Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site ucbvax.BERKELEY.EDU
Path: utzoo!watmath!clyde!burl!ulysses!ucbvax!brahms!weemba
From: weemba@brahms.BERKELEY.EDU (Matthew P. Wiener)
Newsgroups: net.arch,net.math
Subject: Re: IEEE f.p. standard
Message-ID: <13340@ucbvax.BERKELEY.EDU>
Date: Mon, 21-Apr-86 21:22:07 EST
Article-I.D.: ucbvax.13340
Posted: Mon Apr 21 21:22:07 1986
Date-Received: Wed, 23-Apr-86 22:02:40 EST
References: <271@euclid.warwick.UUCP> <186@ima.UUCP> <277@pyramid.UUCP> <289@zuring.UUCP> <2048@lanl.ARPA>
Sender: usenet@ucbvax.BERKELEY.EDU
Reply-To: weemba@brahms.UUCP (Matthew P. Wiener)
Organization: University of California, Berkeley
Lines: 14
Xref: watmath net.arch:3148 net.math:3112

In article <2048@lanl.ARPA> jlg@a.UUCP (Jim Giles) writes:
>Since PI/2 is not exactly representable, there is no representable value of
>X which will produce COS(X)=0 for all rounding modes.  Similarly, no value
>of X will produce SIN(X)=1 for rounding modes which truncate down.  This is
>an example of the type of problems that occur when dealing with
>trancendantal numbers on with only a subset of the rationals to calculate
>with.

This particular instance is trivially curable by using functions sinpi(),
etc.  which return sin(pi*( )), etc.  It's a bad habit to implement a
function in one particular way solely out of respect for the classical
notation.  I prefer implementations that respect standard usage.

ucbvax!brahms!weemba	Matthew P Wiener/UCB Math Dept/Berkeley CA 94720