Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!world!iecc!compilers-sender
From: richarda@cec825.atl.hp.com
Newsgroups: comp.compilers
Subject: INFs and NaNs
Keywords: arithmetic, dataflow
Message-ID: <91-06-024@comp.compilers>
Date: 17 Jun 91 19:03:50 GMT
References: <91-06-016@comp.compilers>
Sender: compilers-sender@iecc.cambridge.ma.us
Reply-To: richarda@cec825.atl.hp.com
Organization: Compilers Central
Lines: 14
Approved: compilers@iecc.cambridge.ma.us

In article <91-06-016@comp.compilers>, bill@hcx2.SSD.CSD.HARRIS.COM writes:
>As far as I know, there is only _one_ kind of mathematics.  The FORTRAN
>standard is referring to real mathematics, not any particular means of
>approximating mathematics.

Dealing with error values (such as INF, divzero, etc.) does not necessarily
preclude one from "mathematicality."  I seem to recall that the folks working
on dataflow architectures rigorously defined the math for propogating error
values so that traps, etc., would not be necessary.  I don't recall the author
but I believe the title of the paper  was Error Values in Val.

-- 
Send compilers articles to compilers@iecc.cambridge.ma.us or
{ima | spdcc | world}!iecc!compilers.  Meta-mail to compilers-request.