Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!purdue!spaf
From: spaf@cs.purdue.edu (Gene Spafford)
Newsgroups: gnu.gcc
Subject: Re: FSF Hypocrisy
Message-ID: <6862@medusa.cs.purdue.edu>
Date: 1 Jun 89 03:22:24 GMT
References: <2129@internal.Apple.COM> <1107@mailrus.cc.umich.edu>
Sender: news@cs.purdue.EDU
Reply-To: spaf@cs.purdue.edu (Gene Spafford)
Organization: Department of Computer Science, Purdue University
Lines: 68

Here's another non-technical posting.  If there was a gnu.discuss
or something similar, I'd use it.  Until then, this follows a
previous posting here.....

In article <1107@mailrus.cc.umich.edu> honey@citi.umich.edu (Peter Honeyman) writes:
>wouldn't it be great if we could use fsf stuff without having to deal
>with fsf politics?

He got some posted abuse for that.  However, his point bears some
further discussion, or at least thought.

Right now, it looks like a number of researchers I know, including a
few here, are going to have to try to get financing to build a TRUE
PUBLIC DOMAIN C compiler.  We can't use gcc for our research.  Let me
explain. 

Like most researchers, I don't want to spend a lot of time building
new tools when existing ones are available.  I also don't want to
spend a fortune buying a commercial tool, if I can help it.  Now
suppose I want to build something that needs to instrument some C code
in some way -- maybe I need to include certain kinds of profiling
commands, or new mechanisms for debugging.  Gcc looks attractive -- it
runs on my machine, and it doesn't cost my research grants any money.
But what happens if I use it?

Well, ignoring the fact that it's still in Beta test after months and
months and months, and updates (and new bugs) just keep coming, there
is problem of the copyleft.  If I use the gcc compiler for my project
and want to provide it to some companies to test against "real world"
data, they won't take it.  The folks at Bellcore, AT&T, IBM, DEC, and
a bunch of others have already told me that they won't.  Why?  Because
their corporate lawyers don't want to take any chance at all that
internal project code is "contaminated" with copyleft software.  They
may have no plans whatsoever to market or otherwise commercially use
my prototypes, but they would "contaminate" their labs if they
accepted gcc-based software.  That leaves me out in the cold.  I can't
get them to test my prototypes under real conditions, and they won't
be interested in funding my research if they can't test the results.

Now, if I really wanted to force the issue, every lawyer I've talked
to says that the copyleft wouldn't last a minute in court.  But I'm
not interested in pursuing that, and neither are the major companies
-- it's bad publicity.  Especially so since RMS has threatened to fold
up shop and stop work on GNU software.  A lot of people would lose on
that.  Besides, none of those companies wants to fight that hard for
software that really doesn't have anyone committed to its correct and
continued functioning...except political fanatics/idealists; I'm
certainly reluctant to base anything major on a compiler that has such
spotty quality control.  Of course, if a government agency decides to
challenge the copyleft, that's it -- they don't care about public
sentiment in a case like this.

If I had access to a stable compiler that I could use, that had a
copyright but waived all restriction on further use of the code, I
would snap it up in a minute and dump gcc -- even if it meant I had to
scrounge up a few $1000 to pay for it.  There are differing realities
in the world; idealism is a wonderful thing but it doesn't always
match the majority of our realities.  Some of us have to deal with the
world as it is, not how we might like it to be.  I think it painfully
ironic that the politics of an effort to create an environment of
free, shared software should lead some of us working on the next
generation of software tools to wish that we could buy some software
we need instead of use the free software available.
-- 
Gene Spafford
NSF/Purdue/U of Florida  Software Engineering Research Center,
Dept. of Computer Sciences, Purdue University, W. Lafayette IN 47907-2004
Internet:  spaf@cs.purdue.edu	uucp:	...!{decwrl,gatech,ucbvax}!purdue!spaf