14th December 2015
I wrote this ages ago, but it makes a good blog entry.
Why did I start up the Unix Heritage Society? I think the seed was placed way back in 1982 when I spent two weeks at a summer school for high-school students at the University of Wollongong. I remember being introduced to this amazing system called Unix, learning a language called Pascal, and being able to write(1) to the other kids who were logged in.
This brief introduction to Unix left me longing for more, but I was stuck with my high school's Apple ][ and BASIC until I got to university in 1984 where there was a TOPS-20 system. TOPS-20 was kinda cool, but it still wasn't this magical system to which I'd been exposed. I then found out that, the year before, the university had decommisioned a PDP-11 which was running Unix. So I'd missed out again. I still remember looking at the silent PDP-11 which was now housed behind glass in the "computer museum" and musing over what I had missed.
Eventually the university regained Unix capabilities with a Pyramid 90x running OSx (a dual-universe system: System V and BSD). During this time I shared a group house with a bunch of other geeks, one of whom explained to me the elegance of RISC design, but also had a fondness for the elegant orthogonality of the PDP-11 family. Again, I felt like I'd missed out on something.
//I have an e-mail record of purchasing and reading `Life with Unix' by Libes and Ressler around May 1990. At the time I thought it was "a book well worth owning, lots of little anecdotes about Unix, its history, the people etc."//
I think the next thing that sparked my interest in the history of Unix was my discovery in April 1993 of a PDP-11 simulator written by "der Mouse". An e-mail to der Mouse at the time says:
I just picked up and compiled your PDP/11 simulator, which was on larry.mcrcim.mcgill.edu. It was dated 1990, and I was wondering if you have been working on it since then, and if your latest work was available anywhere? I may be fortunate enough to buy an old PDP/11 from a university here with a non-DEC operating system on it, and I'd like to move the disk images over to run on your simulator before the old DEC stops working.
I remember having a friend at the Australian National University who was doing research on optical illusions, and was using a PDP-11 in a full-height rack with two RL02s to control a video card which was generating the optical illusions. The system was definitely running Unix: it could have been V7M, Ultrix-11 or 2.9BSD; I don't know. The system was going to be decommisioned, and I was interested in acquiring it, or at least siphoning the software off so I could run it on a simulator. My e-mail records for this period also show me asking around for a V7 Unix binary of kermit. Frank da Cruz answered that he didn't have a binary, but John Mackin from the University of Sydney did provide a binary. And in May 1993 I received my first e-mail from Steven Schultz, who was later to provide me with much help. For various reasons (cost of moving the system, no place to put it, no 3-phase power) I didn't ever acquire that PDP-11.
//Aside: By this stage of my computing life I had already worked with Minix and 386BSD, and in 1987 I had even rewritten much of Doug Comer's Xinu OS in 6502 assembly code so that it ran on the Apple ][ platform. So I was already fairly cognisant of the workings of a multi-tasking multi-user OS. //
//During this period, there was no readily-accessible Web or FTP resources in Australia (the WWW didn't yet exist, and Australia didn't get connected to the Internet until around 1990. When I was working on Minix, I kept my own archive of large Minix Usenet postings (i.e. patches, new application source code etc.), and when I switched over to 386BSD and FreeBSD, I did the same. Thus, I had already begun to be an "archiver".//
I need to scour back through my e-mail records, but from 1993 onwards I began the task of finding disk images of Unix systems to run on the PDP-11 simulators which I was collecting (der Mouse's simulator, Eric Edward's simulator, John Wilson's Ersatz-11 simulator). Of course, the task was difficult because of the Unix license conditions.
Some relevant dates from my e-mail archives include:
Steven Schultz had sent me 2.9BSD in April 1994, but only as a tarball and a set of boot records, not as a disk image. I'd told him that the university I worked for (UNSW) had a Unix source license. The e-mails show that I had a hard time bootstrapping the system on my simulators.
In March 1995 I sent an e-mail to Torsten Hippe at Potsdam University: "This may seem a strange request. I am interested in the `good old days' of Unix when it ran on PDP-11's. Our campus owns a source license for 7th Edition Unix, but our backup tapes are incomplete. There is no WECo boot tape. I can prove that I have copies of the 7th Edition Unix source code. Could someone put me out of my misery by giving me a RK05/RL02 disk image with a bootable 7th Edition UNIX". After several attempts at transfer (e-mail, ftp etc.) I finally obtained a bootable RL02 disk image with Seventh Edition on it. This strictly wasn't virgin V7, as that system did not support RL02s. It was probably the V7M system put out by Fred Canter at DEC.
In October 1995 I emailed Steven Schultz saying "My skeletal start [of a Unix archive] is available at http://minnie.cs.adfa.oz.au/PUPS ... I'd like to try and act as the central repository for this project, as I have licenses etc.".
I had taken up a research programmer's position in Canberra in 1992, and as you can see by 1993 I was asking around for old Unixes. It was a startling discovery, then, to find that in the computer room not 10 metres from my office we had a dozen tapes with backups of 6th Edition and 7th Edition systems. None of them were complete, and certainly none of them constituted a bootable system. I also found a copy of our System V source license, so I then knew that I could legally obtain old Unix binaries and source code.
As indicated by the e-mails above, I began to construct a Unix Archive and PDP-11 support group by October 1995. The archive was kept hidden and only the mailing list and documentation was publicly available. I would ask around for new additions to the archive, and occasionally people would donate new things to add to the collection.
In 1997 I began a petition to get the old Unixes released under some form of hobbyist license. This began with a questionnaire. This was followed up by a petition to the Santa Cruz Operation (copyright owners of the Unix source code) in early 1998. I was also in communication with Dion Johnson at SCO during this time, and it was he who provided the internal pressure to push the proposal along.
In March 1998, SCO released a hobbyist source license for Unix Editions 1 to 7, 32V and System III. The cost was US$100, and license holders were only permitted to exchange source code and binaries with other Unix source license holders. The hobbyist license was for personal, non-commercial use.
On the 13th May, 2000, SCO released a free hobbyist source license for Unix Editions 1 to 7, and 32V. The terms were again for personal, non-commercial use. The license was placed on their web site, and is still visible on the Wayback Machine.
In June 2000 the PDP-11 Unix Preservation Society (PUPS) was augmented by a new mailing list for The Unix Heritage Society (TUHS). The aim of TUHS was to cover all old Unixes, whereas PUPS was specifically aimed to deal with Unix on PDP-11 issues. The membership overlap is around 60%.
On the 24th January, 2002, Caldera released Unix Editions 1 to 7, and 32V under a BSD-style license.
The Caldera license meant that I was finally able to make much of the archive of Unix material publicly available. This is now known as the "Unix Archive", and it is now mirrored by several sites world-wide.
Postscript: I wrote the above before the lawsuit between the new SCO Group (TSG) and IBM. That farce is superbly documented on Pamela Jone's blog groklaw.net. There were no winners in the saga, and unfortunately old SCO's name was tarnished irrevocably.
Unix has essentially been superseded by Linux, but the Unix Heritage Society still tries to collect and curate the old systems, stories and artifacts from when Unix was a dominant system in the IT industry. I'm sure I'll write about some of the early systems that we have been able to resurrect. I'm looking forward to Unix's 50th anniversary in 2019.