Re: Missing functions tanPi, asinPi and acosPi

[Christoph Quirin Lauter <christoph.lauter@xxxxxxxxxxx>]

Jeff Kidder wrote on 12.04.2008 01:15:
> On Apr 11, 2008, at 3:43 PM, Vincent Lefevre wrote:
> > On 2008-04-11 11:56:08 -0700, Jeff Kidder wrote:
> > > tanPi was removed at some point.
> >
> > removed? In which draft was it present?
...
> There really are quite a lot of functions that are not included by name.
> I just ran across sqrtp1m1(x) = sqrt(x+1)-1. We don't have secant. ....
>
> And yes, this is an argument made by those who argued against this whole 
> endeavor.

No, I think what this all tells us is that would have adopted 
nevertheless an approach where a compliant correctly rounded 
implementation for any mathematical function can be derived from the 
mathematical properties of the function by simple rules. I have always 
been against huge table, that, anyway, leave out an infinite set of 
functions.

See

http://perso.ens-lyon.fr/christoph.lauter/revisionIEEE754.txt

and

http://perso.ens-lyon.fr/christoph.lauter/revisionIEEE754_2.txt

for details on out proposal at the time.

Regards,

Christoph