solved by EDP, rather than a dot product correctly rounded, or faithfully rounded, to the working precision (CRDP and FRDP respectively). In current discussion, he cites the same two applications he did in late 2009: Rump operator & logistic map. And Lefevre, Neumaier & Rump all say these can be solved nearly as well using FRDP.

This makes a bad case and gives a distorted view of my contributions. I can only repeat myself: read the mails and the attachments listed above! The title of the position paper for motion 9 mentions already another important application of the EDP.  How is the "say" to be interpreted? I am not aware of any mail or paper of the colleagues which shows this. If there are any, please let me know.  We had the EDP in the XSC-languages since 1980 and the toolbox volumes impressively prove its usefulness. For other fascinating applications see [5] on the poster.

My mail of May 31 contains the following text:
Interval arithmetic brings guarantees and safe bounds into computing. These bounds frequently are overly pessimistic. The EDP is the simplest and most general tool for getting close bounds. It brings accuracy and speed and it is free of exceptions. A combination of both is what is needed from a modern computer. Interval arithmetic and the EDP form one unit. 
We did not get the EDP from IEEE 754. So it is most natural that it is provided in IEEE 1788.

This text should be used as replacement for  (1) above in motion 45.

In numerical analysis the dot product is ubiquitous. It appears in the arithmetic for interval vectors and interval matrices over the real and complex numbers, for instance. Many of the advanced applications discussed in Chapter 9 of my book are based on applications of the EDP. The simplest and fastest way for computing a dot product is to compute it exactly. Designing a solution is easy not to say simple with the tools that are availble today. The IFIP letters should not be wiped out so easily.

I am not against looking for a compromise. But I think a minimum of fairness should be maintained. It was much easier in the 1980ies to convince a number of companies (IBM, Siemens, Hitachi, NAS) to provide the EDP on their computers than a group of anxious and worried mathematicians. I am absolutely convinced that we shall get the EDP, if we require it. We weaken our position considerably if we just recommend it half-heartedly. 

Let me finally make a personal remark. I have supplied plenty of material supporting my view of the matter. I know, the truth sometimes takes time to find its way. But I really don't want to find myself in a fight with colleagues who I so far considered being my friends. So I am serously wondering whether I should leave P1788. If there are further questions you still might  consult my book: Computer Arithmetic and Validity, De Gruyter, second edition 2013. If you need a copy please let me know. I would be happy to send you one. 

With best regards
Ulrich Kulisch

-- 
Karlsruher Institut für Technologie (KIT)
Institut für Angewandte und Numerische Mathematik
D-76128 Karlsruhe, Germany
Prof. Ulrich Kulisch

Telefon: +49 721 608-42680
Fax: +49 721 608-46679
E-Mail: ulrich.kulisch@xxxxxxx
www.kit.edu
www.math.kit.edu/ianm2/~kulisch/

KIT - Universität des Landes Baden-Württemberg 
und nationales Großforschungszentrum in der 
Helmholtz-Gesellschaft