Re: Draft: P1788.1 Standard for Interval Arithmetic (Simplified)
On 2015-07-30 11:22:24 +0000, Corliss, George wrote:
> I fully agree with Vincent on the value of hex.
>
> However, I cannot miss the opportunity to remind us that (EXCEPT FOR MATHEMATICS) very few numbers are known to more than a few places. Hence (EXCEPT FOR MATHEMATICS), no intervals are thin.
>
> Of course, mathematical analysis is important, too, and there, thin and 1 ULP intervals are appropriate.
Some problems, not necessarily purely mathematical, can be solved by
mixing different methods (and different software). For instance, one
may have a problem that needs two steps:
1. Compute some value x.
2. Compute f(x).
It may happen that one can do both (1) and (2) in interval arithmetic.
But one may also want to regard x as exact for (2), even though there
may be some error attached to it (this can be dealt with separately).
--
Vincent Lefèvre <vincent@xxxxxxxxxx> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)