What if computers went back to the '70s too? A Wizard whispers in your ear: "The password of Sheffield Library Packet Switching Service is ABC1234XYZ." That would be a conversation thirty years ago, on the Multi-User Dungeon, or MUD. The Dungeon was actually a minicomputer at the University of Essex. I won't tell you the name of the wizard, because he's a big noise in …
While Unix were constantly releasing new software/hardware to stay at the bleeding edge of technology, Vax was being stabilised.
Result? We have a VMS box which hasn't been switched off in 16 years.
No downtime, AT ALL.
It runs 24/7/365 and runs perfectly. Can't say that for any of our many Unix boxes. So before you use Vax to suggest old and decrepit technology, try bearing in mind that the old technology worked and still does.
If it ain't broke, don't fix it.
I remember where the local not-so-big iron was in Grimsby--I think it was owned by Imperial Foods, or leased by them or whatever.
The office block, last time I passed, seemed to be occupied by some sort of training company, and some of those seem to be only sustained by government-funded programmes for the unemployed.
"While Unix were constantly releasing new software/hardware to stay at the bleeding edge of technology, Vax was being stabilised."
LOL what?
"Unix" is not a company, but a an attempt at a cheap OS cobbled together at Bell Labs in 1969. It was a Doc Brown-style laboratory, complete with source code and was "ported" to various hardware. Since AT&T failed miserably in its attempts to sit on its "Intellectual Property" and since networking code was included in the BSD implementation, the laboratory never needed to be closed.
"VAX" is the DEC computer architecture, a hardware-software coevolution created by like Apple is famous for. Came complete with a wall of manuals that no-one ever read, and any access to the VAX hardware or OS code was guarded in-depth by operators, department beancounters and DEC T&Cs. Well, these days you can get OpenVMS I hear, as long as it's for "non-commercial" purposes and you register at HP or something. But who wants an OS that is handed out like a restricted firearm.
Indeed! I can remember when 2/6 was called "half a dollar" becaus a dollar was worth about five bob. It was a bit more than that in the seventies.
Oh, and RSX-11M is certainly the best operating system I've ever worked on and that includes VAX/VMS. You could do so much so efficiently (with the lovely Macro-11 assembler language) in such a small amount of memory.
And of course with PDP11s you could actually manipulate storage locations using the key switches on the front panel, so you could fix a bug in situ just by zapping a few instructions. Wonderful!
As AC said, VMS was quite somehting. Many years ago I inherited a VAX11/750 to manage along with one running Ultrix (remember that?) and a heap other Unix boxes. The best advice I ever got about running the VMS box was not to touch it. Utterly solid. Unlike the Unix boxes.
What is such a huge pity is that the son of VMS turned out to be anything but stable. (And I don't mean OpenVMS.) I think one lesson is that some ideas and designs simply don't scale.
The history of, and the long shadow cast over modern computer engineering by, the VAXen really is a story worthy of telling.
"And the reason it was not launched was stupid: it was seen as a rival to the success of the 990 mini."
That's only stupid with hindsight. The traditional product lifecycle was spend some time developing the product and then sell it until it became obsolete. It is obvious *now* that in the IT business, products are obsolete even before the final stages of development and the only way to run the business is for products to be continuously under re-development and you sell them as soon as they are stable enough for the cost of customer support to be less than the profit on the sale. (That's why we have so much crap software.) However, even as late as the 1970s, that might not have been obvious to the senior management of any company.
Laugh at these poor souls if you will, (I did. I particularly enjoyed the suggestion that 95% of the computer's time was spent working out how much to charge for the remaining 5%. Makes modern OS schedulers look *very* efficient by comparison.) but do try to understand that the world really has changed since then. These weren't 21st century IT execs with under powered products. They were 19th century execs exploring a strange new business territory where the normal rules just didn't seem to work any more.
One of my lecturers once was involved in developing a compiler on Perkin Elmer hardware - not sure whether that was genuinely British or OEM'ed murrican hardware - circa 1983 I'd say.
Though I cut my teeth on HP minis, the VAXed were really very very good.
Need patriotic icons for long-gone computer systems.
>"A Wizard whispers in your ear: "The password of Sheffield Library Packet Switching Service is ABC1234XYZ.""
:-) I used to use a wide-open JANET node at UCL, no password needed at all: just dial-in,
PAD> call a040000496000001
and you were there. I forget where I got the phone number from, possibly one of my teachers at school told me, or maybe I just read it on a BBS somewhere. It was a whole load cheaper than playing Compunet MUD, but I liked the bugs on c'net MUD more.
Networks were fun then. You could wander around systems and they all had guest/news/games accounts to log in and give it a try. Once you'd found the numbers of the first few systems, you could browse around them and find directories with lists of other systems to log into. I used to play on CERN's mainframes now and again. It was really a case of EU-vs-US in those days though: as soon as you tried to get across the pond you'd run up against some kind of bitnet or timnet gateway that wouldn't let you through without an account, because over there it was all charged-for.
I still hold to the principle that the network is a public space and you have a right to wander anywhere that doesn't have an explicit "keep out" sign. An entire generation of coders grew up learning by exploring and sharing, and the skill base of the whole industry was the better for it. Now that everything's going proprietary, closed and IP/DRM-happy, how are people supposed to have that kind of 'apprenticeship' any more? And then we wonder why all MS' software is so shit.
(Hacker/pirate icon narrowly defeats smiley-face-nostalgia-for-the-old-days icon in a closely-fought battle!)
Hmmm... I think the first 16 bit microprocessor was the Ferranti F100L, which I reckon must have been in silicon by 1974 'cos that's when I left Ferranti. It was a simple single-accumulator thing, and did arithmetic with a fast *serial* ALU (saved space and transistors).
i "grew up" in my i.t. career working under vms, from 1.5 thru to 7.3
i mostly came thru data centre operations, thru senior operator, shift leader, then operations analyst (untold DCL routines i wrote)..and eventualy (sort of) a "systems analyst" (in the famous words of roy chubby brown "i dunno what it is loov, but i'll have a f*cking good look at it for ya!")..
as stated in the first post, vax/alpha/vms reliability was/IS the best. "5 9's" reliability (99.999%), with the alpha in 92 providing 64 bit computing and clustered hardware providing ultimate fault tolerance via soft switchable fail over..(..and this worked a treat, i saw it a few times).
M$'s "wolfpack" and "MSCS" , where are they now?!
personally, i think M$ had vms murdered, by allowing one of it's stooges (Compaq) to buy it up for $7bn in 97 and quietly suffocate vms, much to my eternal anger..i believe vms may have been sold, or it's still sitting in some dusty corner at HP central (the people who bought Compaq)..and, as seems to be the eternal case, the bigger the company, the worse it gets, so it's as good as dead at HP.
what ever became of the 64bit alpha chip i don't know, but i do remember there was a version of 64 bit windows that would run on an alpha, like 8 years ago..but that was quietly suffocated as well ^^
in this day and age of *requirements* for this type of simplicity of platform (ffs, how many versions of the IX's are there???) and UNTOLD reliabilty and ease of use (as anyone who ever worked on them will tell you, DCL is a *joy* to code, F$GETDVI anyone?!), why the hell wasn't vms the platform of choice for all critical apps?
(i don't know if this is still the case but apparently the NYSE still runs on vms, tho that could have changed it was a while since i heard that one..apparently they refused all overtures to get them onto something rubbish and i don't blame them!)
so, kids..before someone spouts off and tells you how "marvellous" any form of unix or linux is, remember, as far as server OS's go, you can't beat vms!
it's awesome.
cheers,
bill
p.s. stufff and nonsense (but nothing about vms!): http://www.eupeople.net/forum
Minicomputers are truly fascinating, not least because they kicked-off the trends towards smaller, more affordable computers, but also because they set the scene for todays computers.
I'm fortunate, I have a microvax-II and a pdp-11 on which I ran the RT-11 (MSDOS-ish) operating system and Fig-Forth until the 36A PSU blew-up! It's amazing what people tolerated for computers and how much usefulness they squeezed out of such puny hardware. I love the way the processors are on hundreds of seemingly identical DIL chips; connected by thousands of wires strewn across multiple, enormous circuit boards. I love them so much I have a design for an original-speed Nova clone in roughly 10 DIL ICs; complete with a toggle-switch front panel. I'd build a Novella(tm) computer for anyone who was interested ;-) !
But DEC were certainly calling their crazy 12-bit PDP-8 machines computers as early as 1966 as the wikipedia article documents!
-cheers from julz @P
And what about ICL? Everyone in the UK wanted them there except the UK government who seemed not to care...
And yes, the VAX was superb to work on! As a consultant, the environments I worked in varied from legal, engineering, system support, satellite comms - it did it all and it worked. The command set (Digital Command Language, DCL) was consistent in its qualifiers - unlike unix and DOS. This made it easy to use and some of our users developed quite complex programs to aid their work. The only fault was that I could not have one at home!
And bear in mind that the PDP-11 (I worked on that too) celebrated its' 21st birthday since it was still being installed into body scanners in the 90s, it being simple, very fast and utterly reliable.
Mike
Hey! Does this mean I'm now over the hill because I cut my teeth on RSTS on a PDP-11, then learning 6502 Assembler and Basic on an Ohio Scientific C3-A? Then SYS III on a DEC?
Sun Micro's 3/60 workstations while in school?
Or I am too young because I remember where I was when the Morris worm hit?
Yeah, I started young. Had my Cat 300 Acoustic modem then the 1200 "digital" (no hand set)
Remember all those home brew kits? Heathkit's H8, IMSAI, CP/M machines? Oh yeah. the memories....
I worked for Plessey (Poole) on the System X project for BT, and yes it had a lot of computing power as well as the specialised switching hardware. We tried to sell it overseas, but no-one wanted it. Huge amounts of the software dev was on Billing, so that trend wasn't unique to the mainframe bureaux.
Plessey would never have built a commercial computer product - they were into cost-plus contracts and they had absolutely no concept of managing product costs. It wasn't a company, it was an institution. GEC and STC (the other System X development partners) were even worse, we thought we were much more dynamic than them.
I designed a lot of the hardware and some of the software for the System X sub-system that linked the Operator Consoles into the network. It actually used microprocessors, and not just any old micro - it was Intel 8086 based! That turned out to be handy experience to have, and in 1983 I found myself in Birmingham working on PCs for Apricot. Good times, mostly.
A DEC PDP 11/40 with 28KWords of Ram and a Single RK05 Disk Drive.
Run DOS V8 and get the odd F342-Odd Address or other Trap 4.
Guess what I used for my Degree Project?
The PDP 11 range was great but already showing signs of age when the Venerable VAX was Released in 1977. 32Bits - Brilliant.
You could have 20 people logged in and running applications on a 1Mb Machine. Can't do that today can we?
I remember fondly the 'pig'. This was the 240v->110v Inverter we had at Dec Arkwright Rd in Reading. The first VAX we had was a 110v beast.
Guy, I remember Prime computers. Napier College (now University) had one. You'd painstakingly type your program into the terminal, type "compile with pascal" (or "c with pascal" if you were feeling lazy) and then find your job at the end of a queue of 300 users. After a few pints and a couple of rounds of pool at the student union you'd come back to find that you'd missed out a semicolon in line 23 and the whole thing had failed.
The year I left they ditched the Prime and bought PCs with turbopascal.
and i'm proud to say that i have worked a fair bit on a PDP-11 (and later on DG AOS/VS systems) in the days of old, and have been in, and grown up with, the computer industry ever since..
my own first computer was a Sinclair ZX81. oh what fun you can have with simple things.
the DG 'minis' with the dumb terminals were the best fun of all (when you were a bureau operator that is)... proc on client's (remote) console anyone?? lol.
oh how things have changed.
i have to say that although modern computers and internet etc have made things far easier (on the user end at least), computers nowadays are definitely a lot less reliable (and trustworthy).
... so you see, i'm not just a troublesome spotty commentard after all..
lolz.
If you have a fascination for how computing was in the 70s, or grew up with them as I did, you can get that retro experience with SIMH, a simulation platform for a variety of classic computers. I have both a VAX VMS and an HP2000 Access system running on my Mac (both OS are licensed for for free for hobbyist use). More info at http://simh.trailing-edge.com/
Incidentally, Essex ran MUD on a Dec System 10 and not a PDP-11. Oslo also had a copy IIRC. I recall BT PSS demo accounts were regularly traded for communications.
Actually, early DEC computers *were* personal. Right back to the PDP-1, which very much resembles a primitive PC. A single box, with a single CRT screen and a single keyboard for input. One person sits in front of it and controls the whole machine. Programs are typed in, executed, and give results in real time. Quite unlike the mainframes that predominated in the 1960s. The PDP-1 has an excellent claim to be considered the first PC, except perhaps on grounds of price. It simply cost too much for anyone but an organization or a rich individual to acquire - but still vastly less than a mainframe. In fact, DEC always offered personal computers right up till the day it disappeared into Compaq. You could take a PDP8, or a PDP11, and use it with a single terminal. And the VAX and Alpha ranges incorporated workstation versions that were quite like today's PCs - only with much more stable (though far more expensive) software.
Interesting. A story by an author who I have always respected - Guy Kewney - and not a mention of one of today's big computer heavyweights that was to end up buying the remnants of DEC (in the form of Compaq, who took DEC over) - HP.
I worked for HP during the late 70's and thru the 80's. The HP1000 real time computer was a very definite competitor for the DEC PDP's. 16 bit architecture and all that, the F-series 21MX with its floating point processor plus a 256Kb memory store and 20Mb hard disk as big as an average washing machine supporting 5-10 users concurrently was a well configured system in those days. And here I am 20 years later wondering how I can make my 64bit CPU with 4Gb RAM and 2Tb of hard disk go faster. Oh, how the world changes.
Good work Guy, but I'm afraid on this occasion I can only sympathise in so far that you seem to have only covered half the story.
And Paris because....well just because.
But you haven't mentioned the other time sharing phenomenon, that of dialling up the provider in another city from a dumb terminal. I remember doing manufacturing process statistics by dialling a remote computer (on one of those new fangled Subscriber Trunk Dialling phone lines) and spending half an hour or more feeding preprepared paper tape into a dumb terminal, with the answers coming back to be printed on a device similar to a Creed teleprinter (but 7-hole tape not the 5-hole beloved of the telex system)! Despite the trunk call charges, and the bureau charges it was still said to be economic.
My abiding memory of it is that they told us when we started at college it was unhackable and 'military strength'.
To its credit, it did take me a week to break into it... in the process causing a crash of the kernel that apparently corrupted the main disk (that's what you get for only skim reading the manuals I guess).
It was a cool OS. If it worked on intel hardware I'd probably still have it on a VM somewhere. It's priviliege system is something that was only vaguley copied on NT later... and DCL has never been bettered.
I can remember designing for the TMS9900 around 1980 - it was my first job out of university. I had to design both the hardware and the software - in those days, you were expected to be competent in both. The equipment we were designing was an industrial control system. Originally it was powered by a Texas 990 mini-computer, but we built out own CPU boards as soon as we could get hold of the 9900 chips - a big, expensive 64-pin chip in a white ceramic package.
They had a totally eccentric architecture, apparently deliberately designed to make them totally incompatible with everyone else's products. Once you started down the TI path, it was extremely difficult to switch to anything else. The address bus was numbered back-to-front (or was it the data bus - I can't remember now) and the peripheral chips would only work with the TI CPU's. They had the weird feature of keeping all the registers in RAM, apart from a single pointer register, so you could do a "context switch" (jumping to a different program), simply by changing that register. The theory was, apparently, that the speed of RAM was increasing so fast that the external RAM-based registers would soon be faster than internal ones. Later on we went to the cheaper TMS9981 and then the 9995, which was, I think, the end of the line for that range. The 9995 was a good processor, but it's non-standard architecture counted against it and it never really caught on.
Later, we designed a system which had two PDP11's working as intelligent disk drives (coupled to the infamous 300Mb disk packs) and a rack full of AMD bit-slice processors doing all the number-crunching - all controlled by a 6809! It was a data compression system using discrete-cosine transforms and was used by a big publishing house (Time-Life, I think), to send their pages by satellite to all the different printing presses. I think the FBI later bought one for storing fingerprints!
Happy days! Electronics isn't nearly such fun now.
VMS came with (most) of the source code on microfiche, couple that with the Internals guides and you could really learn how the beast worked. And don't forget MicroVMS, it came on around 50 floppies of which at least one was going to be a dud.
Dialup over 300 baud could be exciting. The safest way to do anything was to create a command file and execute it. The alternative was to have noise on the wire complete commands for you with random additions. Sometimes those commands worked leaving you pale and wondering what you had just done as SYSTEM.
ICL was a creation of govt, and through its creation ICT's engineering discipline and innovation was frittered away. But in the 1970s I don't remember a shortage of PDP-11s - I was in effect the buyer of them in a University that I will not name because the central computing service boss thought that users getting their hands on computers would diminish his empire. It did.
...after I started working on Burroughs mainframes a PC appeared on someones desk at work,
To this day I can remember thinking this means trouble..and in the days before IP networks were common you needed a different cable hanging out of the back of the PC to talk to each mainframe...I think they called it progress!
Just rebuild YET ANOTHER LINUX BOX - whilst reading this on the other screen, and it took me back. First job I ever had was with GEC, pursuading the boss that with all the money he spent buying time on the central mainframe, he could have his own PDP-11. Anarchy! The wrath of Weinstock fell heavily in those days and the arguement was long and hard. In the middle of this the DEC Salesman whispered "what about a VAX instead?" and so we got an early VAX 11/780.
What a simply wonderful machine? The power was awesome. A macro assembler that has never been beaten.
Just to let you know that MUD 2, the updated version of MUD - with 30 years of development is available to play for FREE at mudii.co.uk on port 23. This is the UK based version. There is also a version based in Canada at mud2.com. The Essex version of MUD, or British Legends as it is now know, is also available to play for free at british-legends.com.
You haven't lived till you've died in MUD!
----*
I worked with VAX/VMS for about 20 years. A more reliable and secure OS you couldn't
wish for. Unlike what passes for an OS on the laptop I'm typing this out on at the moment.
I got out of the business when Unix and Windows took over. It was just too depressing.
Like going back to the stone age. At least VMS felt like it was actually designed and not just thrown together.
Good grief, I remember that slogan. Whew. Them were the days....
What hurt me was buying a PDP11 (128K - yes, K!! words of ferrite core) and a couple of RL01 10 megabyte (megaword??) drives for 10 quid for the whole shebang, then selling it at a car boot sale for the same amount later. Reason? 'Cos when I span up the drives, the telly went on the blink...Missus NOT happy...
Article never mentioned the LSI-11, which was, like a microVAX, a PDP11 on a chipset.
Sod me, I thought the Intel 4004 was tricky....
I'm really surprised that only one comment has mentioned ICL. The article talks of "...companies like Univac, Burroughs, NCR, Control Data Corporation, General Electric, RCA and Honeywell". And that, presumably, means ICL.
Perhaps I'm the only person reading this who named his son after an operating system. I suppose he could have been called TOS (Tape Operating System) or DOS (Disc Operating System) or Pick or CP/M but wasn't. He's called George after the best operating system of the lot---from ICL.
GEC attempted to make general purpose minicomputers, the 2000 and 4000 series. Default login was OPER/OPER.
Most of them were bought by GEC Marconi companies, who had to jump through numerous hoops to justify using a VAX rather than the much crappier internal product.
The reason there was no UK general purpose computer industry was largely the amount of government pork that was doled out to build largely useless weapon systems on a cost-plus (or quasi-cost-plus basis). This work was easier and more profitable than making things ordinary businesses wanted to buy.
I think this still continues (ID cards, nuclear power stations, Eurofighter, aircraft carriers).
24..23..22..21..20..19..18..17..16..15..14..13..12..11..10..09..
08..07..06..05..04..03..
Tests completed.
>>>boot dua0
(BOOT/R5:0 DUA0
2..
-DUA0
1..0..
%SYSBOOT-I-SYSBOOT Mapping the SYSDUMP.DMP on the System Disk
%SYSBOOT-I-SYSBOOT SYSDUMP.DMP on System Disk successfully mapped
%SYSBOOT-I-SYSBOOT Mapping PAGEFILE.SYS on the System Disk
%SYSBOOT-I-SYSBOOT SAVEDUMP parameter not set to protect the PAGEFILE.SYS
OpenVMS (TM) VAX Version V7.3 Major version id = 1 Minor version id = 0
%WBM-I-WBMINFO Write Bitmap has successfully completed initialization.
$! Copyright 2001 Compaq Computer Corporation.
%STDRV-I-STARTUP, OpenVMS startup begun at 30-NOV-2008 23:03:09.42
....
$ show default
SYS$SYSROOT:[SYSMGR]
= SYS$SYSROOT:[SYSMGR]
= SYS$COMMON:[SYSMGR]
$ show sys
OpenVMS V7.3 on node CI4IC4 30-NOV-2008 23:07:40.21 Uptime 0 00:04:44
Pid Process Name State Pri I/O CPU Page flts Pages
00000201 SWAPPER HIB 16 0 0 00:00:00.34 0 0
00000205 CONFIGURE HIB 8 5 0 00:00:00.50 112 175
00000206 LANACP HIB 12 50 0 00:00:02.62 331 1027
00000208 IPCACP HIB 10 6 0 00:00:00.10 99 177
00000209 ERRFMT HIB 8 22 0 00:00:00.12 130 211
0000020B OPCOM HIB 9 54 0 00:00:00.31 211 261
0000020C AUDIT_SERVER HIB 9 53 0 00:00:00.59 563 769
0000020D JOB_CONTROL HIB 10 27 0 00:00:00.16 191 342
0000020E SECURITY_SERVER HIB 10 38 0 00:00:01.30 1779 1687
0000020F TP_SERVER HIB 10 24 0 00:00:00.72 205 317
00000212 SYSTEM CUR 7 215 0 00:00:01.73 1463 356
$
This is running on a SIMH emulator presently running under NetBSD (-current) on an oldish HP i386 workstation. It is actually quite usable if one wants to remember the days of yore... I still haven't got a license for it, but one can login on the console and at least play with the system.
Ive looked at a lot of sites on the PDP-11 and I have a hard time figuring out what the specs were. The only thing Ive come up with is that it had 64k of RAM memory. What kind of permanent memory did it have? Just tape? Was the memory or OS accessed from a line editor, a teletype or console terminal? I know its hard to calculate processor speed as it had a lot of tiny processors, but what was its approximate speed? Thats really interesting to hear some of you have a working PDP-11 in your possession. And not to start a flame war, but which was better, Unix or VMS?
I'd guess the PDP-11 was the first 'real' computer -- bus based, extensible, accessible and so on. A nice bit of kit, made the PDP-8 look really ancient. I don't see much change since then, we seem to be stuck in an eternal timewarp where today's hot patents look like the stuff the mainframe builders were developing so they could cope with physically large computers for the target clock speed (same thing these days except everything's chip scale).
You should mention ICL and maybe its MICOS(?) processor -- a good seller, a microcoded processor capable of running different instruction sets. ICL was never going to make it into the microprocessor age, though -- can't think down to that level (or see that today's inadequate silicon is going to spaw tomorrow's version which will wipe the floor with you.)
The three day week wasn't about inflation. It was about coal. Prices were rising -- helped along by joining the EU, incidentally -- but wages were being held down by government order ("to stem inflation"). Various groups of workers didn't like it, including the miners. Rather than deal with them the government put the country on a three-day week to conserve fuel. I loved it -- three days is about the right length for a workweek, IMHO.
you know what the biggest thing is that i see everyone posting comments on this thread is??
they all forget about the bean counters! why else would DEC have not gone ahead with their plans contrary to the fact that they would have another machine that would have "suppressed" their sales on their "main machine" at that time???
and let's not overlook IBM's huge failing where their "experts" figured that there would only be several hundred PCs desired when, in fact, what they did was to promote "compatible" machines which the market took to heart and ran with which also resulted in numerous other computer manufacturers loosing their marketshares... and it all comes down to the basics of bean counters and poor mismanagement :(
imagine where we'd be today if the same growth curve had been evident in the sixties... heck, just look at the space industry for a similar and very familiar failure :( :( :( ;(
And they were particularly important in the Finance industry.
There used to be a hardware emulation here, running on some other kind of mainframe, that ran the passbook savings account process
It's not that they weren't interested in General Purpose computing: they were just scared of IBM, which, to be fair, as well as being large, was a very tough company, and fought dirty.
Exactly the same situation here! Loved using the PRIME. While you are waiting for your program to compile used to hit the Imperial or some other MUD for a few hours.
Our computer department bought a number of Atari STs about this time which was great because it was really easy to set up the F keys as macros and thus we created the first MMORPG bots. Just hit F1 50 times and go for a cig to find you had gained a level and a wad of cash. ;-)
@Mike Joseph
Yes, ICL had some great kit. George IV running on a 1900-series was way ahead of its time, which is why so many installations of the replacement 2900s spent their entire life running 1900 emulators!
@B
As to what killed VMS, I think it was the same thing that is causing Solaris so many problems today. It was designed by really clever engineers who knew exactly what an OS needed to do, and they filled it with all sorts of nice features (like $getdvi!). Unfortunately that made it slow. Just as I remember being astonished by how much faster VMS 2.0 on a 780 was compared to compiling Fortran on an ICL 1906S, the first Sun workstations flew compared to VMS 4.7, because SunOS 3.x did so much less than VMS.
Now Solaris has been filled with many of the nice features from VMS (always amuses me when something 'new' is added, and I think "Hmm, VMS had that 20 years ago") and now Linux is eating its lunch because people think "ohh, Debian is soooo much faster". Of course it is, it does sooooo much less, but if it's doing all you ned it to, so what? Robustness and features are important to some parts of the market, which is where VMS had (has) a good hold, but not everyone's willing to pay the price up front, even if they learn to their sorrow later that they should have done :)
No doubt in 20 years we'll be reminiscing about how good Linux was, and why it's so sad that it's being superseded by the latest 128-bit mobile phones running some Chinese quantum-effect fuzzy logic OS or somesuch. And some of us will still have Solaris systems running in the basement, just for fun...
I've worked on a number of VAX machines, and even re-homed a few of them as they became surplus to requirements. While DEC may have intended them to run VMS, I only ever saw them running Unix, because quite frankly anyone who *likes* DCL and macro assembler is sick in the head. The whole architecture was a blind alley though - the most CISC ever, just a couple of years before the RISC revolution. What DEC definitely got right was the build quality. My MicroVAX is built like a tank, and I only upgraded the hard drives to get more capacity.
Oh yeah, and the twat who wrote the first comment may want to read Wikipedia himself, or better still a book like the one written by Maurice Bach, where he'll learn about the history of Unix, which was effectively a company at one point - Unix System Laboratories.
"They bought the biggest iron they could afford, and installed giant mainframes (with roughly the power of a modern iPhone)"
In those days you had to rent someone else's computer, but you could run whatever software you liked on it. Now you have an iPhone in your pocket, but you need Steve Jobs' permission to so much as install an app.
"They bought the biggest iron they could afford, and installed giant mainframes (with roughly the power of a modern iPhone) in big, cyber-scifi offices."
Much less power than an iphone! Anything that can do video is way more powerful than a 70's mainframe. A Vax could just manage audio, but not video.
I don't have figures for an iphone, but a Nokia 9300 is worth 17 Crays.
"The Dungeon was actually a minicomputer at the University of Essex. I won't tell you the name of the wizard, because he's a big noise in the computer business these days"
Lorry is a big name in the computer business? Does that mean he's given up his dreams of Amway greatness?
Hah!
The technical issues, managment issues, and the accidents of history of the PDP-11, VAX-11, and Alpha make for unbeatable reading. Probably the most valuable paper is this:
What Have We Learned from the PDP-11 - What We Have Learned from VAX and Alpha, by Gorden Bell and W. D. Strecker. It makes very interesting reading, and pulls no punches.
http://www.computerhistory.org/collections/accession/102630388
A favorite quote (one I used to show to my computer architecture class to emphasis the critical nature of the issue):
By the mid-1980’s, there was a general consensus in DEC that for a given amount of CPU logic in a given technology, a RISC processor could achieve (at least) twice the performance of a CISC processor.
Great stuff, Guy!
Back in the late 1980's I used Compunet. Another user on that system, John Marchant, who went by the name of Gnome, had uploaded the extremely fascinating story of his life with old computers including the Elliot 803 and Molecular 18 Mini-computer before he ended up using Commodore 8-bit computers like the PET. I kept a copy of his story and in 2004 managed to contact him. He gave permission for me to upload his story to my own web site but then his email address stopped working so I lost contact. If anyone out there knows if he is still around then please contact me.
His story is here:
www.davidviner.com/gnome.php
David
The term "Half a Dollar" for a half-crown is much MUCH older than the seventies and has nothing to do with its worth.
Imagine, if you will, in the days before the Normandy Landings, the masses of American servicemen (widely reported as "overpaid, oversexed and over here") wandering around the pubs clutching strange coins. Oddly enough, a half-crown is much the same size as a half dollar coin. What do you think the Yanks called them? And given the proximity of the servicemen of good old Blighty, how long do you think it was before the term "jumped the populations" as it were?
I have no documented proof of this, but common sense prevails and I can't think of another reason why my Grandfather, who had never been to America in his life but had served in both world wars, never called a half-crown anything but "half a dollar", especially since he never had a good word for anything American.
The reason the Compaq wanted to buy DEC had a lot less to do with DECs technology, and a lot more to do with their small army of Field Service Engineers - back then company's like Compaq (and later HP) figured that the real growth opportunities were in Services, rather than Hardware. They all wanted to be the next IBM.
Now they seem intent of shedding as many people as they can get away with
RSTS/E was the most elegant OS I've ever used. VMS was a close second. Amiga OS always appealed to me as its command-line was basically VMS. The PDP-11/70 that I used in college had a whopping 32K of RAM. The day before a programming assignment was due we would have 30 or more people logged in at once, and it ran like we had the machine to ourselves. The VAX I used a couple years later, with 16MB of RAM, would start coughing with 20 users. The University's workhorse was an IBM 4341. It was a pain in the tuckus to use, and complained loudly when too many people needed it at once. Finally around 1987 they bought a 4381 to run administration stuff so we peons (I mean students) could use the whole 4341 for programming projects.
Sure, my Mac with OS-X is pretty, but I feel I could get a lot more done on that PDP. I'm sure it is just my perception. I have many more things going on on the Mac that I could not even dream of when I was using the DEC, but I don't feel nearly as productive trying to program these days.
---
Oh, it was Windows NT that ran on the Alpha. Of course finding Windows applications that were compiled for it was tricky.
Well, in 1968 I designed the MMU on the 1904A (and fixed a bug in the 1906A hardware along the way) - and then it was 2 years before George 4 was finished... (and if you know my name, would rather you didn't publish it against the nickname that I'm using here).
As for the question about what the spec of the PDP-11s included, there were DECtapes (small, rugged spools of mag tape, even students could use them reliably - the unit was just another panel in the cabinet). There was a 64K fixed disc, then RK large diameter single platter removables (pretty good). Paper tape, of course (fanfold). Teletype, and also the DECwriter (console like a Teletype, but dot matrix).
The universal "Periperal Interface Processor" on early DEC operating systems.
Something like:--
pip \c \r DD1:FROM DD2:TO would copy from one directory to another
Something like:--
pip \d \r DD1:FROM DD2:TO would delete the entire contents of both directories
I actually saw someone get this mixed up and delete 6 months worth of programming!
Lets get some background here :-
Digital Equipment Corporation (DEC) ran several core product lines through the 1960s to the mid-1980s when it re-focused initially on VMS and later Tru64/Digital Unix on VAX and Alpha hardware. Setting aside the early PDP and 'FlipChip' hardware products the DEC products that have had the longest lasting influence would seem to be to be :-
PDP-8 family (the classic early mini-computer of the late 1960s and early 1970s)
PDP-11 (including LSI-11) families which is perhaps THE prototypical mini of the 1970s with applications ranging from real-time process control to significantly large multi-user systems (11/70 at the top-end) which fulfilled that market which today we would recognize as 'midrange'.
PDP-6/PDP-10 : which were DEC's range of 36-bit large-scale multi-user timesharing systems from the mid-1960s through to 1982-3 when the 'Jupiter' follow-on to the KL-10 was finally canceled in favour of the 'Venus' product that was eventually to ship as the VAX-8600 (This was a betrayal by Digital that many members of the 'Large Computer Group' (LCG) community have still to forgive!).
It is a little recalled fact that the DECsystem-10 and DECSYSTEM-20 range was (world-wide) the most widely used timesharing system in the higher education sector in the 1970s and it wasn't until quite late in the 1980s that the number of HE seats was overtaken by 32-bit systems (mainly VAX/VMS). In the UK as well as the University of Essex, DEC-10s or later DEC-20s, were to be found in at least the following institutions: Leeds Uni (KI-10 Serial 695), York Uni, Hatfield Poly, Trent Poly, York Uni, The Open University, RGIT Aberdeen, Glasgow CoT, Dundee Uni, Birmingham Uni and doubtless many others that I can no longer recall. Equally it is probably not realised that there were a significant number deployed in industry too - to my knowledge there were several time sharing bureau and similar (Compushare, ADP, On-Line Systems etc) based on the family and in several places in UK industry - ICI for example ran a large SMP DECsystem-10 and several DECSYSTEM-20s.
The purchase of Digital by Compaq in the late 1990s seems to have brought to a close an era of this strand of innovation - it was always clear to those of us around at the time that the midrange systems was never going to be a major focus for Compaq and it was people, services and knowledge that Compaq bought. Once HP had later swallowed Compaq it was all over - who knows now what the future will hold for HP/EDS and the questionable Itanium hardware base?
(no icon for Ken Olsen and not a mention of 'Snake oil' anywhere!)
Mine's the one with part of a KI-10 console sticking out of the pocket [ 30 / 30 : DEPOSIT THIS ]
deCastro left DEC, the story goes, not because Olsen didn't want a 16-bit minicomputer, but because Olsen didn't want *his* 16-bit mini. deCastro's team's design was passed over for the design that was to become the PDP-11.
The story continues that deCastro took the failed design, allegedly a 16-bit extension of the PDP-8 architecture, and went off to start DG, the design becoming the first NOVA. The story's probably not completely true, but there was a lot of bad feeling between Olsen and deCastro.
Though I worked at DG doing communications hardware, I always preferred the PDP-11 architecture.
//tombstone for dead minicomputer market
It's still alive and well and supported by HP. We're up to version 8.3 now. We have a cluster of four 64 bit Alpha servers with four years uptime (I believe the uptime on the one we have in Australia is closer to six years). Solid as rock and it still has a surprisingly large installed base I think you'll find.
"... the most CISC ever..."
Actually the VAXen were microprogrammed beasts and at the core you had a RISC processor. You could, at one time, buy some software from DEC to roll your own microcode. So instead of having all these macros in RISC assembler to perform common operations they just appeared as part of the micro-coded instruction set.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
