[om] Re: OM-services

[Richard Fateman]

Paul Libbrecht wrote:
> 

> 
> Floating polynomials with an accepted (or declared
> after computation) inaccuracy should also be reached
> I believe.

I would not even assume this.  There is a difference in the
bigfloat arithmetic of Mathematica, Maple, Macsyma, Reduce, Derive, Mupad, Axiom
 such that
getting the precisely identical answer from every straight-line
program ... a sequence of adds and multiplies ... would be
difficult. Getting the precisely identical answer for integer
or rational operations would be, I think, easy.

If your specification is loose enough that you are willing
to admit any answer as long as the error is not greater than
some bound, I think this is alleged to be the case in
Mathematica 4.0 but not Mathematica 3.0 for N[..] sometimes.
I suspect that no other system has this particular facility
built in, but it could presumably be done in any of them. A
rigorous bound on the error in evaluating a polynomial can
be computed by merely evaluating another (related) polynomial
whose coefficients are all positive.  


> 
> But I know the war is going to be difficult and your
> experience is certainly one of the most (the only one ?)
> concrete around here. May I request another chance ?
> 

I can't stop you from trying!

RJF
--
om at openmath.org  -  general discussion on OpenMath
Post public announcements to om-announce at openmath.org
Automatic list maintenance software at majordomo at openmath.org
Mail om-owner at openmath.org for assistance with any problems