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

Greg Ewing greg@cosc.canterbury.ac.nz
Thu, 03 Oct 2002 15:24:16 +1200 (NZST)

Raymond Hettinger <python@rcn.com>:

That is unless you wanted more precision than comes with a float. Normalization to some arbitrary maximum denominator magnitude would be a more general purpose solution.

Wouldn't a float type with specifiable precision do just as well for that, and be more efficient?

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 +--------------------------------------+