The Berkeley Software Distribution
==================================
UNIX is always litigious
by Bradford Morgan White
Feb 5, 2024

The first public presentation of UNIX was made at the Symposium on
Operating Systems Principles at the IBM Research Center in Yorktown
Heights in October of 1973. Dennis Ritchie is quoted as saying it was
beautiful day, and Ken Thompson layered his own memories with a thick
coating of modesty:

> The audience was several hundred. I was pretty nervous. The
> response was the normal, polite applause. I don't recall any
> questions.

[The IBM Research Center in Yorktown Heights, image from IBM]

<https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,
q_auto:good,fl_progressive:steep/
https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2F
images%2Fb9ace03e-5fa3-43df-875b-6989e389b156_1200x675.jpeg>

In contrast to Thompson stating that he didn't recall any questions...
the two were immediately asked for copies of the operating system,
and this wasn't a simple matter for AT&T. The American Telephone and
Telegraph company had been established as a legal, nation-wide
monopoly in the USA via the Kingsbury Commitment in late 1913. This
position was further cemented during World War I when the United
States' federal government nationalized the phone system. Following
the end of the war, the phone system went back into the hands of AT&T
and the company achieved some rather remarkable regulatory capture
with the Willis Graham Act of 1921 and the Communications Act of
1934. This complicated legal history presented a very serious
question to AT&T's legal department when people began asking for
UNIX: were computer operating systems part of the common carrier
services of the phone company and therefore required to be
distributed? If they were not, then the company needn't distribute
UNIX at all, but if they were indeed, then it was only a matter of
time before the FCC would force AT&T to distribute UNIX. In the end,
the decision was made to distribute UNIX to universities and research
centers at the cost of the media plus shipping. Somehow, quite
magically, this resulted in a nice round number of $150.00 (or around
$927 in 2024) for Katholieke Universiteit in Nijmegen, The
Netherlands in December of 1974. There are some rather important
points within the license that Katholieke was granted. Licensees were
granted source code for the operating system as computer systems of
the day weren't standardized in any meaningful way. The license then
granted free use and modification within the university, but
disallowed any spread outside. Specifically, the license mentioned
that employees and students had access.

[UNIX license from Western Electric to Katholieke]

<https://www.abortretry.fail/api/v1/file/
ea99ff77-8f06-4c00-ae95-2a1f60410a44.pdf>

At this point in the computer industry, user groups were somewhat
common. IBM had SHARE, and that had inspired similar groups around
DEC, Burroughs, Rand, and so on. It was therefore somewhat natural
that a group would form around UNIX. Thus, Mel Ferentz and Lou Katz
organized a meeting of UNIX users in New York on the 15th of May in
1974. Around twenty people were in attendance, and by this time there
were just over thirty UNIX installations outside of AT&T and its
subsidiaries. This user group grew to become USENIX over time.
Following the user group's formation, a mailing list started. From
the first list on the 30th of July in 1975, we have the following
organizations listed as installation/user sites: AT&T, Brooklyn
College, Carleton College, Case Western Reserve University, The
Children's Museum, City University of New York, Columbia University,
Duke Medical Center, East Brunswick High School, Harvard University,
Hebrew University of Jerusalem, Heriot-Watt University, Johns Hopkins
University, Knox College, Naval Postgraduate School, Oregon Museum of
Science, Polytechnic University of NY, Princeton University, Rand
Corporation, St. Olaf College, Stanford University, The Spence
School, University Catholique de Louvain, University of Alberta,
University of California (Berkeley), University of Manitoba,
University of North Carolina, University of Saskatchewan, University
of Texas (Dallas), University of Toronto, University of Utah,
University of Waterloo, and the University of Wisconsin. As we know
from the license granted to Katholieke, there were more UNIX user
sites than this, but they weren't members of the user group (or at
least not at first).

<https://www.abortretry.fail/p/the-history-of-gm-naa-io-and-share>

[Dennis Ritchie (standing), Ken Thompson at the teletype, PDP-11
1972]

<https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,
q_auto:good,fl_progressive:steep/
https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2F
images%2F3926f41f-c121-450b-8e10-e46ccc1c5448_652x522.jpeg>

In this early time period, UNIX only ran on the PDP-11, but that
changed at Princeton where UNIX was ported to the IBM 360 in 1976.
The next target was the Interdata 8/32 in 1977 which was undertaken
by Ritchie and Steve Johnson (author of yacc, lint, pcc). But,
porting efforts took off only after John Lions at the University of
New South Whales wrote a commentary on the UNIX sources and
distributed them as a book, Code and Commentary, for teaching
students about operating systems. Western Electric tried to stop
dissemination, but this was apparently impossible. Likewise,
modifications of UNIX began circulating following the release of
Lions' book, and a culture we would recognize today as “open source”
began to develop helped in no small part by AT&T's policies regarding
UNIX that essentially stated UNIX would have no advertising, no
support, no bug fixes, and payment in advance. To color the time
period more thoroughly, Bob Kahn and Vint Cerf published the first
description of TCP/IP in 1974 and by January of 1976 there were sixty
three hosts on ARPAnet, and UNIX while being used globally would run
only on hardware that cost over $9000 (around $48000 in 2024).

Professor Robert (Bob) Fabry was at the Symposium on Operating
Systems Principles where UNIX had first been announced and he was
very excited. Returning to UC Berkeley where he was then employed, he
assembled a group to purchase a PDP-11/45. As this was a large
purchase, he coordinated the departments of computer science, math,
and statistics. With the machine purchased, Fabry then ordered a tape
of UNIX from Thompson. The actual installation of UNIX was first
undertaken by Keith Standiford in January of 1974. In 1973/1974, it
was somewhat routine for Thompson himself to be involved in nearly
every UNIX installation for a licensee. The folks at Berkeley seemed
to be interested in doing everything themselves, but things didn't go
well. Eventually, Standiford reached out to Thompson, and Thompson
would connect to the University's 11/45 over a three hundred baud
acoustic coupler to remotely debug crash dumps from New Jersey. I
personally like to imagine that it was a Novation CAT 300, but I
haven't been able to find a model number of the modem, and I haven't
found any reference to what system Thompson was actually using. Plus,
the Novation CAT 300 wouldn't be released for several years.

Following the purchase of the PDP-11/45, the departments involved
began having issues with scheduling time on the machine. Berkeley
bought several more computers. One of these was a PDP-11/70, and its
arrival coincided with the arrival of Thompson as a visiting
professor. Thompson, Bob Kridle, and Jeff Schriebman then setup V7
UNIX on the 11/70. Shortly after the installation was completed, two
graduate students, Chuck Haley and William Nelson (Bill) Joy, arrived
on campus. They were intrigued by the computer system, and they began
hacking on Thompson's Pascal compiler. Within a few weeks (from what
I have been able to find), the teletypes attached to the 11/70 were
replaced with ADM-3 screen terminals.

[ADM-3, image by Chris Jacobs, CC BY-SA 3.0]

<https://substackcdn.com/image/fetch/f_auto,q_auto:good,
fl_progressive:steep/
https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2F
images%2F98f22ee6-2f34-4ffd-b8bd-b37fb059cbe5_1887x1860.jpeg>

For Bill Joy, using ed or em on a screen terminal wasn't really
sufficient. He took a detour from hacking on Pascal, and he created
the ex editor. Together with Pascal, the V7 UNIX at Berkeley was
notably better than other UNIX systems of the time. In early 1978,
Bill Joy began offering the Berkely Software Distribution. The first
copy we know to make it out of Berkeley was to Tom Ferrin at UCSF on
the 9th of March in 1978. The license was signed on the 13th, the
media was an 800 bpi tape, and on the tape was the “Unix Pascal
system” and the “Ex text editor.” Credits were made to W.N. Joy, S.L.
Graham, C.B. Haley, K. Thompson for Pascal, and to W.N. Joy for Ex.

[Bill Joy, Silicon Valley Visionary, on the Future of Batteries ...]

<https://substackcdn.com/image/fetch/f_auto,q_auto:good,
fl_progressive:steep/
https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2F
images%2Fd5b8a872-0970-4832-8b53-86c19f7c8543_1200x630.jpeg>

Ben Collver <bencollver@tilde.pink> writes:

> The Berkeley Software Distribution
> ==================================
> UNIX is always litigious
> by Bradford Morgan White
> Feb 5, 2024

Great article. Thanks for posting.

[...]

> While 4.2BSD was successful and mostly well received, it did get some
> complaints. The majority of the complaints were centered on
> performance. The team then spent two years improving performance,
> refining the networking stack, and they felt that they were ready to
> announce an impending release at USENIX in June of 1985. This didn't
> go well. The fine folks of Bolt, Beranek, and Newman (part of the
> steering commitee) noted that 4.2BSD had shipped without the final
> version of the their networking code, and was instead using a heavily
> modified version of their initial prototype. After some bickering
> back and forth, DARPA provided both network stacks to Mike Muuss
> (author of ping) of the Ballistics Research Laboratory for
> testing.erkeley's code was better. 4.3BSD was released in June of
^^^^^^^^^^^^^^^^^

Berkeley's code was better.

[...]

> Of course, this didn't sit well with Unix System Labs (AT&T
> subsidiary). Shortly after BSDi began sales, they received a cease
> and desist from USL with a particular request to stop using the phone
> number that included “it's UNIX,” as the ownership of the UNIX
> trademark was firmly in USL's hands. BSDi complied changing their
> advertisements, the number, and explaining that BSD wasn't precisely
> UNIX. This, however, wasn't quite enough. USL brought a lawsuit
> against BSDi seeking an injunction against the sale of BSD/386. As
> part of the suit, USL claimed that BSDi's product contained USL code
> and trade secrets, and that further sales of BSDi's product would
> irreparably harm USL. BSDi then claimed that they shouldn't be held
> liable for any code contained in Berkeley's original source
> offerings, but that they were completely willing to discuss the six
> added, BSDi-original files. BSDi's argument won and USL was required
> to restate their complaint or have the case dismissed. USL then filed
> suit against both BSDi and the University of California with roughly
> the same complaints but this time seeking anle and distribution of both
^^^^

It turns out some passages have been left out when you pasted the
article here. This whole period should have been:

USL then filed suit against both BSDi and the University of California
with roughly the same complaints but this time seeking [an injunction
against the sale] and distribution [...].

> On the 21st of December in 1992, Novell announced that it would be
> acquiring Unix System Laboratories including the UNIX copyright,
> trademarks, and licensing contracts. The LA Times stated that this
> transaction was completed with an exchange of stock wherein all of
> the shares of USL would be traded for twelve million three hundred
> thousand ser which USL would be a wholly owned subsidiary of Novell.
^^^

Should have been ``shares''.

The LA Times stated that this transaction was completed with an
exchange of stock wherein all of the [shares of USL would be traded
for twelve million three hundred thousand shares of Novell, and after]
which USL would be a wholly owned subsidiary of Novell.

> From: <https://www.abortretry.fail/p/the-berkley-software-distribution>

On 2024-02-19, Julieta Shem <jshem@yaxenu.org> wrote:
> It turns out some passages have been left out when you pasted the
> article here. This whole period should have been:

Thanks for catching the paste-o's. My original text file has those
passages in it. I am surprised to have this trouble with xclip and
paste. Next time I'll just :r the file into my buffer.

Best regards,
-Ben

On Sun, 18 Feb 2024 18:27:08 -0000 (UTC), Ben Collver wrote:

> A few months after the NetBSD group formed, the FreeBSD group formed.

One wonders why they had to fork off in different directions, and not
continue development from a common base. An overly-centralized development
model, perhaps?

Consider that there are at most maybe half a dozen BSD variants currently
alive, as compared to 50× that number of Linux distros. Yet it is easier
to move among Linux distros than it is to move between BSD variants.

And those Linux distros cover a wider variety of usage scenarios than the
BSD variants can offer. So Linux can offer variety with minimal
fragmentation, while the BSDs suffer from more fragmentation while
offering less variety.