Wow, did that bring back memories. In 1980 I got a job at Bell Labs as a systems administrator on a PDP-11/70 running Unix-RT. That was my introduction to Unix and C programming and I've been doing both ever since. Thanks for the trip down memory lane. (PS: Zork -- best game ever!)
I have fond memories of porting an extended version of Colossal Cave from Cambridge University, FORTRAN IV source supplied on magnetic tape, to run on a PDP-11/34 under RSX-11M. Address space challenges solved with paged 'overlays'. After the computer was finally turned off, while it played Take Me Out To The Ballgame through rather expensive 12 bit D/As, I salvaged the front panel with power switch and keypad to remind me of those good times and programming all-nighters.
I worked for DEC in the '70s. They made a lot more money servicing customer's computers than they did selling the computers. Good luck getting the drive working!
Which was possibly why they had no problems with publishing all their hardware schematics, and letting other companies make copies. Nowadays the company lawyers would be on you like a ton of bricks.
You are correct, apologies for getting the history a bit wrong. It can be a bit tricky to get some of the details correct while being filmed. I would still argue that what was running on the PDP-7 was quite far off from what Unix really stood for.
I believe the Unix kernel was ported to PDP-11 assembly language first, only later rewritten in C. In the early days, some userspace tools were written in B.
The PDP-11 was used in the Xerox range of large laser printers starting with the 9700, and continuing through to the 4635. I was fortunate to support this device until 2001 - writing assembler patches to correct problems. I first started writing code for the PDP-11 at a different company in 1973. Lovely machine and instruction set.
Well, if it's a twisty maze of little passages, or a maze of twisting little passages, as long as the passage are all different, I can find my way out. But if they're all alike, fuggedaboutit.
I also have a PDP-11/73 and it runs BSD 2.11, has 4M of RAM and an ethernet card. I remember telling a mate about it years ago, whose house I was renting at the time. He was living in Canada while I was in Australia. He was so chuffed about the machine that he got me to give him an account on it so he could log in remotely and play with it.
ADVENT xyzzy plugh. Thanks for the video. Wonderful reminder of a great time in computing and in life. I worked for DEC in the 80s and have fond memories of working on PDP-11/34/40/45/70 and other models. At times we would need to wire wrap back planes and such (dual DMR11s?). We would boot some of the machines with the front switches. You could get fast after a while and actually memorize the boot sequence. Once running, it was fun seeing the front lights move in one sequence for RSX-11M and another sequence for RSTS.
I worked for Floating Point Systems, which made array processors and used PDP-11s, among others, as front ends. Then I worked for Prime. I used to set the front lights to show the user number so I could tell who was hogging the machine. It was a big deal when the front switches were replaced by a Z80! Of course you still had to type "boot 14114," the boot address.
Zork was written at MIT by the dynamod (dynamic modulation) group on a PDP-10 in the early to mid 70's (I was playing it in '76). The DM group was working on natural language processing, and making a game seemed to be the best way to test if the computer "understood" what was typed.
A megabyte of ram could cost upwards of $40,000. Triple that for today's dollars. You had to be clever to keep your program from paging. You also kept things you used a lot grouped in one page so it would would have a higher chance of not having been paged out the next time you got a time slice. If the OS had to go out to paging, you forfeited the remainder of your time slice.
Paging was a tricky thing. When we first installed MACSYMA on our VAX-11/780, it behaved very badly. Then our sysadmin edited the .EXE file, changing the page fault cluster size from its default of 16 down to 1, and that improved things a lot.
Spent many hours of my computer science degree on a VT100 terminal playing through Zork running on a DECSystem-10 in the early 80's. (DECSYSTEM in Zork refers to DECSYSTEM-10 and DECSYSTEM-20 mainframes). I think I still have the maps I laboriously created.
ODT was originally the "Octal Debugging Technique" (and, of course, came on paper tape). By the time that the Qbus machines arrived it was part of the firmware and was referred to as "Console ODT" - the meaning of the acronym also started to morph into "Online Debugging Technique" in some of the DEC manuals (although it was still essentially the same thing and everything was in octal). Note, however, that the 'T' always stood for "Technique" not "Tool" no matter what Wikipedia says ...
I can trace my "love" of A3 graph paper from my youth .... mapping Zork. Great video, even better peek at a Zork newbie effort. Thank you Computerphile!
I still have my hand drawn map, along with incantations and sequences of directions for “all alike” passages. That would have been drawn up in the 80s!
Brings back many happy memories of when I worked in DEC Ballybrit Galway and assembled and tested 11/23 machines for the European Market back in the early 80's.
Love that amber terminal. Have a Wyse 65 with an amber tube and DEC-style keyboard personally; got it hooked up to my Cisco equipment for programming and monitoring them.
Golly this takes me back, we used a lot of 11/73s back then running RSX11M+ but my experience goes right back to running 11/45s with RK05 disks. I've serviced and repaired many RL01/2 drives and actually put my back out of action lifting them in and out on my own, pity I'm too far away to help you. Thanks for a great video !!
The very first computer game I ever played was Colossal Cave Adventure on a PDP-10 timeshare machine in 1979 at the University of Sussex. Zork was a successor. Also CTRL-J is still the linefeed command in emacs and most Unix terminals.
You do know that the ascii value of CTRL-J is a linefeed aka newline so its not so much an "emacs" "command" as it is the value of newline. ^A-^Z are ASCII 01-1A
The PDP-11 was my computing life from 1978 to 1987 in real-time factory applications. We had an 11/40 & 11/34 when I started, later 11/70 and finishing with 11/23, 11/73. 11/53 & 11/83. It was never an **ix o/s, usually RSX11M or MicroRSX but with occasional forays into RSTS/E, RT11 and even DOS11 which was out of support even then
Ahhh, you need to spend some time with Zork, my good man. Soon you'll have a pad of graph paper and you'll be scribbling little maps... it was a very fun game for it's day
Happy memories of amber monochrome screens. :) And friends pressing the hold screen button so it looked like your COBOL program was still compiling...........
Hey, _Zork_ (or _Dungeon,_ in reference to _D&D_ )! That's one I played when I was younger, though it was the Infocom version that was turned into a trilogy so they could fit each part on a disk for microcomputers. They also created a bytecode for it, called Z-code, in order to make it cross-platform and reduce the size, and a programming language, Zork Implementation Language (ZIL), which looked a bit like Lisp. As I recall, there are a few substantive differences in this version, i.e. beyond the mention of DEC and such.
Hey!! This is so nice to see! I've tried to restore a PDP-11/03 in my highschool with a bunch of friends, but the serial card clock was off, and we couldn't get it to boot from our emulated tape disk with RT-11. PDP's are beautiful machines to learn how computers work. I love this!
The other PDP-11 I mention in the video is an 11/02, very similar to the 11/03... Need to sort it out at some point. Shame about the serial card you had issues with.
Aaron S. Jackson Sorry, my other comments had links and they were blocked. Here we go again... By the way, you comment made my day :) I found info on that third switch! Line-Time Clock One of the less understood features of the PDP-11 seems to be the Line-Time Clock, or LTC. This is a signal that is normally generated by the DEC power supply. The power supply samples the A.C. power line, at a voltage the system can handle, and then converts it to a square wave for use by the processor. In the Americas, the signal's frequency will be 60 Hz, while in Europe and Asia, 50 Hz is more common. The LTC signal is injected onto the BEVNT line on the backplane, where it generates a high priority interrupt in the processor. This interrupt is used for updating the system clock/calendar, among other things. Some operating systems, such as Unix, will not run without it.
@@AaronJackson1 Hi Aaron, can you please expand on the 11/02..? I am also a "5-digit DEC badge # person" (13xx8) but never heard of the 11/02.. I do still have 11/03-23-53-73 bits n pieces in my collection, need to start putting together a couple of working systems. Thanx for the memories and the inspiration.. lm
@@lmantuano6986 hey! There's not much info about it, and it may not be called the 11/02 at all (but I think it is?). It was an OEM CPU board which was sold for industrial control systems etc
I used to play Zork on the Apple II. On that platform the game was divided into three parts. Some years later I obtained “cheatsheets” for the games from CompuServe, and I finally managed to finish the games!
It's kind of a pity that they eventually dropped the classic front panel, with the lights and switches--they were increasingly useless at this point, but the middle models of PDP-11 had a front panel specifically designed for aesthetics, and I've always thought it was quite beautiful in a 1970s-modern way, with switches in pink and purple shaped to match the bezel edges.
I repaired as field engineer PDP or VAX systems from DEC. DEC was the powerhouse to networking and provided working turnkey systems that was a stable platform and affordable compared to other systems.
Wow this is one of my favorite videos from one of my favorite channels. 💜 Assembly language and instruction codes: reminds me of teaching computer architecture and organization.
As soon as he mentioned the RL-02 drive, my ears pricked up! I worked for DEC around 1980, and did the Q-bus LSI-11 systems course, RL-xx drives, and other hardware including the PDP-8s, VT-52, VT-100, LA-36 etc...! Good times. My home system was an 11/23, RL-01s, RT-11 and RSX-11, a couple of VT100 clones... What a nerd! One of the sites I supported was the nuclear medicine and cardiology departments of a large hospital... they used their PDPs to map out the dungeons with a huge blackboard map.
I am so glad they got mattius lind in on this one. He really is the modern authority on DEC systems. I have a Heathkit H11 pdp11/03 computer and he helped me with the disk drive. So cool to actually see him in person and i am surprises he is my age.
This is a Q-Bus machine. The Unibus backplane was often wire wrapped and could be modified in interesting ways. I've worked on and with PDP-11/10 PDP-11/05 and the original PDP-11/20 (the original). Those had bit switches.
I worked for DEC for many years. RSTS/E was my baby in the earlier years - I miss it so much. DND under RSTS/E was a line character map game, great fun.
This machine is not the early days of computing. The PDP-11 was a sophisticated machine and the 11/73 was almost the pinnacle of the 11 line. The CPU was a microprocessor in contrast to earlier 11s with TTL CPUs. When this machine was built the IBM PC was already a product and UNIX had migrated to the VAX and other architectures.
Always good to see some old hardware working again. You know you are getting old when the current generation of geeks haven't played Zork! (or ADVENT / Colossal Cave as it was originally). ;)
On the Commodore PET's it was, if I recall correctly, re-badged as some variant of HHGTTG. I never got past the Babel fish (which was really early in the game). I much preferred playing space invaders, which was all done with the extended character set working as sprites; and which was written in assembler. It cheated by setting the display to a given memory address (the display was just a memory location in ram) and then created the next frame in a different memory location, then poked the video address to that memory location and repeated that all over for each frame. Was surprisingly fast, and smooth, given its processor worked at 1Mhz.
I had forgotten about Bureaucracy! I seem to remember it was infuriatingly tedious. Somewhere I still have the official Infocom collection box set on 5.25in floppy.
I started in 1976. From what I remember there were many games DEC had that those of us that worked there, played quite often. The first of these Zork games was simply called Adventure. The second game that morphed into Zork was simply called Dungeon. I think I still have my map from playing Dungeon. Without a physical map it was impossible to play the game. It took me months to finally finish the game. When they got into the PC line, I believe they reworked Dungeon into Zork. I played Zork a few times but I still thought Dungeon was better. Anyway, brings back a lot of memories. Sorry, managed to see the whole video and noticed that you were attempting to play Dungeon in fact. The basics of the game is to move into rooms and if there is a description in the room, pay attention to that description. Pick up objects and look at them. You can also carry objects with you. SAVE often because if you die, and you will die, you can restart your journey without starting from scratch. There is a doorway in the house. It is under the carpet. Open the hatch and you will then enter the Dungeon from what I remember. I believe there is a sword and a torch that are either in the house or nearby. You will need the torch to navigate the dungeon. Be careful though as it can be blown out. Again, save OFTEN. Remember, the game was created by a bunch of nerds at MIT, so you have to think like them.
Going by the game source code (which can now be compiled and run on ITS with the use of a PDP-10 emulator), it was called both Zork and Dungeon, possibly at various points in time. When Supnik got his hand on it, he chose to call it just Dungeon. The MIT people eventually went with Zork.
I used to play Zork with my dad, we never got very far. I remember the first time the character got eaten by a grue i had dreams about it. I also remember getting ambushed by a thief by a chasm thing.
I played hangman on VM/CMS on S/370. I also wrote a number of EXEC scripts. In fact, the very first program I ever wrote was in EXEC on CMS. This was in the days of Space Invaders on the Atari VCS-2600 and Zaxxon, the arcade game original. And Saturday morning cartoons that weren't homosexual or diverse or trans anything.
Lovely, this one seemed easier to operate than the more standard way of doing things in the 70s. dip switches, punch cards and a line printer was probably the more common way of interacting with the PDP11 back then.
Zork was based on walking around "The Mill" corporate headquarters in Maynard. Most of the places in the game were riffing off of real places. Yes, there was a dam, a forest, a dynamo room. There was Bat hunting in the attic (Grues). Maze of twisty passages all alike were cubicles.
I'm not really familiar with Zork, but this sounds a bit strange. Zork was written by some people at MIT. Sure, Bob Supnik worked in The Mill, but his translation was faithful to the original.
This is very interesting. This internship doesn't come up Tim Anderson's History of Zork. And in the Get Lamp interview, Bob Supnik said they visited him at DEC. It seems likely he would have mentioned if they interned there. But enough speculation. I'll confirm this with the Imps.
I got this from one of the Zork creators: "The original Zork map was drawn largely in one lump at MIT without anyone's thinking about The Mill." and "The maze of twisty passages was a direct reference to the maze in the Crowther / Woods Adventure, not to cubicle farms in Maynard."
I have a 5 digit DEC badge# and I can verify a lot of the stories about the mill. When I went to the school complex in Bedford, Mass I would make as many "runs" to the mill as I could. Very neat place.
We had 3 weeks training in Maynard for the 11/70. Learned about multi keyboard control using RSTS - Basic. Couldn't think of a use for it in our business, but came up with a pac man type game where we could all sit in the class and chase each other around, collecting points in a maze. Long before the real thing came along.
Honestly with the first Zork it took the better part of my entire first sitting to figure out that you can actually get inside the house. Mainly because I was never punished for wandering around by being eaten by a Grue. Personally when I was a kid I just skipped to Zork 2.
Yeah, Zork, loved that game, text adventure games. It was awesome, still got my hand drawn map somewhere. 'Cause that is how we rolled .... "GRAPHICS!" Pull-eaze, plebeians. ;)
The VT420 is younger than the machine I believe - it's close anyway. I had a PDP-11/40 for awhile, but keeping it happy was more than a hobby, so it went to another collector. Today I still have some DEC stuff - A PDP-11/84, a few microvaxen... Emulation can get you the flavor, but the real hardware is still amazing. Someone should make a modern FPGA based replica for the next generation. I don't have a lot of love for the old disk drives - those can go away.
At 11:26 the leaflet says "Programming Technology Group" at MIT. The rest of the world says it was the Dynamic Modelling Group at MIT. The "somewhat paranoid DEC engineer who prefers to remain anonymous" is now generally known to be Bob Supnik.
The group changed name over the years. Early on it was "Dynamic Modeling, Computer Graphics, and Networking". Networking was dropped, becoming "DMCG". Then just "Dynamic Modeling". At some point it was "Programming Technology Division", perhaps reflecting the groundbreaking research done there by Barbara Liskov's CLU group. Supnik was semi anonymous for some time, but more tight lipped about who got him the Zork souce code files. Later he's on record that is was Ted Hess, and Hess also confirmed this publicly.
@@larsbrinkhoff Thanks. One more question I have is, did any or all of the Tim Anderson, Marc Blank, Bruce Daniels, and Dave Lebling help themselves to time and materials from one or more DARPA contracts to develop Zork?
I was under impression PDPs were somewhat tall fridge-type machines, didn't realize later ones were coming in 'desktop' form-factor ... o_O thanks for showing this!!
The pdp Basic language we used limited program size to 32k (16k words I think). It had a great send/receive function that was very fast, so we set up a home grown multi key - random/sequential database program as a separate process and could query quite nicely. This saved about 8k in the application program. The language required a lot of goto's which made me appreciate C.
PDP BASIC-Plus had GOSUB, but GOTO was used for IF-THEN-ELSE type of constructs. Every line has to have a line number, and there are no "blocks" of code (like using brackets). So, typical would be 11000 REM Do some stuff 11100 IF A=2 GOTO 11600 11110 REM A not 2, so do some stuff ... 11590 GOTO 12000 11600 REM A is 2, so do some other stuff ... 12000 REM Rest of program
Are you talking about Systime’s ACCESS indexing system to enabled keyed access to files? So that we could calculate which 512 byte block to read from the file?
at 12:50 or so he doesn't notice he missspelled a command... that always amazed me about this game, seemed to work on the tads and inform engines like this
"And this... I don't really know what it does, but if I don't enable it, I can't boot into Unix." Why do I feel like that is just typical computer science? It's like when read someone else's code which isn't well-commented: "There's five lines of code here that... I don't understand at all... but if I bypass it, the application crashes, so I'll just leave it in..."
Incidentally, it's probably the Line Time Clock, a reliable clock source based on mains power frequency that predates widespread availability of RTCs, and is a prerequisite for booting UNIX and a few other things. My Heathkit-encased 11/03 has an LTC that is hardwired on and has to be shut off by pulling a jumper on the PSU, but I know they added front-panel switches for the LTC in later revisions.
IceMetalPunk From the name "AUX" and the hint at a clock relationship, I would suspect it's not as much precision timing as it's the tick interrupt that is used for process switching as well as basic timing. I could easily imagine a UNIX kernel getting stuck quickly if process/thread switching is not happening.
The manual for the 11/23-plus says about the aux switch: "OFF - In the· normal factory configuration, setting this switch to off turns off the system ac power. ON - In the normal factory configuration, setting this switch to on turns on the system ac power. If the HALT switch is also up, the system automatically boots at this time."
Wear a grounded wrist strap when working on your computer and put it on a grounded mat not that plastic static generating table top. When I worked at DEC you would have been fired for not following correct anti static precautions.
Interesting video, what strikes me is the number of IC's used, the boards are packed to the brim with chips. But I guess this machine was highly integrated compared to its predecessors, here you have the "J-11" chipset that makes it possible
Awesome, I used to go and play Dungeons at a friends place as he had a PDP11... Memories.... Just a thought but considering this is such a treasure perhaps you could take the time to treat it with the anti-static respect it deserves before you destroy it....
I almost bought a pdp-11 in the 80's for running a multiuser bbs, but thought about the noise & electric bill. a pdp111-70 with 2 rl03 disl drives and afew extra disk packs. I do have a old HP D class unix server, 712 workstation & dumb terminals
Around 1985 I played a two player video game on VT52 terminals connected to a PDP11 clone running RSX-11 . Each player could control 10 tanks (rotate them, start, stop, shoot). It was a simple game but super fun. I do not remember all the details. I was thinking to rewrite it but I cannot find the original or at least some screenshots about the original. Just by pure chance I am wondering if someone played that game and eventually has access to archives to find the source for it or some details about how it worked. Thx.
this iteration was pretty small and made later on. The PDP11/34's that I'm familiar with are about the size of a dishwasher maybe a little smaller. That does not include disk drives, mag tape readers, paper tape reader and other supporting equipment. Ours were in a 48VDC plant so they had a rack of equipment just for power since they ran off mains power normally.
During post high-school studies I was a student operator for a PDP-11/34 that ran RSX-11. It occupied two full height 17" rack mount units. It included two removable disk pack units and one system disk pack.
Thank you. Are there any other restrictions to the order that boards can be plugged into the backplane bus? What if you plugged in two processor boards? Another topic that could be a future video: how does UNIX/C make architecture neutral that which differs across hardware/processor architectures? e.g. system calls + parameter passing, memory mapping / pages. Did they define all this when they first implemented UNIX in assembly, or did it evolve (and break backward compatibility) when UNIX was made portable? How / when was a convention established for calling functions written in different langauges in an architectual neutral way?
Doesn't the aux switch enable/disable the line time clock (LTC), which AFAIK is a clock sync generated by either the PSU, or i believe, the mutifunction card. Again, not too sure what it does, but it's probably for timing purposes independent of the CPU clock
Ha it's so funny our railway system is so ancient that it uses a PDP 11 emulators to run the software that runs Control system for our Metropolitan railways
wow! I worked on a PDP 11 in 1981 whilst doing computing science, it was cutting edge then. Its amazing how computing science has progressed!
Wow, did that bring back memories. In 1980 I got a job at Bell Labs as a systems administrator on a PDP-11/70 running Unix-RT. That was my introduction to Unix and C programming and I've been doing both ever since. Thanks for the trip down memory lane.
(PS: Zork -- best game ever!)
Being a unix sysadmin at Bell Labs in the early days must have been amazing!
So nice to see a PDP-11 up close, even if it's a later model. And it runs BSD2.11. Well done.
I have fond memories of porting an extended version of Colossal Cave from Cambridge University, FORTRAN IV source supplied on magnetic tape, to run on a PDP-11/34 under RSX-11M. Address space challenges solved with paged 'overlays'. After the computer was finally turned off, while it played Take Me Out To The Ballgame through rather expensive 12 bit D/As, I salvaged the front panel with power switch and keypad to remind me of those good times and programming all-nighters.
I worked for DEC in the '70s. They made a lot more money servicing customer's computers than they did selling the computers. Good luck getting the drive working!
Which was possibly why they had no problems with publishing all their hardware schematics, and letting other companies make copies.
Nowadays the company lawyers would be on you like a ton of bricks.
Unix was actually created for the PDP-7; it was ported to the PDP-11 after being ported to C from PDP-7 assembly.
You are correct, apologies for getting the history a bit wrong. It can be a bit tricky to get some of the details correct while being filmed. I would still argue that what was running on the PDP-7 was quite far off from what Unix really stood for.
Aaron S. Jackson Why?
I believe the Unix kernel was ported to PDP-11 assembly language first, only later rewritten in C. In the early days, some userspace tools were written in B.
Wasn't it called Unics on the PDP-7? I could be wrong, but I don't think it got the moniker UNIX until it was written in C and ported to the PDP-11.
“Universal Node Information Exchange” - in Fiction.
The PDP-11 was used in the Xerox range of large laser printers starting with the 9700, and continuing through to the 4635. I was fortunate to support this device until 2001 - writing assembler patches to correct problems. I first started writing code for the PDP-11 at a different company in 1973. Lovely machine and instruction set.
Did they shrink the PDP-11 down to a single board microntroller-like device to control the printer?
Brings back memories. That darn maze of twisty little passages !
Well, if it's a twisty maze of little passages, or a maze of twisting little passages, as long as the passage are all different, I can find my way out. But if they're all alike, fuggedaboutit.
I also have a PDP-11/73 and it runs BSD 2.11, has 4M of RAM and an ethernet card. I remember telling a mate about it years ago, whose house I was renting at the time. He was living in Canada while I was in Australia. He was so chuffed about the machine that he got me to give him an account on it so he could log in remotely and play with it.
ADVENT xyzzy plugh. Thanks for the video. Wonderful reminder of a great time in computing and in life. I worked for DEC in the 80s and have fond memories of working on PDP-11/34/40/45/70 and other models. At times we would need to wire wrap back planes and such (dual DMR11s?). We would boot some of the machines with the front switches. You could get fast after a while and actually memorize the boot sequence. Once running, it was fun seeing the front lights move in one sequence for RSX-11M and another sequence for RSTS.
I worked for Data General. :-)
I would say “my condolences”, but DEC is gone, too...
They're ALL gone.
I worked for Floating Point Systems, which made array processors and used PDP-11s, among others, as front ends. Then I worked for Prime. I used to set the front lights to show the user number so I could tell who was hogging the machine. It was a big deal when the front switches were replaced by a Z80! Of course you still had to type "boot 14114," the boot address.
How many elections did you help steal with CES?
Zork was written at MIT by the dynamod (dynamic modulation) group on a PDP-10 in the early to mid 70's (I was playing it in '76). The DM group was working on natural language processing, and making a game seemed to be the best way to test if the computer "understood" what was typed.
A whole megabyte of RAM?! What extravaganze! What on earth would you need that much space for?
With 16 users logged in, that's only 64kb per user, before OS and IO overhead.
The per-process virtual address space was only 16 bits.
A megabyte of ram could cost upwards of $40,000. Triple that for today's dollars. You had to be clever to keep your program from paging. You also kept things you used a lot grouped in one page so it would would have a higher chance of not having been paged out the next time you got a time slice. If the OS had to go out to paging, you forfeited the remainder of your time slice.
Paging was a tricky thing. When we first installed MACSYMA on our VAX-11/780, it behaved very badly. Then our sysadmin edited the .EXE file, changing the page fault cluster size from its default of 16 down to 1, and that improved things a lot.
It was enough to go to the moon, it was enough for us :E)
I used to love working on those amber screen VT terminals. Much easier on the eyes than green screens.
Spent many hours of my computer science degree on a VT100 terminal playing through Zork running on a DECSystem-10 in the early 80's. (DECSYSTEM in Zork refers to DECSYSTEM-10 and DECSYSTEM-20 mainframes). I think I still have the maps I laboriously created.
Zork was amazingly advanced for its time. You could input very complex commands indeed.
ODT was originally the "Octal Debugging Technique" (and, of course, came on paper tape). By the time that the Qbus machines arrived it was part of the firmware and was referred to as "Console ODT" - the meaning of the acronym also started to morph into "Online Debugging Technique" in some of the DEC manuals (although it was still essentially the same thing and everything was in octal).
Note, however, that the 'T' always stood for "Technique" not "Tool" no matter what Wikipedia says ...
in the 70's on the pdp-10 we had liza, the online shrink. when you ran it it said "the doctor is in" , "how do you feel" it was written in LISP
Superb. I would love to get one of the small models of PDP-11 like this one. That SCSI to SD is fantastic.
There's a rather nice miniature reproduction one that is a Pi behind the scenes, colourful switches and all.
Considerably cheaper than a real one.
I can trace my "love" of A3 graph paper from my youth .... mapping Zork. Great video, even better peek at a Zork newbie effort. Thank you Computerphile!
I still have my hand drawn map, along with incantations and sequences of directions for “all alike” passages.
That would have been drawn up in the 80s!
Brings back many happy memories of when I worked in DEC Ballybrit Galway and assembled and tested 11/23 machines for the European Market back in the early 80's.
Love that amber terminal. Have a Wyse 65 with an amber tube and DEC-style keyboard personally; got it hooked up to my Cisco equipment for programming and monitoring them.
Golly this takes me back, we used a lot of 11/73s back then running RSX11M+ but my experience goes right back to running 11/45s with RK05 disks. I've serviced and repaired many RL01/2 drives and actually put my back out of action lifting them in and out on my own, pity I'm too far away to help you. Thanks for a great video !!
The very first computer game I ever played was Colossal Cave Adventure on a PDP-10 timeshare machine in 1979 at the University of Sussex. Zork was a successor. Also CTRL-J is still the linefeed command in emacs and most Unix terminals.
You do know that the ascii value of CTRL-J is a linefeed aka newline so its not so much an "emacs" "command" as it is the value of newline. ^A-^Z are ASCII 01-1A
The PDP-11 was my computing life from 1978 to 1987 in real-time factory applications. We had an 11/40 & 11/34 when I started, later 11/70 and finishing with 11/23, 11/73. 11/53 & 11/83. It was never an **ix o/s, usually RSX11M or MicroRSX but with occasional forays into RSTS/E, RT11 and even DOS11 which was out of support even then
I also have RT-11 on an RX02 floppy for this machine, but I personally think UNIX is much more interesting.
Ahhh, you need to spend some time with Zork, my good man. Soon you'll have a pad of graph paper and you'll be scribbling little maps... it was a very fun game for it's day
Still didn't teach you the difference between its and it's, did it.
It didn't teach you to end a question with a question mark, either. The game was *still* fun, regardless of my grammatical prowess.
Well, we see where the priorities were. Certainly not in, say, not putting an estimated 39 space characters after a full-stop.
Anvilshock, I wonder what error snuck through the proof-read, requiring the edit... pedantry is quite fun.
N, N, W, W, Down, Get Screwdriver, S, S, put screw driver in slot.... LOL
Happy memories of amber monochrome screens. :) And friends pressing the hold screen button so it looked like your COBOL program was still compiling...........
Hey, _Zork_ (or _Dungeon,_ in reference to _D&D_ )! That's one I played when I was younger, though it was the Infocom version that was turned into a trilogy so they could fit each part on a disk for microcomputers. They also created a bytecode for it, called Z-code, in order to make it cross-platform and reduce the size, and a programming language, Zork Implementation Language (ZIL), which looked a bit like Lisp. As I recall, there are a few substantive differences in this version, i.e. beyond the mention of DEC and such.
Hey!! This is so nice to see! I've tried to restore a PDP-11/03 in my highschool with a bunch of friends, but the serial card clock was off, and we couldn't get it to boot from our emulated tape disk with RT-11. PDP's are beautiful machines to learn how computers work. I love this!
The other PDP-11 I mention in the video is an 11/02, very similar to the 11/03... Need to sort it out at some point. Shame about the serial card you had issues with.
Aaron S. Jackson Sorry, my other comments had links and they were blocked. Here we go again...
By the way, you comment made my day :)
I found info on that third switch!
Line-Time Clock
One of the less understood features of the PDP-11 seems to be the Line-Time Clock, or LTC. This is a signal that is normally generated by the DEC power supply. The power supply samples the A.C. power line, at a voltage the system can handle, and then converts it to a square wave for use by the processor. In the Americas, the signal's frequency will be 60 Hz, while in Europe and Asia, 50 Hz is more common.
The LTC signal is injected onto the BEVNT line on the backplane, where it generates a high priority interrupt in the processor. This interrupt is used for updating the system clock/calendar, among other things. Some operating systems, such as Unix, will not run without it.
Augusto Figueroa I thought so, but didn't want to say it in case I was wrong haha :)
@@AaronJackson1 Hi Aaron, can you please expand on the 11/02..? I am also a "5-digit DEC badge # person" (13xx8) but never heard of the 11/02.. I do still have 11/03-23-53-73 bits n pieces in my collection, need to start putting together a couple of working systems. Thanx for the memories and the inspiration.. lm
@@lmantuano6986 hey! There's not much info about it, and it may not be called the 11/02 at all (but I think it is?). It was an OEM CPU board which was sold for industrial control systems etc
I used to play Zork on the Apple II. On that platform the game was divided into three parts. Some years later I obtained “cheatsheets” for the games from CompuServe, and I finally managed to finish the games!
It's kind of a pity that they eventually dropped the classic front panel, with the lights and switches--they were increasingly useless at this point, but the middle models of PDP-11 had a front panel specifically designed for aesthetics, and I've always thought it was quite beautiful in a 1970s-modern way, with switches in pink and purple shaped to match the bezel edges.
PDP-11, what sweet memories!
I repaired as field engineer PDP or VAX systems from DEC. DEC was the powerhouse to networking and provided working turnkey systems that was a stable platform and affordable compared to other systems.
Wow this is one of my favorite videos from one of my favorite channels. 💜 Assembly language and instruction codes: reminds me of teaching computer architecture and organization.
You should team up with CuriousMarc to bring those old disk drives back to life
As soon as he mentioned the RL-02 drive, my ears pricked up!
I worked for DEC around 1980, and did the Q-bus LSI-11 systems course, RL-xx drives, and other hardware including the PDP-8s, VT-52, VT-100, LA-36 etc...!
Good times. My home system was an 11/23, RL-01s, RT-11 and RSX-11, a couple of VT100 clones...
What a nerd!
One of the sites I supported was the nuclear medicine and cardiology departments of a large hospital... they used their PDPs to map out the dungeons with a huge blackboard map.
I am so glad they got mattius lind in on this one. He really is the modern authority on DEC systems. I have a Heathkit H11 pdp11/03 computer and he helped me with the disk drive. So cool to actually see him in person and i am surprises he is my age.
This is a Q-Bus machine. The Unibus backplane was often wire wrapped and could be modified in interesting ways. I've worked on and with PDP-11/10 PDP-11/05 and the original PDP-11/20 (the original). Those had bit switches.
I remember playing Zork on a PDP 11 in the Switching Control Center ( Southern Bell ). We were running Columbus UNIX.
Thanks for the memories. I worked on these in the 80's along with vaxes
I worked for DEC for many years. RSTS/E was my baby in the earlier years - I miss it so much. DND under RSTS/E was a line character map game, great fun.
My first job was with Pro Press.
We published The RSTS Professional (later, the VAX Professional).
I jumped ship from RSTS to VMS in 84.
PIP
This machine is not the early days of computing. The PDP-11 was a sophisticated machine and the 11/73 was almost the pinnacle of the 11 line. The CPU was a microprocessor in contrast to earlier 11s with TTL CPUs. When this machine was built the IBM PC was already a product and UNIX had migrated to the VAX and other architectures.
Always good to see some old hardware working again. You know you are getting old when the current generation of geeks haven't played Zork! (or ADVENT /
Colossal Cave as it was originally). ;)
On the Commodore PET's it was, if I recall correctly, re-badged as some variant of HHGTTG. I never got past the Babel fish (which was really early in the game).
I much preferred playing space invaders, which was all done with the extended character set working as sprites; and which was written in assembler. It cheated by setting the display to a given memory address (the display was just a memory location in ram) and then created the next frame in a different memory location, then poked the video address to that memory location and repeated that all over for each frame. Was surprisingly fast, and smooth, given its processor worked at 1Mhz.
That might be the Infocom game, they did various Zorks, HHGTTG, Lurking Horror etc.
Zadster agreed. The hhgttg game was actually written by Douglas Adams. He later did another one (Bureaucracy?)
I had forgotten about Bureaucracy! I seem to remember it was infuriatingly tedious. Somewhere I still have the official Infocom collection box set on 5.25in floppy.
Zadster yes, it was a bit too much like real life. Apart from being continually licked by a llama.
Ah, the PDP-11. I programmed it in Maynard when there was only an assembler and paper tape reader and punch. Those were the days.
Those seas of decoupling caps are just sooooooo pretty! hehehe
EDIT: Keep us posted about those drives, should make a passionating video :)
I started in 1976. From what I remember there were many games DEC had that those of us that worked there, played quite often. The first of these Zork games was simply called Adventure. The second game that morphed into Zork was simply called Dungeon. I think I still have my map from playing Dungeon. Without a physical map it was impossible to play the game. It took me months to finally finish the game. When they got into the PC line, I believe they reworked Dungeon into Zork. I played Zork a few times but I still thought Dungeon was better.
Anyway, brings back a lot of memories.
Sorry, managed to see the whole video and noticed that you were attempting to play Dungeon in fact. The basics of the game is to move into rooms and if there is a description in the room, pay attention to that description. Pick up objects and look at them. You can also carry objects with you. SAVE often because if you die, and you will die, you can restart your journey without starting from scratch. There is a doorway in the house. It is under the carpet. Open the hatch and you will then enter the Dungeon from what I remember. I believe there is a sword and a torch that are either in the house or nearby. You will need the torch to navigate the dungeon. Be careful though as it can be blown out. Again, save OFTEN. Remember, the game was created by a bunch of nerds at MIT, so you have to think like them.
Going by the game source code (which can now be compiled and run on ITS with the use of a PDP-10 emulator), it was called both Zork and Dungeon, possibly at various points in time. When Supnik got his hand on it, he chose to call it just Dungeon. The MIT people eventually went with Zork.
I used to play Zork with my dad, we never got very far. I remember the first time the character got eaten by a grue i had dreams about it. I also remember getting ambushed by a thief by a chasm thing.
I played hangman on VM/CMS on S/370. I also wrote a number of EXEC scripts. In fact, the very first program I ever wrote was in EXEC on CMS. This was in the days of Space Invaders on the Atari VCS-2600 and Zaxxon, the arcade game original. And Saturday morning cartoons that weren't homosexual or diverse or trans anything.
Amber monochrome CRTs are the coolest , white and green are for noobs . My wyse is green and i wish i could find a NIB amber tube for it
Ever stared at an Amber monitor for a long while? It screws up you color perception.
Nah, monochrome CRT's were for cheapskates, the real winners are color CRT's.
I'm a fan of both green and amber but there really is something cool about amber.
I've not heard about that. I'll have to try it.
Depends on the use
What a gorgeous little computer.
Lovely, this one seemed easier to operate than the more standard way of doing things in the 70s.
dip switches, punch cards and a line printer was probably the more common way of interacting with the PDP11 back then.
Excellent video! I'd love to have a PDP11 myself.
Amazing for being one of the very first video games ever
Great content, please more videos with the great prof. Brailsford :)
the "0" function is very important @least it was at the time .
Zork was based on walking around "The Mill" corporate headquarters in Maynard. Most of the places in the game were riffing off of real places. Yes, there was a dam, a forest, a dynamo room. There was Bat hunting in the attic (Grues). Maze of twisty passages all alike were cubicles.
I'm not really familiar with Zork, but this sounds a bit strange. Zork was written by some people at MIT. Sure, Bob Supnik worked in The Mill, but his translation was faithful to the original.
That's right Lars but those persons interned at DEC. The essential story line matches various oddities in the Maynard Mill complex.
This is very interesting. This internship doesn't come up Tim Anderson's History of Zork. And in the Get Lamp interview, Bob Supnik said they visited him at DEC. It seems likely he would have mentioned if they interned there. But enough speculation. I'll confirm this with the Imps.
I got this from one of the Zork creators: "The original Zork map was drawn largely in one lump at MIT without anyone's thinking about The Mill." and "The maze of twisty passages was a direct reference to the maze in the Crowther / Woods Adventure, not to cubicle farms in Maynard."
I have a 5 digit DEC badge# and I can verify a lot of the stories about the mill. When I went to the school complex in Bedford, Mass I would make as many "runs" to the mill as I could. Very neat place.
We had 3 weeks training in Maynard for the 11/70. Learned about multi keyboard control using RSTS - Basic. Couldn't think of a use for it in our business, but came up with a pac man type game where we could all sit in the class and chase each other around, collecting points in a maze. Long before the real thing came along.
Honestly with the first Zork it took the better part of my entire first sitting to figure out that you can actually get inside the house. Mainly because I was never punished for wandering around by being eaten by a Grue.
Personally when I was a kid I just skipped to Zork 2.
Cutting holes through your ears. Genius.
Judging someone by their appearance. Genius. Also you don't cut anything, it's a slow, painless stretching process.
Yeah, Zork, loved that game, text adventure games. It was awesome, still got my hand drawn map somewhere. 'Cause that is how we rolled .... "GRAPHICS!" Pull-eaze, plebeians. ;)
The VT420 is younger than the machine I believe - it's close anyway. I had a PDP-11/40 for awhile, but keeping it happy was more than a hobby, so it went to another collector. Today I still have some DEC stuff - A PDP-11/84, a few microvaxen... Emulation can get you the flavor, but the real hardware is still amazing. Someone should make a modern FPGA based replica for the next generation. I don't have a lot of love for the old disk drives - those can go away.
Real hardware is amazing ... for those who can afford the space, and the electricity bills.
That's what got me. I had a chance to get PDP-11/44 for free, but I didn't have 3-phase power in my apartment :)
Great episode. Very cool.
At 11:26 the leaflet says "Programming Technology Group" at MIT. The rest of the world says it was the Dynamic Modelling Group at MIT. The "somewhat paranoid DEC engineer who prefers to remain anonymous" is now generally known to be Bob Supnik.
The group changed name over the years. Early on it was "Dynamic Modeling, Computer Graphics, and Networking". Networking was dropped, becoming "DMCG". Then just "Dynamic Modeling". At some point it was "Programming Technology Division", perhaps reflecting the groundbreaking research done there by Barbara Liskov's CLU group.
Supnik was semi anonymous for some time, but more tight lipped about who got him the Zork souce code files. Later he's on record that is was Ted Hess, and Hess also confirmed this publicly.
@@larsbrinkhoff Thanks. One more question I have is, did any or all of the Tim Anderson, Marc Blank, Bruce Daniels, and Dave Lebling help themselves to time and materials from one or more DARPA contracts to develop Zork?
I sure wish I could use department of defense money as venture capital I never have to pay back too.
I remember playing Zork on the old Apple //e my dad had bought in the 90s.
I was under impression PDPs were somewhat tall fridge-type machines, didn't realize later ones were coming in 'desktop' form-factor ... o_O thanks for showing this!!
The PDP-11/40 and PDP-11/45s that I used circa 1975-1982 were about a half a standard rack. All TTL logic, and the schematics came with them.
Great video. Great Machine. Zork looks fun :)
Zork was a commercial version of Dungeon written in Fortran-11 on PDPs
One of my favourite computers the PDP-11/73 - I remember the serpentine bus setup, can't remember if this was QBUS or UNIBUS.
QBUS
Those terminals were so nice to use. I have an amber monitor on my Apple ][, it's really easy on the eyes.
ODT is “Octal Debugging Tool”. Used it heavily in the early 80’s
Currently building a PiDP11 and looking forward to running BSD2.11 on it!
On top of Unix and C, I believe the first Oracle version (2?) was developed and released on a PDP-11
The pdp Basic language we used limited program size to 32k (16k words I think). It had a great send/receive function that was very fast, so we set up a home grown multi key - random/sequential database program as a separate process and could query quite nicely. This saved about 8k in the application program. The language required a lot of goto's which made me appreciate C.
No gosub I guess ?
PDP BASIC-Plus had GOSUB, but GOTO was used for IF-THEN-ELSE type of constructs. Every line has to have a line number, and there are no "blocks" of code (like using brackets). So, typical would be
11000 REM Do some stuff
11100 IF A=2 GOTO 11600
11110 REM A not 2, so do some stuff
...
11590 GOTO 12000
11600 REM A is 2, so do some other stuff
...
12000 REM Rest of program
Are you talking about Systime’s ACCESS indexing system to enabled keyed access to files? So that we could calculate which 512 byte block to read from the file?
he typed north wrong and it still worked lol
Jay Mee "drink watrer"
*too much (just following the example and parsing stuff too) :p
Jan Hoekstra this is wrong
Kurt Angerdinger I have Never forgotten “ope doo”
at 12:50 or so he doesn't notice he missspelled a command... that always amazed me about this game, seemed to work on the tads and inform engines like this
Great video!
"And this... I don't really know what it does, but if I don't enable it, I can't boot into Unix." Why do I feel like that is just typical computer science? It's like when read someone else's code which isn't well-commented: "There's five lines of code here that... I don't understand at all... but if I bypass it, the application crashes, so I'll just leave it in..."
Incidentally, it's probably the Line Time Clock, a reliable clock source based on mains power frequency that predates widespread availability of RTCs, and is a prerequisite for booting UNIX and a few other things. My Heathkit-encased 11/03 has an LTC that is hardwired on and has to be shut off by pulling a jumper on the PSU, but I know they added front-panel switches for the LTC in later revisions.
IceMetalPunk From the name "AUX" and the hint at a clock relationship, I would suspect it's not as much precision timing as it's the tick interrupt that is used for process switching as well as basic timing. I could easily imagine a UNIX kernel getting stuck quickly if process/thread switching is not happening.
The manual for the 11/23-plus says about the aux switch: "OFF - In the· normal factory configuration, setting this switch to off
turns off the system ac power. ON - In the normal factory configuration, setting this switch to on
turns on the system ac power. If the HALT switch is also up, the system automatically boots at this time."
I find aesthetics of this terminal really pleasing, can anyone direct me to a similar theme for a linux terminal
There's a PDP running a laminator at CBA (Brazilian Aluminum Company / Votorantim). I think it's still there - it's a huge laminator.
Wear a grounded wrist strap when working on your computer and put it on a grounded mat not that plastic static generating table top. When I worked at DEC you would have been fired for not following correct anti static precautions.
I’m fairly confident my local library network used to use these for the library database....
Kind of interesting that it was originally called Dungeon and not Zork, never knew that.
IIRC, the original one for the PDP-11 was called 'ADVENT', but IIRC it was initially written
on the PDP-10.
Interesting video, what strikes me is the number of IC's used, the boards are packed to the brim with chips. But I guess this machine was highly integrated compared to its predecessors, here you have the "J-11" chipset that makes it possible
I managed a PDP-11/73 for a Mfg Engineering department in the 80's
Awesome, I used to go and play Dungeons at a friends place as he had a PDP11... Memories.... Just a thought but considering this is such a treasure perhaps you could take the time to treat it with the anti-static respect it deserves before you destroy it....
Zork I hadn't heard of. I did see Adventure in the list. XYZZY.
Very curious about this
I love the amber screen.
I see Zork.
I click.
lol ZORK Welcome to Dungeon. This version created 10-SEP-78 ...now THAT is what I call retrogaming :D :p
UNIX was developed on PDP-7 not PDP-11. All that said, MY first experience in 1979 with UNIX was with Version 6 with PWB on a PDP-11/70.
Proper academic - doesn't bother ironing his shirt even though he's going to star in a video!!
Just what I thought!
Oh no, they didn't iron their shirt. The absolute horror. Hold the front page.
I meant it as a sideways compliment :-)
Me too :D
Nothing wrong about not ironing your shirt :)
Could computerphile make a video about the implications of the EU's new data-privacy policies?
I almost bought a pdp-11 in the 80's for running a multiuser bbs, but thought about the noise & electric bill. a pdp111-70 with 2 rl03 disl drives and afew extra disk packs. I do have a old HP D class unix server, 712 workstation & dumb terminals
The creation date on that Zork is a month and a day after i was born
Meanwhile it was about 16 and a half years before I was born!
Congratulations?
Step one of identity theft - done.
And I was almost 13, and about to get my first computer.
And you've been playing it ever since.
I gave up after draining the dam. Didn't know where to go next and just left it at that
Around 1985 I played a two player video game on VT52 terminals connected to a PDP11 clone running RSX-11 . Each player could control 10 tanks (rotate them, start, stop, shoot). It was a simple game but super fun. I do not remember all the details. I was thinking to rewrite it but I cannot find the original or at least some screenshots about the original. Just by pure chance I am wondering if someone played that game and eventually has access to archives to find the source for it or some details about how it worked. Thx.
I always imagined a PDP-11 was bigger than that... like the size of a dish washer or something. It's micro computer sized.
This was a later machine. Earlier ones could occupy several racks.
this iteration was pretty small and made later on. The PDP11/34's that I'm familiar with are about the size of a dishwasher maybe a little smaller. That does not include disk drives, mag tape readers, paper tape reader and other supporting equipment. Ours were in a 48VDC plant so they had a rack of equipment just for power since they ran off mains power normally.
During post high-school studies I was a student operator for a PDP-11/34 that ran RSX-11. It occupied two full height 17" rack mount units. It included two removable disk pack units and one system disk pack.
Thank you. Are there any other restrictions to the order that boards can be plugged into the backplane bus? What if you plugged in two processor boards?
Another topic that could be a future video: how does UNIX/C make architecture neutral that which differs across hardware/processor architectures? e.g. system calls + parameter passing, memory mapping / pages.
Did they define all this when they first implemented UNIX in assembly, or did it evolve (and break backward compatibility) when UNIX was made portable?
How / when was a convention established for calling functions written in different langauges in an architectual neutral way?
Doesn't the aux switch enable/disable the line time clock (LTC), which AFAIK is a clock sync generated by either the PSU, or i believe, the mutifunction card.
Again, not too sure what it does, but it's probably for timing purposes independent of the CPU clock
Ha it's so funny our railway system is so ancient that it uses a PDP 11 emulators to run the software that runs Control system for our Metropolitan railways
How awesome would out be if you got some of the devs of these text adventures?
Did someone say zork! Nostalgia level over 9000..... or 999.
Vi-Kata XiphiqiX Don't you mean 0777777 ?