Re: Alternate floating-point results under directed rounding
Dear Van
Van Snyder wrote on 2008/11/10 Mon PM 08:17:31 GMT:
...
> > Actually I say it should return NaI. Empty can be a genuine, correct
> > result, whereas NaI is _always_ an indicator that something went
> > wrong.
>
> I don't understand the advantage of NaI as the result of a computation
> such as 2.0*[0.25*HUGE,0.75*HUGE], instead of returning [0.5*HUGE,inf]
> with overflow signaling.
...
The above would not give NaI. It has the same effect as
intval(2.0)*[0.25*HUGE,0.75*HUGE]
= [2.0,2.0]*[0.25*HUGE,0.75*HUGE]
= [0.5*HUGE,inf]
as you suggest. I'm sorry if I wasn't sufficiently clear. I was talking about point values that result from a computation (as
Baker was discussing); which are large enough that they might overflow; and that then become part of an interval
expression. Something like x*[0.5,1.5] where x might exceed HUGE.
Cheers
John
-----------------------------------------
Email sent from www.virginmedia.com/email
Virus-checked using McAfee(R) Software and scanned for spam