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

Re: Motion 45



Dear Prof Kulisch,
I think your concern is moot, contrarily setting an explicit +0 enforces that there is only ONE zero in p1788
yours Juergen WvG

Am 01.08.2013 05:42, schrieb Ulrich Kulisch:
It was clear to me from the beginning that I shall vote No on Motion 45. But the motion was accepted. So reading the text of the motion more carefully now a number of inconsistencies occur to me. I wonder whether somebody of the Yes voters can help me staighten these out.

The product of two real numbers is a particular form of a dot product. Now consider the interval product with a>0 and b<0:
[0, a] × [b, 0] = [a×b, +0]   or    [0, 0] × [c, d] = [+0, +0]
By motion 45 the zeros in the results have to carry a plus sign. This looks very strange to me in particular since a×b<0. A signed zero does not make sense in interval arithmetic. An interval with zero as one bound has already a direction.

Similar for division by an interval where a bound is -oo or +oo:
If in the result of an interval division one bound is zero, like [a, 0] or [0, b], the other bound shows immediately whether the elements are positive or negative. So there is absolutely no need for indicating this by a minus or plus symbol in front of the zero.

A real interval is defined as a closed and connected set of real numbers. In case of unbounded intervals -oo and +oo are just used as bounds. They are themselves not real numbers! So in case of interval division the rules for computing the bounds of the result do not strictly follow the pattern of IEEE 754 arithmetic for extended real numbers. For instance 0×(-oo) = (-oo)×0 = 0×(+oo) = (+oo)×0 = 0 in interval arithmetic and not NaN like in IEEE 754 arithmetic.

Well defined interval arithmetic over the real numbers (or the subset of floating-point numbers) including CA and the EDP leads to a calculus that is free of exceptions. We must not make the mistake defining interval arithmetic over all IEEE 754 datums!

The development of IEEE P1788 suffers from the fact that many members of the group take IEEE 754 arithmetic as basic truth for scientific computing. This may be true for floating-point arithmetic. But it is not true for interval arithmetic. The two should be separated more clearly. Within IEEE P1788 we have the unique chance of defining interval arithmetic including CA and the EDP as a powerful but simple calculus that is free of exceptions.

What in Motion 45 under 1. is called the "broader aim" actually is a much narrower aim.

I shall comment on a number of misunderstandings concerning CA and the EDP in another mail.

With best regards
Ulrich


--
-                Prof. Dr. Juergen Wolff von Gudenberg
     o           Lehrstuhl fuer Informatik II
    / \          Universitaet Wuerzburg, Am Hubland, D-97074 Wuerzburg
InfoII o         Tel.: +49 931 / 31 86602 Fax ../31 86603
  / \  Uni       E-Mail:wolff@xxxxxxxxxxxxxxxxxxxxxxxxxxx
 o   o Wuerzburg