RE: [Axiom-developer] infinite floats domain

[Bill Page]

On June 12, 2005 6:18 PM Tim Daly wrote:

> you're right, of course. 
> my mind is mush at the moment. 
> i've been driving all week.

Yah, that'll do it to ya ... :) How's CMU?

Seriously, I think that there *are* some issues that
should be dealt with here. There may very well be some
situations that could make more efficient and/or effective
use of Axiom's "infinite precision" floats. I expect that
just as in the case of "infinite" integers, there may be
some better ways and some worse ways of doing calculations
with this objects.

For that matter, mathematically speaking just what is Axiom's
Float type? Is it formally equivalent (though obviously not
computationally equivalent) to FRAC INT? How does it relate
to standard (IEEE ?) definitions of floating point? How does
it differ mathematically from the reals, c.f. RealClosure,
etc.

Regards,
Bill Page.