[Python-Dev] Re: PEP239 (Rational Numbers) Reference Implementation and new issues

[Greg Ewing]

Andrew Koenig <ark@research.att.com>:

> The real point is to have small, efficient rational representations
> for floating-point numbers, even those with large exponents.

But how long are they going to *stay* small and efficient,
once you start doing arithmetic on them?

The problem as I see it is not the magnitude of the number, but that
the number of significant bits grows without bound, if you never throw
any of them away.

Greg Ewing, Computer Science Dept, +--------------------------------------+
University of Canterbury,	   | A citizen of NewZealandCorp, a	  |
Christchurch, New Zealand	   | wholly-owned subsidiary of USA Inc.  |
greg@cosc.canterbury.ac.nz	   +--------------------------------------+