Rebooting A Retro PDP-11 Workstation - and Its Classic 'Venix' UNIX (blogspot.com) 35
This week the "Old Vintage Computing Research" blog published a 21,000-word exploration of the DEC PDP-11, the 16-bit minicomputer sold by Digital Equipment Corporation. Slashdot reader AndrewZX calls the blog post "an excellent deep dive" into the machine's history and capabilities "and the classic Venix UNIX that it ran." The blogger still owns a working 1984 DEC Professional 380, "a tank of a machine, a reasonably powerful workstation, and the most practical PDP-adjacent thing you can actually slap on a (large) desk."
But more importantly, "It runs PRO/VENIX, the only official DEC Unix option for the Pros." In that specific market it was almost certainly the earliest such licensed Unix (in 1983) and primarily competed against XENIX, Microsoft's dominant "small Unix," which first emerged for XT-class systems as SCO XENIX in 1984. You'd wonder how rogue processes could be prevented from stomping on each other in such systems when neither the Intel 8086/8088 nor the IBM PC nor the PC/XT had a memory management unit, and the answer was not to try and just hope for the best. It was for this reason that IBM's own Unix variant PC/IX, developed by Interactive Systems Corporation under contract as their intended AT&T killer, was multitasking but single-user since in such an architecture there could be no meaningful security guarantees...
One of Venix's interesting little idiosyncrasies, seen in all three Pro versions, was the SUPER> prompt when you've logged on as root (there is also a MAINT> prompt when you're single-user...
Although Bill Gates had been their biggest nemesis early on, most of the little Unices that flourished in the 1980s and early 90s met their collective demise at the hands of another man: Linus Torvalds. The proliferation of free Unix alternatives like Linux on commodity PC hardware caused the bottom to fall out of the commercial Unix market.
The blogger even found a 1989 log for the computer's one and only guest login session — which seems to consist entirely of someone named tom trying to exit vi.
But the most touching part of the article comes when the author discovers a file named /thankyou that they're certain didn't come with the original Venix. It's an ASCII drawing of a smiling face, under the words "THANK YOU FOR RESCUING ME".
"It's among the last files created on the system before it came into my possession..."
It's all a fun look back to a time when advances in semiconductor density meant microcomputers could do nearly as much as the more expensive minicomputers (while taking up less space) — leaving corporations pondering the new world that was coming: As far back as 1974, an internal skunkworks unit had presented management with two small systems prototypes described as a PDP-8 in a VT50 terminal and a portable PDP-11 chassis.
Engineers were intrigued but sales staff felt these smaller versions would cut into their traditional product lines, and [DEC president Ken] Olsen duly cancelled the project, famously observing no one would want a computer in their home.
But more importantly, "It runs PRO/VENIX, the only official DEC Unix option for the Pros." In that specific market it was almost certainly the earliest such licensed Unix (in 1983) and primarily competed against XENIX, Microsoft's dominant "small Unix," which first emerged for XT-class systems as SCO XENIX in 1984. You'd wonder how rogue processes could be prevented from stomping on each other in such systems when neither the Intel 8086/8088 nor the IBM PC nor the PC/XT had a memory management unit, and the answer was not to try and just hope for the best. It was for this reason that IBM's own Unix variant PC/IX, developed by Interactive Systems Corporation under contract as their intended AT&T killer, was multitasking but single-user since in such an architecture there could be no meaningful security guarantees...
One of Venix's interesting little idiosyncrasies, seen in all three Pro versions, was the SUPER> prompt when you've logged on as root (there is also a MAINT> prompt when you're single-user...
Although Bill Gates had been their biggest nemesis early on, most of the little Unices that flourished in the 1980s and early 90s met their collective demise at the hands of another man: Linus Torvalds. The proliferation of free Unix alternatives like Linux on commodity PC hardware caused the bottom to fall out of the commercial Unix market.
The blogger even found a 1989 log for the computer's one and only guest login session — which seems to consist entirely of someone named tom trying to exit vi.
But the most touching part of the article comes when the author discovers a file named /thankyou that they're certain didn't come with the original Venix. It's an ASCII drawing of a smiling face, under the words "THANK YOU FOR RESCUING ME".
"It's among the last files created on the system before it came into my possession..."
It's all a fun look back to a time when advances in semiconductor density meant microcomputers could do nearly as much as the more expensive minicomputers (while taking up less space) — leaving corporations pondering the new world that was coming: As far back as 1974, an internal skunkworks unit had presented management with two small systems prototypes described as a PDP-8 in a VT50 terminal and a portable PDP-11 chassis.
Engineers were intrigued but sales staff felt these smaller versions would cut into their traditional product lines, and [DEC president Ken] Olsen duly cancelled the project, famously observing no one would want a computer in their home.
SCO Xenix (Score:4, Interesting)
I’m afraid your dates for SCO Xenix are a bit off as I was a certified Admin for SCO Xenix circa 1983 and it he been available in the Uk for at least a year by then having been made available primarily as a consequence of the Intel 286 processor being released (1982) which supported protected mode and therefore could effectively run a kernel with effective segmentation of he kernel from user space
Re: SCO Xenix (Score:4, Informative)
I'm pretty sure that SCO Xenix from 1983 supported the 8088-based IBM PC/XT (there are extant manuals for it), and the first Xenix that used the 80286 protection features (and correspondingly, required the 80286) was released in 1984.
Re: (Score:3, Informative)
In my defense it was a long time ago in a land far, far away and I was going from memory.
You are correct, amusingly my point about 286 mode even features in the article.
https://en.wikipedia.org/wiki/... [wikipedia.org]
Re: (Score:3)
>I'm pretty sure that SCO Xenix from 1983 supported the 8088-based IBM PC/XT
Using the word "support" quite loosely, yes it did.
They dropped one off at our departments, for reasons I never did learn.
We gathered around, and after waiting a few minutes for login to finish and offer a prompt, we all wandered off.
Re: SCO Xenix (Score:2)
Well, it was released before the 80286-based PC/AT. There was no computer to run this Xenix on besides the PC/XT and perhaps a handful of XT clones, some of them maybe marginally faster than the PC/XT.
If it didn't support PC/XT, then it supported nothing (of course, it would run perfectly well under emulation right now, if it survived. Sadly, it didn't). Maybe it didn't survive precisely because nobody used it. The next PC Xenix, released for the PC/AT as IBM PC Xenix 1.0 , did survive, though.
Such memories (Score:5, Interesting)
>"It runs PRO/VENIX, the only official DEC Unix option for the Pros. In that specific market it was almost certainly the earliest such licensed Unix (in 1983) and primarily competed against XENIX, Microsoft's dominant "small Unix," which first emerged for XT-class systems as SCO XENIX in 1984"
That was just a little before my *ix dive. I was still into Microware's OS-9 at home, which was similar in many ways. But soon after, at work in '89, was Altos Xenix running on 986T systems (the machines I first managed), but then replaced with a bigger machine, Altos 2000 and Altos Unix, later added a single Sun machine running Solaris Unix, then no more Altos and moved on to standard X86 server stuff and SCO OpenServer (Unix). Then at home, somewhere in that timeline, I got ahold of a legit Interactive Unix license/media at a swap meet but had to download X11 from a local college campus to a QIC tape because it was not included and just too big for a modem (what a blast, and it worked really well; yes, I had a QIC drive in my home tower machine). Then came Linux at home with a zillion floppies. And, eventually, Linux at work. And it has been so ever since.
Re:Such memories (Score:5, Interesting)
Oooh, OS-9 was interesting. I worked for a company that developed software for Unix System III (then V), and a computer manufacturer gave us an 6809 based OS/9 system to see if we could port our software -- not that we promised to do that, they just shipped us the machine and asked. To tell you the truth I think they were desperate; there was so much going on at the time and 68000 based Unix systems were starting to come onto the market with a flat 32 bit address space.
For what it was, which was a machine based on an 8 bit processor with a 16 bit address bus, it was incredible. I would dearly have loved to have kept that machine as a home computer. It blew away anything Intel based at the time, which was around 1984. We probably could have ported our software, although the OS was different enough it would have been a major headache. The problem was nobody was going to buy a business system built on OS-9 on a 6809 when 68000 baed Unix machines were on the market. I understand OS-9 had somewhat more success in the industrial control market.
Re:Such memories (Score:4, Interesting)
OS-9 was, in many ways, way ahead of its time, especially for the smaller hardware it ran on. Full multitasking, multiuser, real-time, memory-mapping, process-based, re-entrant code, then GUI/windowing. The 6809 was a great processor for the time. It just didn't catch on with a big enough push into business.
Re:Such memories (Score:4)
I cut my teeth on the 6800, and then graduated to the 6809, with its big-boy features. While I understand the level of enthusiasm around the 6502 because of its widespread presence in home computer games, it was a crippled imitation of the 6800 architecture's advanced features. Then, when I needed to start writing for the 8088 et seq processors, I spent countless hours thinking, "but why, why, why would Intel do things that way when the obvious, better solution is different?" I still have my MC6800 processor programming manual, and keep it as a reference for how to clearly organize detailed technical information. That book was full of power. The Motorola technical teams from that era were on fire.
The biggest feature that I miss from OS-9 is being able to type ".../" to the shell prompt to indicate the grandparent directory (the equivalent of "../../"). Somebody had their thinking cap on when they came up with that generalization of "." and ".." to "...".
Xenix (Score:5, Interesting)
started as a modified Unix V6 or V7 on a PDP-11, and from that version Microsoft created various ports to microcomputers, notably some with Motorola 68000, Zilog Z8000, Natsemi NS16032 and even i8086 (certain non-PC compatible machines from Altos and Intel). The first Xenix for PCs was made by SCO for the PC/XT in 1983 (or 1984), but it was already based on Unix System III. My point is, Xenix started years before it got PC support, and it started on... a PDP-11.
Tom here (Score:5, Funny)
When I left the organization, I saved that PDP-11 image in an emulator VM, which I'm running on my laptop now.
I'm still trying to exit that vi session.
Re:Tom here (Score:4, Funny)
I'm still trying to exit that vi session.
I have to admit (somewhat sheepishly)... I've found myself in that situation once or twice.
Re: Tom here (Score:5, Funny)
See that's your problem. If you used emacs, you'd never need to exit your session...
Re: (Score:3)
I was going to jokingly reply "except to check mail, right?"... but apparently that egregious oversight has finally been rectified!
https://www.maketecheasier.com... [maketecheasier.com]
Truly we live in the future.
Re: (Score:2)
Re: (Score:2)
>"I have to admit (somewhat sheepishly)... I've found myself in that situation once or twice."
Who hasn't :wq :Q :q!!!
It is a bitch when you get into some mode you don't know you are in and keep pressing escape and trying colon over and over. :)
Re: (Score:2)
Re: (Score:3)
Thanks for the suggestion, but when I tried it, I got
The quest goes on.
Re: (Score:2)
Re:VENIX? nah!!!! (Score:5, Insightful)
Re: VENIX? nah!!!! (Score:5, Informative)
There was no VMS on the PDP-11. The closest thing to VMS was the RSX-11 OS.
The PDP-11 was unusually rich in OSes. To name a few, RSX-11, RT-11, RSTS/E, and of course, a multitude of Unixes. Unix itself was born on the PDP-11 (if we don't count the earliest primitive PDP-7 version) and up to V7 Unix was PDP-11 only. Also, there were Ultrix-11, System III, System V, System V Release 2, 1BSD, 2BSD, and a line of 2.xBSD Unixes up to 2.11BSD, which is by the way still getting patches as of 2024 (!)
Re: (Score:2)
Re: (Score:2)
All we had on our PDP-11 was RSX-11
There were several operating systems named RSX-11: RSX-11D begat IAS, and RSX-11M begat RSX-11M-plus which was rewritten for the VAX as VMS. I think there was also RSX-11A, B and C.
My memory may be missing a few bits, but... (Score:5, Informative)
...I remember running Unix on a PDP-11 at Princeton U in 1974-78. I'm pretty sure that the original Bell Labs Unix was developed on the PDP-11 by Brian Kernighan, Ken Thompson & Dennis Ritchie. They may have been hanging out at Princeton at the time.
Again, my memory is a few bits short on the details.
Perhaps this wasn't an official "DEC Unix" as that hadn't been invented yet. Also remember running RT-11 and RSTS on the PDP-11.
Those were the days of great mini computers!
Re: (Score:2)
>. I'm pretty sure that the original Bell Labs Unix was developed on
>the PDP-11 by Brian Kernighan, Ken Thompson & Dennis Ritchie.
actually, initially Unix ran on PDP-7.
I'm not clear, however, whether or not it ran on just that single machine, or others, before porting to PDP-11 .
Re: (Score:2)
The PDP-11 was one of the earliest ports.
What I'd imagine the submitter meant was that Venix was the official PDP-11 port from DEC at the time that ordinary PDP-11 users were allowed to buy. AT&T wasn't really selling Unix until the late 1970s, when it began licensing it to third parties who could resell it, and before that it was mostly a research project whose source code was distributed around the industry and to universities but wasn't for sale anywhere.
Much of Our Computers is from DEC's Doing (Score:5, Informative)
Re: (Score:3)
Digital Equipment Corporation bought us to our computing place. Unix was written on DEC computers. PDPs inspired a lot of computers, such as the MC 68000 seen in the first Macs, all the way up to and beyond the SnapDragons seen in PDAs. DEC's VAX brought virtual memory, and also brought us memory-mapped hardware. It is a tragedy that there was no way forward--for a company that brought us so much. A nod for DEC.
I have many fond memories of DEC. I worked for DEC customers from 1963 to 1977 as a System Programmer, then for DEC until 1992 as a Software Engineer. Looking back my favorite projects were the DN60 communications processor and the EDT text editor. DEC took the small computer market away from IBM, but IBM survived to take the personal comptuer market away from DEC.
PDP-11 history trivia (Score:2)
Re: PDP-11 history trivia (Score:3)
The 11/20 was a very slow and limited machine, though. Its whole address space was only 64Kb. For that reason, only the earliest Unixes ran on it.
Unix preferred faster and more advanced PDPs like the 11/45 and 11/70 models. The latter supported 4Mb of memory in 1975. It was VERY expensive back then.
Thank you, /. editorial overlords (Score:4, Insightful)
Re: (Score:2)
Awesome! Older computer are more human. (Score:3)
I have my puny collection of vintage computers, but sadly the oldest is a 486sx25. My only non DOS computer is an IBM Power PC, which I got for free along with one of the best keyboards in all of existence, the IBM Model F 'AT'. It's proudly plugged into my 486dx 100.