Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!isis!ico!rcd
From: rcd@ico.isc.com (Dick Dunn)
Newsgroups: comp.unix.sysv386
Subject: Re: OS costs
Summary: Let's narrow down the rumors if we can
Message-ID: <1990Sep16.224644.28322@ico.isc.com>
Date: 16 Sep 90 22:46:44 GMT
References: <36054@unix.cis.pitt.edu> <70400021@m.cs.uiuc.edu> <7925@gollum.twg.com>
Organization: Interactive Systems Corporation, Boulder, CO
Lines: 43

david@twg.com (David S. Herron) writes:
[tracing back a couple levels]
> > >[ ... ] MACH is going to become freeware.
> >...I find this difficult to believe, since currently MACH requires both a 
> >Berkeley and an AT&T source license...
...
> Believe it..
> 
> CMU's plan is to do whatever they can to free it from the little snippets
> of AT&T and Berkeley code that are left.

What is the "it" that is being freed?  If you mean the Mach kernel, you're
right, and for the 3.0 kernel it might even be right that there are just
"snippets" of AT&T and BSD code left.  But what about the OS interface and
all the libraries and applications?  In other words, suppose that you *did*
have a completely free Mach kernel...now what about the remaining 99% of
what it takes to make a usable system?  (99% is not an exaggeration.  It's
probably low.)  At the least, any module containing AT&T code has to be
rewritten from scratch.  This is NOT trivial, and there's more than
"snippets" of code to be dealt with here.

This keeps coming up.  The important thing to keep in mind is that a freed
kernel--especially a minimal kernel like the Mach goal--is *not* of much
use by itself.  Freeing the rest of the system is a gigantic task.  For
some comparison, Berkeley says they've got things slightly more than half
freed.  (They claimed "55%" in Anaheim).

Sure...it can be done; it probably will be eventually.  But the "free
UNIX-like system" is *not* just around the corner.  It won't happen in '90;
it probably won't happen in '91 either.

Speaking of Berkeley, could someone (David, perhaps, since you said it?:-)
explain why it would be desirable to free the code from Berkeley licensing?
The BSD license is about as cheap and un-restrictive as they come; it's the
next best thing to PD.  Where possible, they've released pieces of code that 
aren't subject to AT&T license; in any case, you get the code with little
more than a charge for media/copying, and a license that says little more
than "leave our copyright on, acknowledge us, don't hold us responsible,
now go have fun with it."  If CMU is trying to make a system generally
available, I think they'd be foolish to be at cross purposes with Berkeley.
-- 
Dick Dunn     rcd@ico.isc.com -or- ico!rcd       Boulder, CO   (303)449-2870
   ...I'm not cynical - just experienced.