• The Berkeley Software Distribution (1/2)

    From Ben Collver@21:1/5 to All on Sun Feb 18 18:27:08 2024
    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>

    BSD (now referred to as 1BSD) shipped around thirty copies in the
    first half of 1978. Somewhere around June, Pascal had been further
    improved, the C shell had been written, vi had been written (by Joy),
    and termcap had been written (by Joy). These new tools comprised the
    bulk of the Second Berkely Software Distribution or 2BSD. Joy was the
    man running the show, and he'd answer the phone, create the tapes,
    incorporate feedback, package, and mail the software. Seventy five
    copies of 2BSD were sent out.

    The DEC VAX was first introduced in 1977. It was a 32 bit ISA with
    virtual memory. In the first half of 1978, Professor Richard Fateman
    at Berkeley was looking for a machine with a lot of memory for a
    project of his. The VAX-11/780 met his requirements and his budget,
    and he got some folks together to purchase it with a bit of help from
    the National Science Foundation. The VAX did have one disadvantage in
    the minds of those at UC Berkeley; the VAX ran VMS. Luckily, UNIX had
    already been ported to VAX as UNIX/32V (V7 UNIX variant) by John
    Reiser and Tom London at Bell Labs. This port, however, didn't take
    advantage of the main VAX feature, virtual memory, which limited the
    available memory to 1MB.

    This had to be fixed. For Fateman, virtual memory was a requirement,
    and UNIX was a requirement. He then contacted Professor Domenico
    Ferrari about getting virtual memory support in UNIX. One of the
    graduate students working with Ferrari, Özalp Babaoğlu, then set
    about this task. Along the way, Babaoğlu reached out to Joy for
    assistance. Joy helped integrate Babaoğlu's memory system into 32V
    and helped him debug the resulting UNIX variant.

    The new version of UNIX was good. Joy knew that the 32 bit VAX
    running UNIX would render the 16 bit PDP-11 obsolete, and he started
    porting 2BSD to 32V. Peter Kessler and Marshall Kirk McKusick worked
    on porting Pascal, while Joy handled ex, vi, C shell, and many other
    BSD utilities. The group had completed their work by the end of 1979,
    and Joy shipped 3BSD in December.

    At this point, it's important to delineate the family tree. 1BSD and
    2BSD were improvements to UNIX version 6 while 3BSD was an
    improvement to 32V which was itself of a port and modification of
    UNIX version 7. During the creation of 3BSD, 2BSD had continued to
    see additions, fixes, and releases. One important distinction to be
    made is that the filename for the kernel in 3BSD became vmunix for
    virtual memory UNIX.

    Computer hardware and software were highly varied in 1979 and most
    often nothing was compatible with anything else. For the Defense
    Advanced Research Projects Agency, this was an issue. They felt that
    the best they could hope for was unity at the operating system level
    for the growing network of networks, ARPAnet, and to this end they
    chose to standardize on UNIX due to its proven portability resulting
    from having been written in C. This same system would also be used by
    DARPA for work in the VLSI Project. In the autumn of 1979, Fabry
    reached out to DARPA offering 3BSD as the solution to their problem.
    Initially, this wasn't well received, but the success of 3BSD that
    December changed opinions. Fabry secured an eighteen month contract
    with DARPA that began in April of 1980. This contract stipulated that Berkeley's new Computer Systems Research Group would add the features
    to 3BSD needed by DARPA which was left rather open. In a technical
    report on the matter, the government doesn't actually list anything
    too specific.

    [ARPA Standard UNIX Report]

    <https://www.abortretry.fail/api/v1/file/ 2c12ca84-b921-4d68-97a1-c046f60372c4.pdf>

    Fabry hired Laura Tong to be the project administrator, and Fabry
    then set about finding a tech lead. On an evening in early March, Joy
    rang Fabry at his home and expressed his desire to lead UNIX
    development, and Fabry agreed. Unlike earlier BSD releases (and
    on-going releases of 2BSD) a more robust system of distribution
    needed to be in place to handle a higher number of orders. Tong setup
    just such a system, and had Fabry coordinate with Bob Guffy at AT&T
    as well as the university's lawyers to find license terms that would
    satisfy all parties. The system that Joy and the software team would
    create was 4BSD which was available in October of 1980. This release
    brought job control to the C shell initially developed by Jim Kulp
    and integrated by Joy, delivermail (predecessor of sendmail), job
    control signals that worked more reliably (so if you sent SIGHUP the
    job would actually hangup), the curses library, the control-z
    suspend/resume functionality we know today, a filesystem that
    supported block sizes up to 1K, and greater hardware support.
    Notably, 4BSD supported the VAX-11/750. Like prior BSD releases, 4BSD
    included the Pascal compiler and that saw still more improvements,
    and it included the Franz Lisp system by Richard Fateman. 4BSD saw
    nine months as the system de jour, and about one hundred fifty copies
    were shipped. Licenses were per institution and not per machine, and
    estimates are that the distribution was running on about five hundred computers.

    4BSD drew criticism from David Kashtan at Stanford Research
    Institute. He'd run some benchmarks on VMS and BSD and he claimed
    that VMS was the clear winner. This didn't sit well with Joy. He went
    and made a series of performance improvements to BSD (mostly in
    vmunix), and a few weeks later released a paper rebutting Kashtan's
    claims and showing that BSD was every bit the match to VMS. The
    improved kernel was coupled with Kevin Robert Elz's
    auto-configuration and a system featuring them was released as 4.1BSD
    in June of 1981, and 4.1BSD for VAX became 2.8BSD for the PDP-11..
    This version was current for two years and shipped four hundred
    copies. This version was good enough to win CSRG another two years of
    funding from DARPA, and the CS department at Berkeley also saw
    funding.

    Interestingly, the reason for a point release was political. The CSRG
    had wanted to call this release 5BSD, but AT&T blocked that name.
    AT&T were releasing System V UNIX at the time, and they felt that a
    Berkeley release also named “five” would confuse customers. Point
    releases of 4.?BSD were then the agreed upon solution.

    The new funding round came with some more concrete goals.
    Specifically, DARPA wanted a better filesystem with higher
    throughput, support for multi-gigabyte adress spaces, better IPC, and integrated networking. Marshall Kirk McKusick outlines how these
    decisions were made:

    To assist in defining the new system, Duane Adams, Berkeley's
    contract monitor at DARPA, formed a group known as the "steering
    committee" to help guide the design work and ensure that the
    research community's needs were addressed. This committee met twice
    a year between April 1981 and June 1983. It included Bob Fabry,
    Bill Joy, and Sam Leffler of the University of California at
    Berkeley; Alan Nemeth and Rob Gurwitz of Bolt, Beranek, and Newman;
    Dennis Ritchie of Bell Laboratories; Keith Lantz of Stanford
    University; Rick Rashid of Carnegie-Mellon University; Bert
    Halstead of the Massachusetts Institute of Technology; Dan Lynch of
    The Information Sciences Institute; Duane Adams and Bob Baker of
    DARPA; and Jerry Popek of the University of California at Los
    Angeles. Beginning in 1984, these meetings were supplanted by
    workshops that were expanded to include many more people.

    What was developed largely by McKusick, Joy, Sam Leffler, and Rob
    Gurwitz is astounding. These developments were iterated in releases
    4.1a, 4.1b, and 4.1c but ultimately culminated in 4.2 which was
    released in August of 1983. The two largest innovations were the
    Berkeley Sockets API and the Berkeley Fast File System. The sockets
    interface allowed multiple different network protocols to be used at
    any time and exposed them as files. The Berkeley FFS allowed
    blocksizes from 128 bytes to larger than 4096 bytes if needed. Thus,
    with Berkeley's FFS, a system operator could optimize either for disk
    use or for disk performance. 4.2BSD included a full TCP/IP stack and
    NFS support as well. Starting in the early iterated releases, small
    tools like rcp, rsh, rlogin, and rwho appeared to demonstrate
    capabilities. These were intended to be short lived, but those
    familiar with UNIX will certainly recognize them. 4.2BSD also brought
    disk quota facilities, a better install process, better
    documentation, and some new filesystem related system calls.

    Events less visible to users also occurred between 1982 and 1983. Joy
    left the project in late spring of 1982 for Sun Microsystems, but
    still spent a little time that summer working on IPC and reorganizing
    the UNIX kernel sources to isolate machine dependent code. Once at
    Sun, contributions to BSD continued with Sun sending their
    modifications for running BSD on the Motorola 68000 back to Berkeley.
    Leffler then took over Joy's previous duties. Joy's work, however,
    would make BSD the single most easily ported operating system for
    about the next twenty years. Pauline Schwartz was hired to take over distribution duties in April of 1983. In June of 1983, Fabry went on
    sabbatical and Ferrari and Susan Graham took over running the CSRG.
    In 1984, Leffler went to work for Lucasfilm and Mike Karels took over
    as the UNIX dev lead. He'd previously worked on the 2.?BSD series for
    the PDP-11. Later that year McKusick joined CSRG full-time.

    4.2BSD gained over a thousand site licenses in a year and a half, and
    most UNIX system vendors actually shipped 4.2BSD (or a system based
    on it) rather than AT&T's System V UNIX. Likewise, many varieties of
    corporate UNIX would be based upon 4.1c or 4.2BSD such as SunOS and
    DEC Ultrix. An important point at this time is that BSD's codebase
    was largely divergent from that of AT&T while still mostly
    maintaining compatibility. Additionally, AT&T had started shipping
    UNIX without source, and those customers who bought AT&T's commercial
    UNIX would often then separately obtain BSD sources if they wanted to
    modify some aspect of the system. For those using microcomputers,
    Microsoft's (and SCO's) XENIX was the standard.

    Over on the PDP-11, a mixture of 4.1a and 4.1c formed 2.9BSD, but
    4.2BSD never showed up.

    In 1976, comic artist Phil Foglio drew the first versions of Beastie
    as T-shirt art for Mike O'Brien as payment for unlocking a safe.

    [T-shirt art of BSD by Phil Foglio]

    <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%2F1b00a913-e45d-44e7-b206-763ebac84cd4_688x833.jpeg>

    Much of this is largely a pun on services being called daemons, and
    UNIX making heavy use of pipes, with /dev/null being a bit bucket.
    The daemon, Beastie, gained lasting association with BSD via a
    drawing by John Lasseterf Lucasfilm being used on the cover of the
    Unix System Manger's Manual that was published in 1984 by USENIX for
    4.2BSD.

    Cover of the manual, image from jacobelder.com <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%2F0cfe4090-7d76-4d07-acfd-0eadab7936cd_640x905.png>

    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
    1986.

    Keith Bostic joined CSRG in 1986 with the condition that he be
    allowed to continue and to complete a prior project. He was working
    to port 4.3BSD to the PDP-11. This meant that a 250K minimal system
    on VAX would be made to fit in the 64K address space of the PDP-11...
    no one thought this would work. No one. The man was clearly mad, but
    despite what madness may have lurked within him, Bostic made it work
    using a set of overlays and auxiliary processor states, and this
    formed the 2.10BSD release. He also began attending USENIX and he'd
    announce the progress of the removal of all AT&T code from BSD which
    started in 1986 at thirty five percent AT&T license free, and the
    announcement was met with widespread cheers and applause. This became
    important due to price increases from AT&T. As of the 24th of
    February in 1984, the price for a commercial license of UNIX System V
    Release 2 with source for a single CPU stood at $43000 (about $126000
    in 2024), and each additional CPU was a further $16000 (nearly $47000
    in 2024). For educational institutions, the price was lower at $800
    (or $2300 in 2024) and an additional $400 for each CPU.

    [AT&T UNIX prices 1984]

    <https://www.abortretry.fail/api/v1/file/ caa106e2-deee-4f25-b938-54f15bcd1f64.pdf>

    Getting well into the 1980s, it became rather obvious to CSRG members
    that VAX wouldn't go on forever. Computer Consoles Inc had opened a
    development center in Irvine and there they developed a proprietary minicomputer called the Power 6/32 code named Tahoe. This new
    platform was aimed to compete directly with VAX. In need of an
    operating system, CCI turned to CSRG. CCI provided the team several
    machines and the team set about porting 4.3BSD to the 6/32. This
    became 4.3BSD-Tahoe in June of 1988. Importantly, they were able to
    push forward Joy's work of separating machine dependent code from the
    rest of the system, and 4.3BSD-Tahoe introduced an OSI network
    protocol stack, improved the kernel's virtual memory system, and
    introduced more efficient TCP/IP support. Unfortunately, 4.3BSD-Tahoe
    was short lived as CCI changed their company's focus, and Sperry and
    Burroughs released rebranded minicomputers based on the Power 6/32
    platform. Where this effort did live on, however, was in a merger of 4.3BSD-Tahoe and 2.10BSD into 2.11BSD for the PDP-11. The Tahoe
    effort also meant that much of the BSD codebase had been rewritten
    furthering the aim of removing AT&T licensed code.

    BSD was released in source format only. Any prospective user would be
    required to compile his or her system from source entirely. Given
    this, any user would first need to acquire an AT&T source license
    before he/she would be able to make use of BSD. As noted previously,
    UNIX license fees were ridiculous. Yet, the TCP/IP stack in 4.3BSD
    was unique to Berkeley. Several software developers requested that
    BSD's networking code be offered separately from UNIX, and this was
    done in June of 1989 as Networking Release 1. Pricing for a tape from
    Berkeley was $400 (around $983 in 2024), but the license terms
    allowed for free modification, free redistribution, and free
    application to any use case provided that the copyright notices on
    Berkeley's code remain in place, and that products incorporating the
    code mention in documentation that code from the University of
    California and its contributors was included. It was also available
    via anonymous FTP shortly following the initial release. This
    newsgroup post was made on the 7th of December in 1988, but other
    documentation states public availability was November:

    Path:
    utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!
    ames!pasteur!ucbvax!OKEEFFE.BERKELEY.EDU!bostic
    From: bos...@OKEEFFE.BERKELEY.EDU (Keith Bostic)
    Newsgroups: comp.bugs.4bsd.ucb-fixes
    Subject: V1.73 (BSD Networking Software, Release #1)
    Message-ID: <8812070154.AA18358@okeeffe.Berkeley.EDU>
    Date: 7 Dec 88 01:54:54 GMT
    Sender: dae...@ucbvax.BERKELEY.EDU
    Organization: University of California at Berkeley
    Lines: 374
    Approved: ucb-fi...@okeeffe.berkeley.edu

    We are happy to announce the availability of the first release of
    the BSD networking software. It consists of the standard user level applications, (along with their manual pages and some related
    documentation) and some kernel and C library support. It should be
    noted that this software has only been tested for compilation and
    operation on 4.3BSD and 4.3BSD-tahoe. A complete list of files is
    attached to this message.

    The TCP and IP code is approximately the same as that recently made
    available via the ARPANET and Usenet. Several new algorithms are
    used in TCP, in particular Van Jacobson's slow start and dynamic
    window size selection algorithms and Phil Karn's modification to
    the roundtrip timing algorithm. These changes increase throughput
    and reduce congestion and retransmission. Several fixes were made
    in the handling of IP options and other gateway support.

    This software suite is copyright The Regents of the University of
    California and may be freely redistributed. No previous license,
    either AT&T or Berkeley is required. The release costs $400.00 US.
    To request an order form, please contact our distribution office by
    phone at 415-642-7780, or by email at bsd-d...@ucbarpa.berkeley.edu
    or uunet!ucbarpa!bsd-dist, or by U.S. Mail at:

    CSRG, Computer Science Division
    University of California
    Berkeley, CA 94720

    Mike Karels
    Kirk McKusick

    Rather than including the file list, I will simply note that this
    release was very complete. It included arp, ftp clients and servers
    (yes, plural on both), route, telnet, dns tools, ifconfig, inetd,
    pieces of the BSD libc, sendmail, syslog, ping, and uucp.

    While the networking release was being made, 4.3BSD development
    continued. A new virtual memory system was implemented from MACH at Carnegie-Melon with the porting and merging work done by Mike Hibler.
    The interface for that VM system was, however, a purely Berkeley
    design adhering to the architecture descriptions found in 4.2BSD. The
    NFS system was upgraded to be Sun-compatible via the use of Rick
    Macklem's work at the University of Geulph. This became 4.3BSD-Reno.

    The original BSD license included in parts of Tahoe and Reno, and in
    all of NET/1 and NET/2 read:

    Copyright (c) <year> <copyright holder>. All rights reserved.

    Redistribution and use in source and binary forms are permitted
    provided that the above copyright notice and this paragraph are
    duplicated in all such forms and that any documentation,
    advertising materials, and other materials related to such
    distribution and use acknowledge that the software was developed by
    the <copyright holder>. The name of the <copyright holder> may not
    be used to endorse or promote products derived from this software
    without specific prior written permission. THIS SOFTWARE IS
    PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES,
    INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
    MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.

    Keith Bostic was still as avid as ever to rid BSD of AT&T code and
    free it from the heavy costs that came with that code. At a meeting
    of CSRG, he mentioned the popularity of the networking release and he
    proposed an expanded release that would include more than networking. Discussion went on for a while, but eventually McKusick and Karels
    took up the kernel work with Bostic handling the utilities and C
    library. Obviously, this would be a seriously large undertaking for a
    single individual, and Bostic figured that others would be willing to
    help. Bostic then encouraged people to coordinate over the budding
    Internet with people across the globe submitting contributions. He
    also encouraged people to contribute at USENIX. In a little over a
    year, most of the utilities and libraries had been rewritten with
    major contributions coming from Bill Jolitz, Donn Seeley, Trent Hein,
    Vadim Antonov, Mike Karels, Igor Belchinsky, Pace Willisson, Jeff
    Polk, and Tony Sanders. Karels and McKusick hadn't actually expected
    Bostic to succeed, but with the work completed, Bostic walked into
    their office with his head held high and inquired as to their
    progress on the kernel. The two then went off to work on it file by
    file, removing everything originally included in the 32V release, but
    they were short by six files. Rather than getting a new license and
    name created by the university's lawyers, the group reused what they
    had and this became Network Release 2 announced on the 3rd of July in
    1991.

    Path: gmdzi!unido!fauern!ira.uka.de!sol.ctr.columbia.edu!spool.mu.edu! caen!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!ucbvax! OKEEFFE.BERKELEY.EDU!bostic
    From: bos...@OKEEFFE.BERKELEY.EDU (Keith Bostic)
    Newsgroups: comp.bugs.4bsd.ucb-fixes
    Subject: V1.95 (BSD Networking Software, Release #2)
    Message-ID: <9107032314.AA06592@okeeffe.Berkeley.EDU>
    Date: 3 Jul 91 23:14:59 GMT
    Sender: dae...@ucbvax.BERKELEY.EDU
    Organization: University of California at Berkeley
    Lines: 36
    Approved: ucb-fi...@okeeffe.berkeley.edu

    We are happy to announce the availability of the second release of the
    BSD networking software. The distribution includes approximately 75%
    of
    the utilities distributed as part of 4.3BSD-Reno and the C library
    (along with manual pages and some related documentation), and much
    of the kernel. We wish to *strongly* emphasize, however, that
    significant portions of the kernel are missing and that no binary
    support is supplied for any architecture. Please note also that
    this software has only been tested for compilation and operation on 4.3BSD-Reno.

    This release is intended for system developers and others who wish to
    preview or experiment with the most recent Berkeley system. It may
    also
    be useful as an update to earlier BSD or BSD-derived systems, although substantial work will be required to integrate portions of this
    release
    into older systems. This distribution is *not* intended to be used on production systems, nor is it intended for sites without the expertise
    to find and fix problems that are encountered.

    This software suite is Copyright (C) 1991 The Regents of the
    University of California and may be freely redistributed without
    further charge. No previous license, either from AT&T or Berkeley
    is required. The release costs $850.00 US on 6250 BPI 9-track
    magnetic tape or 8mm Exabyte cassette or $950.00 US on 1600 BPI
    9-track magnetic tape. The distribution is approximately 90Mb in
    size. To request an order form, please contact our distribution
    office by phone at 415-642-7780, or by sending email to bsd-d...@ucbarpa.berkeley.edu or uunet!ucbarpa!bsd-dist, or by U.S.
    Mail at:


    [continued in next message]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Julieta Shem@21:1/5 to Ben Collver on Mon Feb 19 12:25:05 2024
    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>

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Ben Collver@21:1/5 to Julieta Shem on Mon Feb 19 16:11:13 2024
    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

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Lawrence D'Oliveiro@21:1/5 to Ben Collver on Fri Mar 29 03:40:28 2024
    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.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)