Path: utzoo!attcan!uunet!samsung!caesar.cs.montana.edu!ogccse!emory!att!dptg!pegasus!psrc
From: psrc@pegasus.ATT.COM (Paul S. R. Chisholm)
Newsgroups: comp.lang.c++
Subject: Re: Naive C++ vs. G++ question
Summary: products vs. programming languages, with a minimum or religion
Message-ID: <4280@pegasus.ATT.COM>
Date: 29 Nov 89 06:12:21 GMT
References: <34623@cornell.UUCP>
Distribution: comp
Organization: AT&T Bell Laboratories
Lines: 54

In article <34623@cornell.UUCP>, beck@hermod.cs.cornell.edu (Micah Beck) writes:
> [C++] is a translator licensed by AT&T, [G++ is] a PD compiler.

C++ is a programming language.  The AT&T C++ Language System, Release
2.0, is an implementation of that programming language.  The price
(except to universities) is *not* cheap!  It's sold only in source
form, largely to companies that in turn will produce end user
products.

G++ is a copyrighted (not public domain), pretty much freely
distributable implementation of a large subset (or maybe a superset of
a large subset) of C++.  (The "pretty much distributable" part refers
to the fact that if you pass along a binary copy, the copyright claims
you're obliged to make sure the recipient can also get source.)  There
are no restrictions on what you do with the output of G++, as long as
you use only the vanilla libraries.  If you use libG++, things get more
complicated.

> I would like to use the Oops library of objects, but it is labeled
> as written for C++.

Yup.  You'll need some implementation of the C++ language.  (It's
called the NIHlib nowadays, which I think is a nice pun.  It was
developed at the National Institutes for Health, and it helps fight the
dreaded "Not Invented Here" syndrome.)

> The licensing fee for C++ seems like an impediment to wide
> distribution.

The licensing *agreement* for the AT&T C++ Language System, Release
2.0, is an impediment to unlimited distribution.  Ditto for the
licenses for Zortech C++ or Sun's C++ compiler or Turbo C, or the
copyright on Dr. Stroustrup's book, or lots of other things in the
world.

> I am tempted to try and use G++ with the Oops classes and hope the result
> will run under both C++ and G++ if I stick to simple language features.
> Is there any hope that this will work?

Some.  Various attempts at sticking to a common subset have been
described on the net; keep listening.

> Should I just stick to ANSI C?

A friend of mine, Tony Hansen (author of THE C++ ANSWER BOOK) has been
marshalling arguments for our management that C++ is "a better C" than
ANSI C, just as ANSI is a better C than old K&R.  The data abstraction
facilities add more value; the object oriented features, still more.

> Micah Beck, beck@cs.cornell.edu

Paul S. R. Chisholm, AT&T Bell Laboratories
att!pegasus!psrc, psrc@pegasus.att.com, AT&T Mail !psrchisholm
I'm not speaking for the company, I'm just speaking my mind.