Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1+some 2/3/84; site dual.UUCP
Path: utzoo!linus!decvax!decwrl!dual!fair
From: fair@dual.UUCP (Erik E. Fair)
Newsgroups: net.micro,net.college
Subject: Re: Free and undirected campus computing facilities
Message-ID: <872@dual.UUCP>
Date: Wed, 21-Nov-84 05:55:31 EST
Article-I.D.: dual.872
Posted: Wed Nov 21 05:55:31 1984
Date-Received: Thu, 22-Nov-84 07:21:53 EST
References: <457@utcsrgv.UUCP> <649@watdcsu.UUCP> <425@watcgl.UUCP> <420@ncoast.UUCP>
Organization: Dual Systems, Berkeley, CA
Lines: 91

>> From: bsa@ncoast.UUCP (Brandon Allbery)
>> Date: Sun, 18-Nov-84 08:38:59 PST
>> Organization: `Stamp Out MicroSoft!' on the North Coast
>>
>> We at ncoast offer a Unix (well, Xenix) system for public access, almost
>> unlimited (have to make room for the owner at work, and, of course, for
>> news :-) for any user who wants on.  (Unfortunately, 3-user Xenix without
>> source is not the best opportunity around...)
>> 
>> Anyway, I learned Unix hacking on here and just put the system back together
>> after a major system crash; others have done so in the past from stuff
>> learned working on here.  Is it possible that privately-owned systems
>> like ncoast could replace campus facilities for this purpose?

The short answer is ``no.''

The problem with small privately owned UNIX systems is universal: no source.
There are three ``public'' access UNIX systems in the Bay Area, of which
two are on USENET: proper and vectron. It so happens that both of them are
DUAL Systems (end plug).

My background in CS and UNIX hacking is from U.C. Berkeley, and I'm happy
to say that things there are improving (albeit *very* slowly) rather than
degrading as MIT and Waterloo are (or did).

When I arrived in the fall of 1980, the Berkeley Computer Club (a.k.a.
the Computer Science Undergraduate Association) had 60 restricted
access accounts on the Cory Hall PDP-11/70 (2.8BSD), and that was it,
unless you wanted to pay your way in the computer center on their
poorly maintained 11/70's (and pay you would!).  There was hot
competition for those 60 accounts, particularly for the top 20, which
were for `consultants' who had fixed hours that they would be in the
terminal room to help other users (and hack whatever; it was an excuse
to be on the machine during prime time hours). In addition to slightly
more machine access, there was a higher disk quota on the consultant
accounts.

The graduates weren't much better off; they all had Cory 11/70 accounts,
but the place to be was on Ernie Co-VAX the CSVAX, UCB's one & *only* VAX.
Vax accounts for undergrads was pretty much unheard of.

The birth of something real for UCB's undergrads came when Onyx Systems
donated one of their first machines to Berkeley to have Berkeley
utilites ported to it. The graduate students tired of it rather rapidly
because it wasn't even as fast as the 11/70, and curious undergrads
started getting accounts and doing strange things to the utilities and
the Kernel. This was encouraged largely by an enlightened graduate
student named Mark Horton (yes, the very same) with some behind-the-scenes
help from Professor Robert S. Fabry.

After many machinations (Mark graduated to Bell Labs, the Onyx had a
head crash and was dead for three months, political maneuvering of
various kinds) the Onyx was set up in the basement of Evans Hall (in a
terminal room that was being abandoned by the Computer Center & EECS)
as the first machine of the `Undergraduate Computing Facility.' The
basic policy was that any registered undergraduate of UCB could come
and get an account by filling out a form and showing his reg card to a
staff member. Of course, there was no way the Onyx could handle the 
total number of undergrads at UCB (~21,000 at that time), so we didn't
advertise our existence very widely (although USENET old timers may
remember the `ucbonyx' system hanging off of ucbarpa in days of yore;
as a minor point of hysterical fact, version B netnews was originally
developed on the ucbonyx, because it was the only place that Matt Glickman
could get an account).

The UCF has since grown to include a Vax730 (another machine that the
grads got tired of real quick, but it's a real vax running 4.2BSD), and
a Perkin-Elmer 3230. There is a dedicated cadre of hackers running
those machines with varied interests. Some are Kernel hackers, others
are Lisping along, still others are doing work in computer graphics.

The point, however, is that the UNIX hackers in the bunch wouldn't have
learned doodly squat without source on a machine that they could hack
and no one would care (much) if it died.  That's how I got my first
exposure to device drivers (trying to optimize the Onyx disk driver)
and to other parts of the v7 kernel. That experience, plus the fact
that I had been system manager for a year, got me my current job at DUAL
Systems when the University and I parted company in January of 1983. It
has stood quite a few of the other hackers who have departed from UCB
in good stead also.

In short, no source means that you miss a large part of what being a UNIX
hacker is all about.

	down with non-student run university computation centers,

	Erik E. Fair	ucbvax!fair	fair@ucb-arpa.ARPA

	dual!fair@BERKELEY.ARPA
	{ihnp4,ucbvax,hplabs,decwrl,cbosgd,sun,nsc,apple,pyramid}!dual!fair
	Dual Systems Corporation, Berkeley, California