Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!ll-xn!ames!amdahl!bnrmtv!connery
From: connery@bnrmtv.UUCP (Glenn Connery)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Datalight? vs. ????
Message-ID: <2485@bnrmtv.UUCP>
Date: Thu, 3-Sep-87 13:08:02 EDT
Article-I.D.: bnrmtv.2485
Posted: Thu Sep  3 13:08:02 1987
Date-Received: Sat, 5-Sep-87 16:47:01 EDT
References: m/pc/general <682@cup.portal.com>
Organization: Bell Northern Research, Mtn. View, CA
Lines: 26
Summary: Yes its a perfectly reasonable compiler

In article <682@cup.portal.com>, Aron_Fingers_Nelson@cup.portal.com writes:
> I was reading an ad by datalight challenging microsoft to a test, does
> anyone know or has heard anything of it? Is datalight a viable
> challenger to turbo c or microsoft 5.0? Is it any good period?

Datalight is a perfectly good compiler, and definitely up for the current
compiler wars.  The basic compiler is very fast and produces very good
code.  The Optimizer is an optional step and unlike Microsoft, when you
turn it off exacts no time penalty.  But with the optimizer on, compile
times are very similar to Microsoft C 4.0 (ick).  However, this is a
serious optimizer--loop invariants, common subexpressions, etc.  The
DataLight Optimizer and Microsoft C 5.0 are breaking a lot of the compiler
benchmarks which do nothing, and therefore compile to nothing.

All that said, I am using Microsoft C and Turbo these days.  Microsoft
largely because I'm used to it, like Codeview, think ANSI C is a godsend,
may be working with OS/2 etc, and Turbo because its highly compatible with
Microsoft, cost very little and compiles much much faster.  Datalight
started with Lattice C as a model but is evolving to the ANSI C stuff
over time.

Glenn
-- 

Glenn Connery, Bell Northern Research, Mountain View, CA
{hplabs,amdahl,3comvax}!bnrmtv!connery