Xref: utzoo comp.lang.scheme:1823 comp.lang.lisp:4017
Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!usc!samsung!munnari.oz.au!goanna!ok
From: ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe)
Newsgroups: comp.lang.scheme,comp.lang.lisp
Subject: Re: Numerical gotcha's
Keywords: Interval-Arithmetic Bignums Rationals
Message-ID: <4361@goanna.cs.rmit.oz.au>
Date: 23 Nov 90 07:56:19 GMT
References: <431@data.UUCP>
Followup-To: comp.lang.scheme
Organization: Comp Sci, RMIT, Melbourne, Australia
Lines: 18

In article <431@data.UUCP>, kend@data.UUCP (Ken Dickey) writes:
> ;; ELK 1.1: f = 1.18059162071741e+21 [Sun3]	??? {floats only}
> ;; THE CORRECT ANSWER IS:
> ;;  (tryit) ->  -54767/66192	{Gambit 1.4; sun3; rationals+bignums}

On an Encore Multimax, T 3.0 gave the correct answer.
However, when the numeric literals were changed to floating point
(e.g. 333.75 instead of (/ 1335 4), 121.0 instead of 121) and the
x, y arguments were written as floats, the answer came out at
roughly -1.1806e21.  Since the Sun3 and the Encore both provide
IEEE 754 arithmetic, I imagine other IEEE 754 machines will give
the same answer.

The moral is (if (and (inexact? x) (not (numerical-analyst? *you*)))
		 (doubt *you* x))

-- 
I am not now and never have been a member of Mensa.		-- Ariadne.