Re: Motion 4: P1788 on non-754?
> Date: Tue, 14 Apr 2009 19:13:39 +0900
> From: Christian Keil <c.keil@xxxxxxxxxxxxx>
> To: stds-1788 <stds-1788@xxxxxxxxxxxxxxxxx>
> Subject: Re: Motion 4: P1788 on non-754?
>
>
> Vincent Lefevre wrote:
> > On 2009-04-13 09:19:34 -0400, Michel Hack wrote:
> >> If the underlying arithmetic is not correctly-rounding, it would be
> >> difficult to define "tightest bound" for the enclosures of primitive
> >> IA operations, which we are considering for the standaed. In this
> >> context I interpret "tightest" as saying that no narrower representable
> >> result would guarantee containment.
> >
> > Is the "tightest bound" required or only recommended? Having it
> > required would be annoying if one wants interval arithmetic on
> > the elementary functions, as IEEE 754-2008 only recommends
> > correct rounding for such functions.
>
> That's something to be decided. The Vienna proposal defines different
> tightness "levels" for arithmetic operations and elementary functions.
>
> Cheers,
> Christian
>
Christian is quite correct, that has yet to be decided
& is, once again, not part of Motion 4.
However, in fairness to this point, I believe that what
Michel is discussing is a 'tightest bound' that would
be optional to the user but an option that we might want
to make mandatory that the implementor provide.
Still, that has yet to be decided...
Dan