Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!algor2!jeffrey
From: jeffrey@algor2.uu.net (Jeffrey Kegler)
Newsgroups: comp.unix.wizards
Subject: Re: What kinds of things would you want in the GNU OS?
Message-ID: <1989Aug4.080124.26687@algor2.uu.net>
Date: 4 Aug 89 08:01:24 GMT
References: <160@uucs1.UUCP>
Reply-To: jeffrey@algor2.UUCP (Jeffrey Kegler)
Organization: Algorists, Inc.
Lines: 51

I have (obviously) taken some time to this over because it is a big
question, and one likely to be important in our field.

In order of importance, I think the features of a GNU OS should be:
1) POSIX and X/OPEN compatibility
2) SVID compatibility
3) Berkeley compatibility
4) Some solution to the problem of kernel size, such as a message
passing mini-kernel with configurable modules.
5) Other new features

I give new features a low ranking, not because I do not think many
would not be nice, or even necessary, but because I think the first
goal has to be to get people using the GNU OS.  Because it will be
open, it will immediately attract a lot of research attention.  Many a
PhD thesis on some aspect of operating systems will feature an
implementation in GNU OS.  So, once GNU OS takes off, a cornucopia of
new features should quickly become available.  When AT&T stopped
shipping source with every UNIX the dark ages of UNIX began.  The day
GNU OS is announced may well mark the end of the Dark Ages.

So the first object should be to get GNU OS a critical mass of
acceptance.  In justifying use of GNU OS for their development to my
clients, compatibility with industry standards will be crucial.  I
gave POSIX and X/Open top billing because they are real standards, not
descriptions of implementations.  I ranked SVID above Berkeley not as
a technical judgement, but because SVID is more widely accepted as
standard commercially, in my opinion.

The kernel size problem has become so major that it outranks all other
new features in my mind.  What with networking, the increasing variety
and complexity of devices, etc., having to put all kernel functions
into a single executable is getting very burdensome.  V6 UNIX was
elegant.  The kernel I now run is an ugly ungainly thing.  I would
like to see the elegance of V6 return (without losing all the new
features added since V6).

Since GNU OS is done with volunteer labor, a certain amount of work on
new research features may be necessary as payment to the type of
talent the project will attract.  Except as necessitated by this
consideration, I think new features should be excluded from the first
release of GNU OS.

From what I can gather of what is going on at FSF, they are well ahead
of me on the above considerations, and I feel that we can look forward
to great things with the release of GNU OS.
-- 

Jeffrey Kegler, Independent UNIX Consultant, Algorists, Inc.
jeffrey@algor2.UU.NET or uunet!algor2!jeffrey
1762 Wainwright DR, Reston VA 22090