If you're reading this and are really interested in early UNIX systems, you have probably read some of the early UNIX books and papers. Things like Don Libes' "Life With UNIX", or Peter Salus' "A Quarter Century of UNIX". These are all great books, and if you haven't read them, many can be had for a song on eBay or as used books on Amazon. Basically these are so cheap that you end up paying only shipping. Great deals.
Some of the earlier UNIX systems from AT&T came with an old text processing/string manipulation/pattern matching language called Snobol. Here is the Wikipedia entry on Snobol.
I remember when almost anything I did with a computer resulted in a sense of wonder and excitement.
I realized yesterday that I rarely feel it these days. While I still learn things in my field every day, I think it's the newness that has worn off. I also think that computing was much more fun and interesting when I was closer to the hardware and to the basics of the operating environment.
I've been collecting old UNIX systems for a while now; mostly AT&T UNIX systems. At various times, I've had Fifth, Sixth, and Seventh Edition UNIX running, then System III, System V Release 2, 3, and 4.
At present, I'm only running three: Seventh Edition, System III, and System V Release 4. These are a good compromise for both examining the evolution of UNIX and for having useful systems at hand.
For anyone interested in the technical history of UNIX, actually looking at the various systems is an incredible experience, made even more interesting by reading the excellent documentation that came with each one, or available on the net.
Each successive version adds a measure of sophistication and polish to the kernel, API, userland functionality, and documentation. Yet each is clearly UNIX, and anyone reasonably familiar with any UNIX variant would probably be able to get around in any of them with a little work.
I will be continuing my exploration in the coming weeks and will write a follow-up to this post.
The command line was great and powerful, but unfulfilling for some tasks. Perhaps the most important of these tasks was editing text files, something UNIX was used for frequently in the early days. One of the original uses of the earliest UNIX systems was to author official AT&T publications, hence the Documenter's Workbench (see the Heirloom Documentation Tools page for more information).
The "standard" editor on UNIX was ed, an ancient, untamable beast; a line editor. For the longest time, ed was the only text editor that was guaranteed to be present on any given UNIX system, and almost certainly the only one available in single user mode, for fixing broken things.
Over the last 15 years, UNIX has quietly evolved into something like UNIX but not UNIX.
In fact, one of the reasons I'm interested in Ancient UNIX is to remind me of and teach me the original UNIX philosophy, which among other things says:
This is the Unix philosophy: Write programs that do one thing and do it well. Write programs to work together. Write programs to handle text streams, because that is a universal interface.
The above is a summary attributed to Doug McIlroy, but there are books devoted to the philosophy.
I don't remember my first experiences with UNIX clearly enough.
It was in 1984 at the Naval Postgraduate School in Monterey, CA, where I was starting the computer science masters program. Initially, we studied basic programming (not Basic programming!), then data structures, then algorithm design, all in pascal on the mainframe. I think I remember keying in pascal source on an IBM 3278 terminal in a large computer center. The operating system was VM/CMS as I recall, and it was my first experience with timeshared systems. I had a lot of fun on that system learning a new text editor (Xedit), two new programming languages (Pascal, and Rexx), and a whole new OS (VM/CMS). This was a wonderful time for me, exploring a new universe of high-powered computing, and so unlike the punch card, Fortran nightmare of my undergraduate days in the late '70s.
The oldest UNIX I used regularly was 4.1BSD on Vaxen and 4.2BSD on Sun workstations in the mid-'80s. I don't remember too much about them as they were my first real experience with UNIX and I didn't have a lot of experience with the operating system back then.
Since then, I've had a lot of experience, mostly with AT&T versions (System V in all its distributions and Solaris), though some BSDish stuff, including SunOS, and various Open Source BSD distributions. And then pretty massive experience with various Linux distributions.
Recently I've become interested in some of the older UNIXs again. I'm running AT&T System V R4 on older hardware for the fun of it, and I have been running V7 and V6 UNIX under Simh emulation for a while.
Tonight I uploaded a V7 distribution for x86 architecture here. It's in the Documents tab of this group. It's distributed in multiple files by Nordier & Associates, an IT consulting company in South Africa. These are clearly some really cool people!
I haven't installed this thing yet, but I plan to just as soon as I can find some hardware to put it on. I'd encourage others too also. All are welcome to comment on their experiences.
