Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!uunet!fernwood!portal!cup.portal.com!pgl
From: pgl@cup.portal.com (Peter G Ludemann)
Newsgroups: comp.lang.prolog
Subject: Re: arithmetic overflow checking
Message-ID: <29741@cup.portal.com>
Date: 9 May 90 03:10:15 GMT
References: <870@cluster.cs.su.oz>
Organization: The Portal System (TM)
Lines: 16

I would assume that not checking for arithmetic overflow is
part of a misguided attempt to be "efficient".  Getting a
wrong answer in half the time does not impress me, but evidently
it does impress some people, especially in a "Very High Level Language".

C, of course, is a primary offender; some C implementations do not
make detecting integer overflow easy.

It would be interesting to find out which Prologs support arithmetic
overflow detection and whether the ISO standard intends to insist on
checking for overflow.  By-the-way, IBM-Prolog checks for overflow
even though it's unlikely (rational numbers that fit into 2**24 bytes
are supported).  Any other Prologs which support overflow detection
and/or big numbers?

- Peter Ludemann	--- standard disclaimer ---