Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ames!think!rst
From: rst@think.COM (Robert Thau)
Newsgroups: comp.misc
Subject: Re: The GNU Manifesto
Message-ID: <14302@think.UUCP>
Date: 1 Jan 88 04:47:25 GMT
References: <153@mozart.UUCP> <19303@clyde.ATT.COM>
Sender: usenet@think.UUCP
Reply-To: rst@pozzo.think.com.UUCP (Robert Thau)
Organization: Thinking Machines Corporation
Lines: 59

In article <19303@clyde.ATT.COM> rcj@moss.UUCP (Curtis Jackson) writes:

... basically, that the politco/economic parts of the GNU Manifesto
don't seem to make a terrible lot of sense.  It's hard to argue; as it
stands, the document is more than a little incoherent.  However, there
are a few potshots in the article which are unfair to Stallman and the
Free Software Foundation.

First off, free software, qua free software, doesn't necessarily lead
to chaos.  GNU Emacs has certainly been around long enough for
multiple, incompatible versions to become a problem, yet they haven't.
One occasionally sees divergent streams of development; however,
changes which are substantial and worthwhile eventually find
themselves on the distribution tape.  (For example, the VMS port had a
separate existence while it was in development, but it has since been
merged back).

There are other large pieces of essentially free software that have
survived like this as well; B News leaps to mind.  In fact, there is
*almost* an example of a free operating system.  While MINIX isn't
free, the source code is readily available as such things go, and
there is already a large community devoted to hacking it.

What the stable free programs I've seen (patch, rn, netnews, GNU
Emacs, Microemacs, nethack, ad nauseam) all have in common are:

  *) an authority figure or organization which takes responsibilty for
     evaluating and incorporating new features, documenting changes,
     and releasing new versions.  This is often the original author,
     but not always.

  -- and --

  *) channels of communication which allow the user community
     (including hackers) to get new versions easily.

(BTW, this is where money sneaks in.  Something has to be motivating
the coordinators-and-distributors-of-releases to do what is ultimately
a difficult, tiresome, and often thankless job.  Devotion to the good
of the community may be enough for some people.  Cash seems to be more
universal). 

Secondly, in re:

>You ever seen what happens to a system when you turn even very talented
>students loose on the system code?  I have...

Stallman has; the MIT AI lab was like this at one point.  What he
doesn't seem to realize is how lucky he was.  Such an environment does
work, sort of, if the people are extremely competent, exceptionally
tolerant of system breakage, and exceptionally intolerant of people
who wade into system code without knowing *exactly* what they're
doing.  For the rest of us, Stallman's attitude towards security is a
little unreasonable.

As for me: I don't really see anything wrong with letting a student
play around with a copy of, say, the C compiler.  However, if it's the
copy that *I'm* using, we're talking justifiable homicide ...

rst