Xref: utzoo comp.compilers:1468 comp.lang.fortran:4038
Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!att!att!ima!esegue!compilers-sender
From: mcohen@amsaa-seer.brl.mil (Marty Cohen)
Newsgroups: comp.compilers,comp.lang.fortran
Subject: Re: IEEE 754 vs Fortran arithmetic
Summary: What is the intention of signed zero?
Keywords: Fortran, arithmetic
Message-ID: <1869@amsaa-seer.BRL.MIL>
Date: 30 Oct 90 13:16:15 GMT
References: <9010230628.AA22160@admin.ogi.edu> <BURLEY.90Oct24025053@world.std.com> <2408@charon.cwi.nl> <1990Oct28.030338.29498@cs.cmu.edu>
Sender: compilers-sender@iecc.cambridge.ma.us
Reply-To: Marty Cohen <mcohen@amsaa-seer.brl.mil>
Followup-To: comp.compilers
Distribution: na
Organization: Army Materiel Systems Analysis Activity
Lines: 11
Approved: compilers@esegue.segue.boston.ma.us

Does IEEE 754 give a purpose for signed zero?
I.E., is it to result from underflow?
If so, then the -pi, +pi result would be appropriate for atan(-+0,x).
--
Marty Cohen mcohen@brl.mil  {uunet|rutgers}!brl!mcohen
Custom House Rm 800, Phila. PA 19106 (215)597-8377
[IEEE 754 doesn't give much of a rationale for anything, it's pretty short.
-John]
-- 
Send compilers articles to compilers@esegue.segue.boston.ma.us
{ima | spdcc | world}!esegue.  Meta-mail to compilers-request@esegue.