Re: Motion P1788/M0009.01_ExactDotProduct

[Ulrich Kulisch <Ulrich.Kulisch@xxxxxxxxxxx>]

George and all:

Roughly speaking: Without an exact dot product interval arithmetic is a 
nice toy. With it, it is a superior arithmetical tool. The exact dot 
product is relevant not just for interval arithmetic. It brings accuracy 
and speed. Realization is not as complicated. The major software vendors 
would like to have it. See the letters of the IFIP Working Group on 
Numerical Software.

So, all those who are still undecided, please vote Yes.

We shall get it if we require it.

Best wishes
Ulrich Kulisch



Corliss, George schrieb:
> Reluctantly, I vote NO.
>
> I prefer to see the exact dot product as recommended, but not required.
>
> Simplicity.  If we are able to craft a standard on which we agree, there is
> no guarantee any commercial entity will implement it.  The simpler it is,
> both in description and in work to implement, the better are the chances
> someone will implement it, if only to have another checkmark in front of
> standards they support.  If the standard to too heavy, vendors can say,
> "We'll implement it as soon as our customers demand it."  That could be a
> long wait.
>

Attachment: DotProd-11.pdf
Description: Adobe PDF document

Attachment: IFIPWG-IEEE-P1788.pdf
Description: Adobe PDF document

Attachment: IFIPWG-IEEE754R.pdf
Description: Adobe PDF document