Re: Table 4 proposal version 0.2...
On 2012-03-23 11:34:02 -0400, Michel Hack wrote:
> Vincent Lef?vre replied to me:
> > > The problem here is that the 754 representations of Zero,
> > > Infinity (and also NaN) are overloaded.
> >
> > and +1, because
> > powr (+1, ?inf) signals the invalid operation exception.
> > because +1 is regarded here as a potentially inexact number
> > (contrary to the usual exact inputs).
>
> Not at all! powr(x,y) is defined as exp(y * log(x)), so it
> is simply an instance of Inf*0 which is NaN. All inputs are
> assumed exact (well, it's exp(NaN) but that is also NaN).
but with the same effect. In general, a function should be
defined from its math properties, not by its (high- or low-level)
implementation.
--
Vincent Lefèvre <vincent@xxxxxxxxxx> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)