Midpoint-radius
(was "Re: floats and intervals, constants, 0.1, 1e400 and Inf")
P1788 members
Svetoslav Markov wrote on 2008/11/11 Tue AM 08:04:50 GMT
> On 10 Nov 2008 at 13:02, Arnold Neumaier wrote:
> > Though briefly discussed, I think nobody has shown strong emotions for
> > satandardizing midpoint-radius interval representations.
> > I advocate against it.
>
> I should like to declare myself as a strong supporter of
> midpoint-radius interval representations.
Partly with my Technical Editor hat on, I give my view that there is a difference between
(a) "I find midpoint-radius (MR) a useful representation in certain applications."
and
(b) "I want to see MR standardised by P1788."
I see some mathematical complications in MR, e.g.
- There is not always a unique finite-precision interval hull of a set, using MR.
- Conversion between MR and the bounds representation usually is inexact in both directions.
However, I know Siegfried Rump uses it extensively in Intlab, and not just for representing complex intervals. Should it be considered as
just a useful internal form used to make certain algorithms more efficient?
I suggest supporters of MR get together and produce a position paper, NOT to draft a MR standard, but to debate the motion
"MR should be standardised by P1788",
present the arguments for and against, and come to a conclusion.
Best wishes
John Pryce
-----------------------------------------
Email sent from www.virginmedia.com/email
Virus-checked using McAfee(R) Software and scanned for spam