Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!mips!pacbell.com!pacbell!att!dptg!ulysses!andante!alice!ark
From: ark@alice.UUCP (Andrew Koenig)
Newsgroups: comp.lang.c
Subject: Re: Floating-point non-exactness
Message-ID: <11132@alice.UUCP>
Date: 3 Aug 90 14:49:06 GMT
References: <25837@mimsy.umd.edu> <26722@pasteur.Berkeley.EDU>
Organization: AT&T Bell Laboratories, Liberty Corner NJ
Lines: 12

In article <26722@pasteur.Berkeley.EDU>, hilfingr@tully.Berkeley.EDU (Paul Hilfinger) writes:

> Assume the machine rounds
> literals deterministically to the nearest representable FP number.

An optimistic assumption indeed.  To do that, you need unbounded
precision in your input conversion routine.  It's hard enough to
do floating point input conversion that always gives you one
of the two nearest neighbors of the input value.
-- 
				--Andrew Koenig
				  ark@europa.att.com