PDP-11 Still Working In Nuclear Plants - For 37 More Years

Taco Cowboy writes "Most of the younger /. readers never heard of the PDP-11, while we geezers have to retrieve bits and pieces of our affairs with PDP-11 from the vast warehouse inside our memory lanes." From the article: "HP might have nuked OpenVMS, but its parent, PDP-11, is still spry and powering GE nuclear power-plant robots and will do for another 37 years. That's right: PDP-11 assembler programmers are hard to find, but the nuclear industry is planning on keeping them until 2050 — long enough for a couple of generations of programmers to come and go." Not sure about the OpenVMS vs PDP comparison, but it's still amusing that a PDP might outlast all of the VAX machines.

Comments

I cut my teeth on that CPU

[Bucc5062]

The PDP-11/45 was the very first computer I ever worked with at College in 1978. God I hate to sound like an old guy with a lawn, but they just don't make like that any more. I learned RATFOR, Pascal, c, and Assembler during that time. Even later on, thanks to my time on the PDP11 I expanded system knowledge working with the HP1000 and its front panel switches.

Good times....good times.

Re:I cut my teeth on that CPU

[hazeii]

I still have a bunch of them; every so often I fire them up and program them in ODT. After programming them in assembler (or raw octal), every other instruction set seems irregular. Putting MOV -(PC), -(PC) at the top of memory and executing it was always fun....

Re:I cut my teeth on that CPU

[Anonymous Coward]

What were you doing chewing on the CPU?

You're supposed to lick them.

Re:I cut my teeth on that CPU

[Giant Electronic Bra]

Hahaha! Yeah, my father designed the original flight information display systems which went into all the big airports in the late 60's/early 70's. I'd go into work with him on Saturday and play around on a TTY with BASIC for the PDP-11/C they had in the office there. In the early 90's we were installing micro PDP-11s at VY to do monitoring of discharge temperatures, that was their state-of-the-art machine at that time! Honestly though, how much CPU do you need to read a DAC and push the data up a current loop? A whole PDP-11/45 must cost $.02 and be the size of a grain of rice today. Why invoke the massive overkill of migrating to a PC?

Re:I cut my teeth on that CPU

[poena.dare]

16 bit... tchka!... you had it easy.

The PDP 8/e was da bomb for:

printing "Hello World"
playing Tic-Tac-Toe
paper tape races

Re:I cut my teeth on that CPU

[Cassini2]

The microcontrollers are not rad-hardened. The PDP with core memory and 54-series TTL logic will probably survive a small nuclear blast. There are no highly vulnerable EMI susceptable components in a PDP that I can think of. In fact, I think the military has used (does use?) this and the earlier DTL technologies in its missile computers.

Re:I cut my teeth on that CPU

[rvw]

I still have a bunch of them; every so often I fire them up and program them in ODT. After programming them in assembler (or raw octal), every other instruction set seems irregular. Putting MOV -(PC), -(PC) at the top of memory and executing it was always fun....

Please explain why it was fun! I know the PDP-11 was still around in 1986, and I have probably worked with it for an assembler assignment once. I do remember being in the terminal room, but that's about it.

Re:I cut my teeth on that CPU

[kevinT]

Sad but true.

I went to work for a company in 1995 that was still using PDP11 as their primary language on old DEC machines. This was a commercial application suite. They got their spare parts at garage sales in the area (yes that is what they did, literally!). My last project was to read a bunch of old 9 track tapes to try to find the source code for a program that needed to be modified.

I put up with that for 5 months and bailed to a much better, higher paying job using C on AIX. I was never so glad to leave a company!

Re:I cut my teeth on that CPU

[Cryacin]

At least they're not running nuclear power plants in action script. (ducks)

Re:I cut my teeth on that CPU

[Cryacin]

What were you doing chewing on the CPU? You're supposed to lick them.

Some like to be chewed, some like to be licked. It depends on the model.

Re:I cut my teeth on that CPU

[kilodelta]

My first brush with PDP-11's was a PDP-11/44. I learned BASIC, PL/I and COBOL on that sucker. Then I got hooked up with a retro computing group and what do they have but a plethora of PDP hardware. It's all based on things called Flip Chips. Tiny boards with discrete components on them.

So diagnosis and repair is pretty easy so long as the parts exist.

Re:I cut my teeth on that CPU

[Giant Electronic Bra]

Yeah, I wouldn't be surprised. People nowadays would be surprised at what you can do with low clock speeds, a simple instruction set, and TTL logic. For basic control functions there's no need for anything too fancy and every extra transister is just another thing that can go wrong. Most avionics systems are still using 1970's era 16 bit processors. They've gotten a lot cheaper, but mostly there's just no need for anything fancier when the job is "monitor these 8 DACs and these 5 discrete inputs in a tight loop, apply this filter, write the results to this UART, and close this valve if the state machine reaches this point", lol.

Re:I cut my teeth on that CPU

[Sique]

That might get flashy.

Re:I cut my teeth on that CPU

[BrokenHalo]

Yeah. The only thing the old core-memory processors were not so good at was dealing with excessive heat. My first computer was a Burroughs B3700 a lot like this but with a teletype master console (which Burroughs called a SPO, for Supervisor Printer Operator).

If the airconditioning broke down in the machine room, we had about 15 minutes to shut everything down before the temperature hit 50 degrees C. [OT: Why, oh why, can /. *STILL* not manage such simple things as html entities?]

The company I worked for got rid of that machine in 1978 (in favour of a Honeywell DPS7), but I remember reading in some computing magazine in 1988 that NASA (IIRC) had ordered several of these machines. I can't find any reference to it now, so that might have been shitcanned. It wasn't very long after that, in any case, that Burroughs merged with Sperry (another of my earlier platforms) to form Unisys.

Re:I cut my teeth on that CPU

[Anonymous Coward]

Many of us still can program in macro-11. I got tired of ODT though and built a symbolic debugger with DDT that ran in another task.

The pdp11 has a number of possible hacks that can be useful in remotely identifying code. Ever try a DIV on SP, which moves the stack and changes
the program counter all at a go? Mov @offset(r5),pc ? That gives control transfer an extra level of indirection. There are others.
Note the DIV hack works only for CPU models with the EIS instructions.

The pdp11/45 has a 300 ns. cycle time though, corresponding to ~3.3MHz clock speed. Current machines have more like ~3.3GHz.
The larger 11s could address 4MB of memory (a lot in those days). Now a 4GB machine is beginning to look small.
I recall when we got a 22MB hard drive (size and shape of a clothes washing machine; we put Dymo labels under the lights
labelled "wash rinse dry" for fun) it seemed vast. Nowadays it is not unheard of for a home machine to have 22TB.

With all those factors modern machines can be said to have grown by a factor of maybe a trillion.

Still the pdp11 was versatile and allows significant and useful code to run. It is usually programmed in assembler (macro11;
nobody uses PAL11R any more I hope) or Fortran or C. (The original pdp11 Fortran was a nightmare of code inefficiency, could
burn 100 instructions to add 2 integers where the hardware could do it in one. Only the later f4p compiler got half decent code
generation.)

I only ever saw mov -(pc),-(pc) used as a fast clear memory; it traps at the end, but won't get any unmapped addresses.

It was more fun to arrange to handle interrupts in supervisor mode. (Did that too. Needs a lil trick to RTI.)

Re:I cut my teeth on that CPU

[Joe_Dragon]

we have homer Simpson for that.

Re:I cut my teeth on that CPU

[Anonymous Coward]

Be very careful licking a modern CPU with a fan on it. Two choices, leave the fan on and fragment your tongue or remove the fan and fry your tongue and perhaps the CPU (turning off a CPU before pleasuring it is, of course, counterproductive).

Re:I cut my teeth on that CPU

[swalve]

Yeah, but ironically fancier is cheaper. And now we have printers that take two minutes to boot up.

Re:I cut my teeth on that CPU

[Anonymous Coward]

PDP-11 is too mainstream, we use Sindac in our plants.

Re:I cut my teeth on that CPU

[Giant Electronic Bra]

Yeah, I know what you mean. A lot of newer technology is just less good. Verizon 'improved' my phone service so now they 'back up' my contact list. Its great, every month or so they restore some old worthless crap over the top of my contact list on my phone. Ain't progress great? Now I carry around an actual physical address book.

Re:I cut my teeth on that CPU

[Anonymous Coward]

This was not my first "real" computer. That would be a Burroughs punch card fed machine. I also used Honeywell on punch cards and teletypewriters. However, the PDP-11 was the first that I used predominately through a video terminal. At the time, the Vax was the new kid on the block, and generally reserved for those more deserving than I. I eventually used both at the same time. The Vax was more usable in many ways, but the PDP-11 was more predictable. Features vs. stability.

Re:I cut my teeth on that CPU

[NeoMorphy]

Edusystem 50 on a pdp-8 via teletype was my first. I remember thinking that when I become successful, I'll get a Decwriter II, those things are awesome!

Re:I cut my teeth on that CPU

[Sponge Bath]

That sounds like hell for a job and heaven for a nostalgia project. You should have dressed up 70s corporate style [1] and had someone take a black and white photo of you at the machine.

[1] Either white dress shirt, short hair and pocket protector OR long sideburns and disco suit, depending on the corporate culture.

Re:I cut my teeth on that CPU

[Anonymous Coward]

Don't forget that the PDP-11 also could communicate to the Analog Computer in the computer lab. Anybody remember setting up differential equations on the Analog computer and letting the PDP-11 read the results so that your program could play with the data? Hmmm, my favorite editors where SOS and EDT.

Re: I cut my teeth on that CPU

[Anonymous Coward]

anything with a feature size at or under 0.25 um is more or less rad hard. that's what we use at the LHC. ESD is another story though.

Re:I cut my teeth on that CPU

[rochrist]

I sill fondly remember the 5MB disk drives with the removable platters the size of manhole covers. Those were the days.

Re:I cut my teeth on that CPU

[ak3ldama]

In fact, I think the military has used (does use?) this and the earlier DTL technologies in its missile computers.

Security Breech!!!

Re:I cut my teeth on that CPU

[ackthpt]

The PDP-11/45 was the very first computer I ever worked with at College in 1978. God I hate to sound like an old guy with a lawn, but they just don't make like that any more. I learned RATFOR, Pascal, c, and Assembler during that time. Even later on, thanks to my time on the PDP11 I expanded system knowledge working with the HP1000 and its front panel switches.

Good times....good times.

May have cut my teeth on an OSI (that's Ohio Scientific) home computer, with a whopping 4K of memory, later expanded to 12K, but honed my skills on a PDP 11/50. When I was hired as a student programmer I was given an admin account, subscription to a couple DEC magazines (which had some great kernal hacks in them, plus a full color map of Zork GUE, which I still have :o) and the privilege of performing the offline backups of the two RP04 drives each Friday. I learned how to dissemble the kernal, where all the fun bits were, how to peek at various terminals (to see what the very few aspiring hackers were up to), wrote honey pots and generally learned everything there was to learn in Fortran IV, RSTS BASIC, UCSD Pascal and assembler (while still an unpaid student I brought the CPU to its knees with an assembler program to calculate Pi to 1,000 places, one of those Oliver Wendell Jones sort of moves.)

The one big shock, going from completely nailed down security on a PDP 11 to Windows was how utterly lax and clueless the Microsoft software architects were on dealing with any kind of security - literally babes in the woods, totally unaware of decades of good security practices of mainframe environments. People who think Bill Gates was brilliant need to look more closely at how oblivious he was to threats before unleashing Windows 95 on the world.

I have a PDP emulator for the PC, but without all my old code, I don't know what I'd really do with it. I once had sources to Mark Turmell's early games written on a PDP, such as Squash, all collision detection peformed with an 80x24 integer array and VT52 cursor controls (I bet nobody heard about those early beginnings before.) Before Atari, C64, Nintendo, etc, we played video games on VT52 and VT100 terminals. :o)

Re:I cut my teeth on that CPU

[knight24k]

My High School was lucky enough to have one gifted to it. Apparently, Hugh Hefner (Yes, *that* Hugh Hefner) was an alumnus of my High School and had his daughter come to the school to present the check. I remember learning Assembly, Fortran, Watfiv, PL\1 and I think Cobol as well as rudimentary LED and circuit logic either on it or the IBM 360 downtown. Yes, once upon a time the Chicago public school system had a very good education system. I even remember breaking into the 360 and messing with other school's (and students) teachers assignments (teacher's password: pencil....really?! and taped to the desk next to their terminal. /facepalm).

Re:I cut my teeth on that CPU

[Anonymous Coward]

The PC increments automatically (ok, actually by 2 because we're dealing with words).

So 'mov -(PC), -(PC) ' starts by decrementing the PC (cancelling the auto-increment); now the PC points to the instruction it just executed (this one). The destination is another pre-decrement , so the destination is the memory address before the instruction.

The net effect is the instruction copies itself down one location and executes the newly copied instruction. From there, it's rinse-and-repeat until all memory is full of this instruction (shortly followed by a crash, the exact nature of which depends on interrupts, traps and the state of the stack).

Essentially it's a one-line, one-opcode, self-replicating program.

Re:I cut my teeth on that CPU

[Anonymous Coward]

Never used PDPs, used Data General Nova & assembly, wore t-shirts, jeans, sneakers, hair & beard was long. If it was snowing hard, I was out skiing not coding. Wonderful. Went to a conference & only cobol slingers were "dressed up"; when I talked about A-to-D conversion, real-time, pattern recognition, process control, digital integration & differentiation, their eyes glazed over. Cobol'ers whole world was wrapped up in huge piles of 11x17 green & white pin-feed looking for that one place to slide in a line of code to make some accountant happy - not my idea of a good time.

Oh, I lied - I did use a PDP-8 or -9, not sure which, to connect over DARPA-NET to maybe MIT to run eliza.

Re:I cut my teeth on that CPU

[Medievalist]

Ah, yes, me too - the RL01 and RL02 drives with the aerodynamic heads. People used to make clocks out of the platters if they had a head crash.

Re:I cut my teeth on that CPU

[nerdbert]

Ah, rad hardened PDPs. Those were the days.

I used to program one and we had one in an accelerator when the PDP11 was state of the art. Every time you injected fresh particles into the beam we'd have to leave the accelerator and hide behind a hill due to the radiation (this particular accelerator was designed to put out a ton of polarized x-rays). We could hide behind the hill, but the PDP11 couldn't. The PDP lasted about 3 years before the CPU died from radiation poisoning. I tried to replace the CPU but DEC wanted more money for the CPU than for an entire replacement motherboard. I tried to explain the AE that I didn't feel comfortable subjecting someone else to a board that didn't have much life left, but they made me return my old board for a new one. I wonder what sucker ever got that nearly-dead motherboard?

You can get rad-hard controllers these days. The company I worked for a few years back had CERN come in and make a ton of parts in our process. We couldn't figure out why they kept coming to us for parts as we weren't anywhere near the lowest cost provider for such a limited run of parts (our NRE was big to keep the low volume guys out), but it turns out they'd done rad-hard tests on a bunch of different CMOS processes and ours was an order of magnitude better than anyone else's. I can assure you we weren't designing for a rad-hard process, it just turned out that way.

Re:I cut my teeth on that CPU

[inasity_rules]

Sounds like a modern Toshiba PLC....

Seriously, though there is enough embedded/PLC coding going on that I imagine people still know what can be done with low clock rates, a simple instruction set, and some clever asm gymnastics. Ever end up scaling a analogue input by multiplying and discarding the low word of the result because the instruction set lacked and efficient DIV instruction? And if I'm that old, I'll go drown my sorrows in alcohol now! :)

Re:I cut my teeth on that CPU

[mcrbids]

For me, it was the VAX 11/750. That computer was gods-level engineering compared to the cheap sh17 we use today.

ECC RAM? Oh please! The VAX would identify bad RAM spots on the fly and remap them as needed, reloading the contents of the RAM from disk as needed and the end user never had to know. There was a utility you could run to identify which RAM card (roughly the size of a dinner plate) to remove, that you could swap out without turning it off!

Similar with hard disks: you could mark a drive to be disabled, and it would move all the files around without shutting down the system (if there was enough space) so you could upgrade the drives.

Kicker: you could hot swap CPUs if you had a multi core system. (The one I used had only a single CPU, about the processing power of an 80286)

Sometimes, I weep a little inside when I see what 1U "enterprise class" hardware today looks like compared to real, manly man stuff like the VAXes.

Re:I cut my teeth on that CPU

[hazeii]

>Mov @offset(r5),pc
Ah, the calling convention for Fortran. I spent a while recoding the hotspots in a Fortran program into Macro 11 - the runtime dropped from several minutes to a few seconds.

Re:I cut my teeth on that CPU

[hazeii]

See this comment for an explanation.

Re:I cut my teeth on that CPU

[west]

Ah, the DG. My first job was working on a Data General.

I always wondered why there was a long pause whenever I called technical support and introduced myself as Tom West (at least until I read "Soul of a New Machine").

Of course, while entertaining, that book wrecked ego surfing for me, as it was used for decades in every Computers and Society course on the planet.

Re:I cut my teeth on that CPU

[Anonymous Coward]

Some of the earlier PDP 11s had memory mapped registers, so you could run small snippets of code in the registers.

Re:I cut my teeth on that CPU

[ackthpt]

My High School was lucky enough to have one gifted to it. Apparently, Hugh Hefner (Yes, *that* Hugh Hefner) was an alumnus of my High School and had his daughter come to the school to present the check. I remember learning Assembly, Fortran, Watfiv, PL\1 and I think Cobol as well as rudimentary LED and circuit logic either on it or the IBM 360 downtown. Yes, once upon a time the Chicago public school system had a very good education system. I even remember breaking into the 360 and messing with other school's (and students) teachers assignments (teacher's password: pencil....really?! and taped to the desk next to their terminal. /facepalm).

I was stunned in the mid 1980s I was still required to learn RPG, PL/1 and COBOL (COmmon Baboon Oriented Language) punching the code onto a deck of cards, which were batch run on an antique IBM 360. Really. We had this spiffy PDP 11 where you could sit down and type on a terminal and get immediate feedback and test bits of code and they were still teaching with card punches.

Rather then sit in the cardpunch lab, as I was already a student programmer, I typed up my code in EDT and then wrote it to tape (TU16) using my own ASCII to EBCDIC converter. I'd then take the tape over to the IBM and run the tape out to a deck of cards on the card punch (nobody had replaced the ribbon in it in years so many of the letters on my cards were unreadable, but I didn't care it generally ran correct on the first pass.) My professors were puzzled why only my cards were nearly blank, when I turned in my decks and printouts, but once they realized they had a smarter than the average bear in their class they just looked at the output and put A on it. I was writing system utilities at the time, included a major rewrite of the college test scoring system.

Oddly, PL/1 would come back to me in the early 90s, when we brought in a PR1ME 750 system (not to be confused with anything DEC) None of the compilers we had, excepting PL/1, would handle input buffers larger than 512 characters, so I converted the test scoring suite to PL/1. Later I'd rewrite it a 3rd time in c on a DEC Alpha.

They finally retired the PDP 11/50 and IBM 360 in 1986, as PCs were starting to cut into mainframe sales and DoD cutbacks were beginning to thin the herd of giant mainframe companies.

Re:I cut my teeth on that CPU

[Giant Electronic Bra]

lol, yeah, done all kinds of crazy stuff. FFT on a processor so slow that to get 100ms sample rates you had to do it in less than 12 clock cycles.

Re:I cut my teeth on that CPU

[hazeii]

The place I was working that used PDP-11's had analogue computers too; they had lamps in the cabinets (to keep the temperate constant), and opening a door made you seriously unpopular due to the drift caused by the air currents.

Re:I cut my teeth on that CPU

[hazeii]

Ever walked down a corridor carrying one that was still spinning down, and tried to turn the corner? :)

Re:I cut my teeth on that CPU

[hazeii]

Yeah, that was the biggest culture shock when PC's came out (and made loads of us think they'd never catch on). We were used to getting snail-mail circulars about bugs that had been found in the OS, along with the consequences, the source code, and the fix or workaround.

Re:I cut my teeth on that CPU

[g01d4]

I took an analog computing course my freshman year in college. My first job included programming (fortunately for me mostly FORTRAN) and system administration of an AD-10 hybrid (digital + analog) computer that was hosted by a PDP-11/34. That PDP also fronted a MAP-300 array processor. When someone suggested paying for a direct hardware interface between the AD-10 and MAP for a specific problem I volunteered that I could do it on our new VAX 11/780 for a fraction of the cost. The solution might take a little longer but since there was no real time requirement there was no point of all the expense and complication (and I'd have to program it :). Anyone remember all the indirect addressing modes of MACRO-11?

Re:I cut my teeth on that CPU

[hughk]

The original PDP-11 standard Fortran would churn out pseudo code. This would be a list of addresses into the library with a link via R4, something like jmp @(r4)+. This was slow but actually quite elegant (easy to switch amongst the innumerable hardware variants via choice of library), but it was hard work to link because the entire program was external references. F4P cost a relative fortune though and I didn't get to use it until a time working at Digital.

Re:I cut my teeth on that CPU

[Anonymous Coward]

Yeah, I know what you mean. A lot of newer technology is just less good. Verizon 'improved' my phone service so now they 'back up' my contact list. Its great, every month or so they restore some old worthless crap over the top of my contact list on my phone. Ain't progress great? Now I carry around an actual physical address book.

So complain then. And definitely do not pay the bill until they fix what they broke.

Re:I cut my teeth on that CPU

[hb253]

breech
/brCH/
Noun
The part of a cannon behind the bore.
Verb
Put (a boy) into breeches after being in petticoats since birth.
Synonyms
posterior - backside - rear - nates

Re:I cut my teeth on that CPU

[Anonymous Coward]

I remember a departmental server, in 1976, with 3 2.4 MB disk drives.

Some of the PDP-11s (and the LSI-11s) didn't have supervisor mode.
I do remember split-ID space, which almost made 64kB data bearable
(but only if you had the later model KT-11C).

Then again, the (pre-Bourne) shell could run usefully in one 8kB "page".
And shell scripts had 'goto's.

As for why it was fun? The instruction set was so regular, I can still
cough up a few instructions, and all the register modes/values, and
write them on the fly. And you could play 'kill the rotating bit' on the
console switch register.

Re:I cut my teeth on that CPU

[Anonymous Coward]

Sounds like a "CAMAC" (?sp), which was some version of a PDP-11. They
were really popular for simple controllers, and used at a lot of physics labs.
IIRC, Cornell had a bunch of them at CESR and CLEO back in the '80s,
and they're prob'ly still there now.

Re:I cut my teeth on that CPU

[Anonymous Coward]

**GEEK DEFICIENCY** Turn in your geek badge at once.

Re:I cut my teeth on that CPU

[datavirtue]

Yeah, does anyone know the specs on the modern YUKs?

Re:I cut my teeth on that CPU

[ak3ldama]

I was fairly certain I speeled it wrong but I just did not care.

Re:I cut my teeth on that CPU

[Anonymous Coward]

With all those factors modern machines can be said to have grown by a factor of maybe a trillion.

And we still are no better off due to sloppy code. Back in the old days you *HAD* to do it right, but not now...

Re:I cut my teeth on that CPU

[Anonymous Coward]

what? perhaps you haven't heard of unix yet?

Re: I cut my teeth on that CPU

[sartin]

Ah, good old 014747.

Re:I cut my teeth on that CPU

[Arker]

"And we still are no better off due to sloppy code. Back in the old days you *HAD* to do it right, but not now..."

Eh, I remember a lot of stuff done very wrong in the old days. But yes, there were more limits on it. You couldnt require 2gb for hello world when top of the line computers only had 2mb.

The trend since then is to every higher level developer languages, ever cheaper "programmers" trained ever more quickly to string together libraries they dont understand into applications they cannot debug to run on machines they will never understand.

Re:I cut my teeth on that CPU

[Anonymous Coward]

I did my PhD in genetics on a PDP-11 (1987-92). I would have been able to cut that down by at least a year if we had a larger machine. I was running FORTRAN 77 code and the statistical package BMDP.

I remember in our genetics department, one of the professors purchases a 286 PC with a 287 math co-processor and it beat the PDP-11 hands down.

Re:I cut my teeth on that CPU

[leedsj]

I once had sources to Mark Turmell's early games written on a PDP, such as Squash, all collision detection peformed with an 80x24 integer array and VT52 cursor controls (I bet nobody heard about those early beginnings before.) Before Atari, C64, Nintendo, etc, we played video games on VT52 and VT100 terminals. :o)

Apologies for being all tech history anoraky, but what year was this?

Re:I cut my teeth on that CPU

[RockDoctor]

The trend since then is to every higher level developer languages, ever cheaper "programmers" trained ever more quickly to string together libraries they dont understand into applications they cannot debug to run on machines they will never understand.

I'll get the sheep dog to herd some kids onto your lawn, while you load up.

Very Cool!

[pauljuno]

It's kind of interesting to read these sorts of news articles. You would have thought they would have replaced these relics long ago. I wonder if the PDP-11's used Macro-11 like the VAX-11s. I remember learning to program on a VAX/11-750 in high school. My first real exposure to "real" computers. Up until then I only used TRS-80's. Thanks for the flashback and making me feel old!

Re:Very Cool!

[Nutria]

The VAX uses the 32-bit MACRO-32 assembler.

Ken Thompson

[Anonymous Coward]

if they could pull him away from Google.

I just read a story that Ken reported that once he and Dennis Ritchie independently coded an assembly language function. Their implementations were each 20 lines of code, and turned out to be identical line for line! Now that's a good creative partnership.

Re:Ken Thompson

[Anonymous Coward]

Not surprised. Form mirrors function. A short function, in any language, written to the same specs by several competent programmers, should all be just about the same. 20 lines of assembler == 2 to 5 lines of C code... :-)

Re:Ken Thompson

[BrokenHalo]

Not surprised. Form mirrors function. A short function, in any language, written to the same specs by several competent programmers, should all be just about the same. 20 lines of assembler == 2 to 5 lines of C code... :-)

True. Except that a modern assembler will accept lower-case as well as upper-case. :P

Re:Ken Thompson

[emt377]

Except that a modern assembler will accept lower-case as well as upper-case. :P

I'm pretty sure the pdp-11 unix toolchain was case sensitive. The DEC ones universally used Radix-50 (three characters per 16 bits, 6 chars in 32 bits) and weren't case sensitive. Same as their OS'es used for file systems, task names, etc. Devices only got three chars (one word) plus a LUN sequential index. This is where 6.3 file names, 3 char device names, etc comes from.

Re:Ken Thompson

[BrokenHalo]

True, but I was more referring to the fact that in a world where our code was input via 80-col cards with a 10-button punch machine, we either said it in upper-case or we didn't say it at all.

Re:Ken Thompson

[TangoMargarine]

Well yeah...if you're talking assembly, depending on the problem, you could probably have literally one ideal solution. Minimize throwing stuff around between registers, and different ops destroy the contents of different ones...

Re:Ken Thompson

[kermidge]

quibble, but true, and makes more sense: form _follows_ function. Put another way, from a correct function, an optimal envelope of form will emerge to express it; some see it in math, I learned it in engineering, some of it later in coding. Don't know if it's quite original with him, but the principle is oft ascribed to Frank Lloyd Wright. (Were I not lazy I'd look that up.)

Re:Ken Thompson

[tengu1sd]

RAD-50. My first civilian job was support a database using RAD-50 on PDPs. We marched customers from PDP to VAX to Alphaservers before the new management announced the SQL/IIS port. They took a company with over 50 percent market share (Realtor/Real Estate agent) desktops to 5 five percent in 10 years. One customer jumped from a PDP 11/84 straight to an Alphaserver 1000.

Re:Ken Thompson

[david_thornley]

The uppercase-only restriction wasn't because of the cards but rather because of the keypunch machine. The 370s I worked with could accept lowercase letters on cards, it was just hard to make them. (The Control Data computers I programmed with card decks, on the other hand, had 6-bit characters and therefore no lowercase - you used the 6/12-bit character set for them.)

That's just cruel

[Chrisq]

... until 2050 — long enough for a couple of generations of programmers to come and go.

By their short lives I imagine that they must make them work in a high-radiation area.

Re:That's just cruel

[Joce640k]

Why do you think they call it a "half life"...?

Re:That's just cruel

[msauve]

"Generation" doesn't refer to an average lifetime, it refers to the average childbearing age.

Re:That's just cruel

[Anonymous Coward]

... until 2050 — long enough for a couple of generations of programmers to come and go.

By their short lives I imagine that they must make them work in a high-radiation area.

Yeah, but they all eventually get super powers.

Re:That's just cruel

[fnj]

So? A "generation" is commonly held to be 30 years; the average child (note: not first-born) being born when the parents are approximately 30. Secondly, TFA specifies two generations "coming and going", which means two ENTIRE generations pass; not just one passing and the second one beginning.

That is 60 years, not 37 years. TFS, if not TFA, which I didn't read, is officially stupid.

Re:That's just cruel

[Anonymous Coward]

Maybe they're talking about the total service life of the machine? The things are already 30 years old, so another 37 would more than cover two generations.

Re:That's just cruel

[tgd]

So? A "generation" is commonly held to be 30 years; the average child (note: not first-born) being born when the parents are approximately 30. Secondly, TFA specifies two generations "coming and going", which means two ENTIRE generations pass; not just one passing and the second one beginning.

That is 60 years, not 37 years. TFS, if not TFA, which I didn't read, is officially stupid.

Commonly by who?

In virtually all cases, generations are pegged at 20 years. The common "Gen X", "Gen Y", etc are all 20 year spans. In fact, virtually every named "generation" of the last century were equal or slightly less than 20 years.

Even if you go by the average age of first birth, in virtually all of the "1st world", its right around 25. The peak averages are barely 30, and globally its in the low 20's, depending on the source.

So by either definition, there's definitely time for two generations ... and if you're talking about the average time in a given position (which is a more meaningful generation when speaking about engineers), you're looking at more like 15 years -- or time for three.

Re:That's just cruel

[SirGarlon]

I will be working until 2050, you insensitive clod! :-)

Re:That's just cruel

[Anonymous Coward]

In virtually all cases, generations are pegged at 20 years. The common "Gen X", "Gen Y", etc are all 20 year spans. In fact, virtually every named "generation" of the last century were equal or slightly less than 20 years.

Even if you go by the average age of first birth, in virtually all of the "1st world", its right around 25. The peak averages are barely 30, and globally its in the low 20's, depending on the source.

You're confusing "a generation" as a period of passing time with a "generation" referring to a group of people born during certain years or decades. As is intuitively obvious to the most casual observer, TFA uses generation to refer to a time span, not a group of people. A generation time span has always been 30 years, regardless of the change in average childbearing age.

Re:That's just cruel

[khallow]

A generation time span has always been 30 years

Except when you're not speaking of generations of humans from birth to birth which is quite frequent. Here, they're speaking of a career generation which is at best a very fuzzy idea since employees don't magically give birth to new trained employees.

Re:That's just cruel

[dAzED1]

"long enough for a couple of generations of programmers to come and go" - funny, seems like they're precisely referring to a group of people. You really think programmer generations are 30 years? We've easily had 3+ major paradigm shifts (ie, generations) in the last 37 years in the - despite what you're saying - very specific group named "programmers." Is there some reason in particular you think the next 37 years won't have 2?

Re:That's just cruel

[dAzED1]

point in fact they do; interns become code monkeys, code monkeys become designers, designers become architects or managers. If you have an IT group without any seniority, you either have a very tiny, or a very inefficient, IT group. Note too that one can be from a younger "generation of programmers" but be, as a human, older than the more senior programmers; some people get started early, some people get started late. It's not really that complicated, however.

Re:That's just cruel

[CAIMLAS]

Well, programmers have even shorter generations than genealogical generations. A generation can refer to any time increment, sort of like 'epoch'. Programming generations are much shorter - maybe 3-5 years, ie how long a programmer really 'lasts' without a skill refresh, how long they stay in a job on average, etc.

Re:That's just cruel

[DerekLyons]

In virtually all cases, generations are pegged at 20 years. The common "Gen X", "Gen Y", etc are all 20 year spans. In fact, virtually every named "generation" of the last century were equal or slightly less than 20 years.

You're confusing two different things, which isn't surprising since they more or less use the same word.

• "Generation", used standalone, is a noun and describes a measure of time expressed in a fraction of human lifetimes and is generally pegged at around 30 years. ("Three generations ago we did this thing or that thing.")
• "Generation ___" is a proper noun and a descriptive term for a generational cohort or a social generation. ("Millennials are all this social trait or that social trait.")

Re:That's just cruel

[msauve]

Nope. You're conflating technological generations with "programmer generations." Are you seriously arguing that a programmer's career only spans 3-5 years? Or even that one can't have a programming career which is equal or longer than a human generation (20-25 years)? I guess Linus is about done. Keeping skills up to date applies to any career.

"Programmer generation" is a red herring, made of whole cloth. There is no such thing.

Re:That's just cruel

[tgd]

In virtually all cases, generations are pegged at 20 years. The common "Gen X", "Gen Y", etc are all 20 year spans. In fact, virtually every named "generation" of the last century were equal or slightly less than 20 years.

You're confusing two different things, which isn't surprising since they more or less use the same word.

• "Generation", used standalone, is a noun and describes a measure of time expressed in a fraction of human lifetimes and is generally pegged at around 30 years. ("Three generations ago we did this thing or that thing.")
• "Generation ___" is a proper noun and a descriptive term for a generational cohort or a social generation. ("Millennials are all this social trait or that social trait.")

No, actually I'm not. The GP was being a pedantic twat by picking a definition of generation that was a) clearly not what the article was talking about and b) incorrect based on his/her own choice of definition.

I'm not the least bit confused about the GP's twatness, the definition of "generation", or the obvious intent of the author of the story in using the word.

Re:That's just cruel

[DerekLyons]

No, actually I'm not.

Yes, actually, you are.

The GP was being a pedantic twat by picking a definition of generation that was a) clearly not what the article was talking about and b) incorrect based on his/her own choice of definition.

And you're an ignorant twat not only by creating a definition out of thin air that has no bearing on or relation to the actual definitions, but by also by being clueless and thick headed enough to not recognize the difference when they're pointed out to you.

Re:That's just cruel

[tgd]

No, actually I'm not.

Yes, actually, you are.

The GP was being a pedantic twat by picking a definition of generation that was a) clearly not what the article was talking about and b) incorrect based on his/her own choice of definition.

And you're an ignorant twat not only by creating a definition out of thin air that has no bearing on or relation to the actual definitions, but by also by being clueless and thick headed enough to not recognize the difference when they're pointed out to you.

Except that you, and he, are incorrect about the generally approved definition of generation, and the average age of procreation, neither of which are 30 years.

I know this may be hard for you to follow but:

1) The definition of generation that you posted is structurally correct but factually incorrect because "30" is not the number that is standardly used by anyone. It may be used by some people who don't know better, but it is *not* the definition, because there is nowhere in the world where the average generation gap is 30 years. The average in most of the western world is around 25, and globally is about 21.
2) The article is clearly talking about employees, and thus the generation it is referring to is obviously not a biological or sociological generation.

Re:That's just cruel

[TheRaven64]

You think they're deploying PDP-11s now? They were installed in the '70s, so have already seen about 40 years and are scheduled to run for another 37, so they'll see at least 70 years of active service, which is over two generations.

That is 60 years, not 37 years. TFS, if not TFA, which I didn't read, is officially stupid.

Glass house, stones, etc.

If it ain't broke...

[intermodal]

Honestly, it's a system that works. Everything is seen as disposable today, but really, the only reasons we end up getting rid of systems that works these days are either because of support issues (i.e. Microsoft's end of life abandonment of security updates for older products) or lack of available replacement hardware to swap in for failed or failing units.

Honestly, without the need for protection from security holes related to the Internet (and the accompanying security patches), most office workers could get by on Windows 2000 machines with Pentium III processors with probably less than 1GB of RAM and Office 2000 for the foreseeable future.

Not saying we haven't made advances, but I'm definitely saying that modern closed-source computing (Microsoft, Apple) is a system of planned obsolescence.

Re:If it ain't broke...

[Joce640k]

These days you can probably replace them with Arduinos.

Re:If it ain't broke...

[Anonymous Coward]

I find that argument fascinating. Without the need for protection from security holes related to the internet. . . . That's like saying without the need for oxygen, I could live underwater. It's a terrible argument you're making there.

Re:If it ain't broke...

[Karganeth]

If it aint broke, encourage people to waste their time learning an entire language only to be used once.What a great use of time. Not.

Re:If it ain't broke...

[Infiniti2000]

If it ain't broke, then why does it need software maintenance? If it needs software maintenance, then by definition something is broke.

Re:If it ain't broke...

[scsirob]

Yup, but then someone will eventually be silly enough to hook it up to Internet and all hell would break lose.
Better keep that PDP-11 around to do it's job.
(BTW, PDP-11/05 with paper tape boot loader was my first encounter)

Re:If it ain't broke...

[dywolf]

its not so much what its capable of doing.
its what its capable of surviving.

these systems are extremely robust and reliable. its like when people wonder why aircraft avionics tend to be so big and expensive when an arduino could probably handle those tasks too (and yes ive heard that too)...same thing. vibration, rough landings, random mechanics using a hammer to get the screws to line up, or overwrenched a cannonplug.

Re:If it ain't broke...

[dywolf]

then by definition something is broke.

No, not at all.
It's called preventative maintenance, and if you have a car you should be familiar with the concept.

Software in critical applications (and what's more critical than running a nuke reactor??) has to work flawlessly. Believe it or not even old critical software can have things identified that need fixed before they become an issue. Or the NRC issues a directive that in the end means the software needs to be updated in order to implement it.

Re:If it ain't broke...

[sjames]

Not really. Some systems really don't need to (and in fact should never) communicate with the outside world.

Re:If it ain't broke...

[fnj]

Basically, abso-damn-lutely. However, the last PDP-11 model was introduced in 1990. I'm not sure when production ceased,but this hardware has to be pretty long in the tooth by now. How long do you reckon the hardware will keep running? How long will repair parts, even down to the IC level, be available? How long will peripherals be available? A PDP-11 still running in 2050 would be like a 1953 computer still running today.

Re:If it ain't broke...

[Archangel Michael]

I carry a more powerful computer in my pocket.

And you're wrong about the computer needs of today's office workers. Most people don't spend all day in Office 2000. The multitasking needs are much greater than a PIII provides. Not to mention the need for multicore processors doing multiple simultaneous tasks. I currently run about two or three dozen different programs all at the same time. I realize that I am an exception and a geek, but I know people who don't know anything who try to do even more. Office, Accounting, Database, Web, Presentation, Desktop Publishing, Email, Skype ....

I remember trying to run Visio on at Win 2k with a gig of ram on PIII, all I can say "underpowered"

Re:If it ain't broke...

[andy16666]

Well, one of the best arguments for upgrading from a system that old is computation power per watt. It costs a lot to run a very old power-hungry system to do something that a tiny micro-controller can do today for fractions of a penny on the dollar in power consumption. You're paying an exorbitant rate per CPU cycle with something like a PDP-11.

That doesn't quite equate to it always being worthwhile to replace a dated system. In some applications, it is critical that a well tested software/hardware combo not be messed with. I suspect this is one of those cases, where the difficulties and cost in maintaining such a system are deemed more economical than the alternatives.

Re:If it ain't broke...

[scsirob]

Sorry to burst your bubble, but that's not how it works in nuclear nor in aviation..

One of the main reasons things are behind in those industries is paper trails. Rules and regulations. It takes forever and lots of money to get this gear certified. Once certified, it takes an act of God to change it.

A PDP-11 isn't much more reliable than any other system. It has unreliable old-style linear power supplies, unreliable backplane connectors and all parts do fail eventually. Just because they weigh a hundred times more doesn't mean they are a hundred times more reliable.

Re:If it ain't broke...

[intermodal]

You seem to be equating this laundry list of things running at the same time with "need". Frankly, I'm not convinced that present-day "need" gets any more accomplished than was performed by what we had ten years ago in most businesses with the "needs" from then.

I don't measure productivity in the number of bits pushed or number of programs used. I measure it in how useful those bits were and how much was usefully accomplished by those programs. You're simply justifying bloat.

Re:If it ain't broke...

[Cyberax]

So what happens once someone accidentally drops a wrench on a PDP machine? How are you going to source replacement parts and where are you going to find expertise to fix it? And what about the price of electricity to power it?

Keeping old outdated equipment just because it works is NOT a good solution.

Re:If it ain't broke...

[scsirob]

This. "Need' is the most overrated word today.
Where are my mod points when I "need" them.

Re:If it ain't broke...

[cardpuncher]

Indeed. I worked on networking products for RSX-11M and there's very little that the '11s didn't do - multi user, memory protection, standard peripheral bus; even Ethernet made it to the PDP-11. And the 11/70 supported more than a dozen development users with 512MB of (core) memory. Nice regular, consistent instruction set. E-mail, chat, even network file access (mostly). Only pain was creating the overlay trees when your application wouldn't fit in 16 bits of address space. It doesn't really take very long to pick up Macro-16 (the assembler language) - there aren't many instructions and they all pretty much work with the same addressing modes. Training people really isn't a problem in this case.

Imagine replacing the systems, though. Apart from the problem of having to recertify all of the software, I imagine there are a lot of sensors connected up to Unibus interface cards which may well have been custom designed. You'd probably pretty much have to redesign the whole thing.

So much easier to keep the kit and adapt the people.

Re:If it ain't broke...

[SJHillman]

"Just because they weigh a hundred times more doesn't mean they are a hundred times more reliable."

But it is a hundred times more satisfying to shove it off a building when it misbehaves. At least, that's how it works with copiers.

Re:If it ain't broke...

[SJHillman]

Why do you change the oil in your car if it still runs fine? Why do you eat when you're not starving? Maintenance is generally just as much about keeping something running as it is about fixing something that's broken.

Re:If it ain't broke...

[cardpuncher]

Woops, I meant Macro-11. Too long ago...

Re:If it ain't broke...

[SJHillman]

Considering you can still buy a brand new 486 today (although it's a little more difficult at the consumer level), I don't think it will be a problem as long as there's demand. Sure, prices might go up as you become the only customer for those parts, but someone somewhere will gladly take your money for them.

It's actually not too hard to keep stuff running for decades as long as you can still source replacement parts. The main reason there's not a lot of Model T's and Model A's on the road is because newer cars are significantly better... not because it's impossible to keep a Model A running. If anything, it's probably easier to keep a Model A running yourself than a 2013 Honda Civic just because you know every little detail about the Model A and how it works, what its quirks are and it's generally just simpler.

Re:If it ain't broke...

[intermodal]

I don't think a nuclear power plant cares about computation per watt. I'm also unconvinced that computation per watt is relevant to most business computer use as long as the job gets done. At that point, all that really matters is the overall power consumption compared with the actual useful outcome of the work itself.

Say what you will about older processors versus old ones, you'll be right when you're actually taxing the processor. We're talking about a PDP-11, a machine that is killed by my smartphone or a Raspberry Pi on every specification. To be fair, many desktops in service around the year 2000 lose both those fights as well. But it doesn't mean that any particular machine can't do what it was originally designed for.

Re:If it ain't broke...

[SJHillman]

I imagine that nuclear plants pay a little less per watt than most other industries do. Power probably isn't as big of a concern here (especially considering the other costs involved with running a nuke plant).

Re:If it ain't broke...

[intermodal]

Have you seen the cabinet on a PDP machine? Is somebody going to open it by accident and then accidentally drop in a wrench?

Also:

Price of electricity at a nuclear power plant? I'm going to go with "zero".

Re:If it ain't broke...

[jkonrath]

If you dropped a wrench on a PDP machine, you're probably going to need to buy a new wrench.

Re:If it ain't broke...

[dywolf]

you're not bursting anyone's bubble let alone mine. i have considerable experience in aviation maintenance, primarily military, to back me up.
even parts for new aircraft, for new civilian light aircraft, have a robustness to them you wont ever find in consumer electronics. its not merely a matter of being hard to get certified due to red tape clogging innovation.

Re:If it ain't broke...

[rnturn]

``So what happens once someone accidentally drops a wrench on a PDP machine?''

That's easy: buy yourself a new wrench.

In a former life, I've built, disassembled, and rebuilt PDP-11s (34s, 44s, 70s, and a their Q-bus equivalents) more times than I care to remember. You rarely needed an actual wrench for anything... Oh, wait... yeah, you used a wrench to adjust the feet on the racks to keep them level. But that was pretty much it.

Re:If it ain't broke...

[jonwil]

If you need a modern "system that just works" in the vein of these old systems, IBM and others will sell it to you.

Comparing a PDP to a modern desktop PC is like comparing a big-rig truck to a Toyota Corolla.

Re: If it ain't broke...

[ogdenk]

The expertise is still out there and there's a dozen places still selling pdp11 parts though they are a bit pricy.

The later QBUS MicroPDP 11's are probably the easiest to keep going and the most modern.

The trouble is finding folks with RT11 or RSX experience willing to work for 35k with no benefits. Far easier to find wannabe sysadmins with half-assed Windows server training in this category.

For the record 2.11BSD for the PDP11 still sees updates.

Re:If it ain't broke...

[NeoMorphy]

My work laptop has a quad-core. It seems that at least one, sometimes two processors are required for the various security software that is always running on it. There's also other "maintenance" software that runs in the background. I can usually count on having one or two processors for work but it seems like Windows XP's scheduler makes the quad feel like a uni-processor.

It would be nice if I could run Linux, installed and maintained by myself, then I would have plenty of processing power, but that will never happen.

Re:If it ain't broke...

[intermodal]

A big-rig and a corolla aren't that different in most respects. most of the systems present are similar, even if they are specialized to their vehicle's intended use. The two simply are tailored to different duties.

Re:If it ain't broke...

[dbIII]

Yup, but then someone will eventually be silly enough to hook it up to Internet and all hell would break lose.

That's right :)
http://en.wikipedia.org/wiki/Colossus:_The_Forbin_Project
"There is another system"

Re:If it ain't broke...

[Sique]

Well, one of the best arguments for upgrading from a system that old is computation power per watt. It costs a lot to run a very old power-hungry system to do something that a tiny micro-controller can do today for fractions of a penny on the dollar in power consumption. You're paying an exorbitant rate per CPU cycle with something like a PDP-11.

I've seen worse. I know of a site where an old phone switch is still in service to handle a single door, e.g. the doorbell and the door lock are connected to the phone switch, pushing the button will ring a specified phone in the building, and then the phone user can unlock the door by dialing a code. The phone switch was once designed to handle up to 750 parallel calls and is powered by 1500 Watts. Due to the replacement of the phone switch with an new one, almost all stations were replaced by new ones, except that single door-opener-phone and the doorbell. As far as I know, the operator of the building seems to be unable to replace the door lock with a new one that can be operated by the new phone switch, thus the old one is still in service, using 1500 Watts x 24/7.

Re:If it ain't broke...

[NixieBunny]

Working in a place like a nuclear power plant will change your perspective on replacement of old stuff. The computer HAS to work. A PDP-11 is simple enough that it can be completely understood, so it's possible for a human to certify that it will work correctly. A modern computer is too complex for a human to understand, so there's no way to be sure that it's going to do the right thing. (If you don't believe me, then tell me every task that your computer's doing and why it's doing them.)

Re:If it ain't broke...

[Anonymous Coward]

If you did not know that has been done before - Yes a PDP-11 was pushed off a building - at Cal.

Go bears.

Re:If it ain't broke...

[UnknowingFool]

All that said you missed the most important detail: "These go up to 11." :P

Re:If it ain't broke...

[RabidReindeer]

its not so much what its capable of doing.
its what its capable of surviving.

these systems are extremely robust and reliable. its like when people wonder why aircraft avionics tend to be so big and expensive when an arduino could probably handle those tasks too (and yes ive heard that too)...same thing. vibration, rough landings, random mechanics using a hammer to get the screws to line up, or overwrenched a cannonplug.

I think it's more of a matter that A, when things are mission-critical, you become conservative and don't swap stuff out every month and B, a large, standardized object is easier to swap out than one of an incompatible collection of newer odd-sized smaller things. Some people could also say C, repairability, but there are limits to what's repairable in-flight anyway, and there are fewer things that need repair when everything's on a single chip that's so small that you could pack 3 primary and 2 spare units into the same-sized space as the older discrete or TTL circuitry with plenty of spare room for shock padding and radiation/pulse shields.

Even the Shuttle gradually migrated to electronic controls. Long after the astronauts' watches packed more computing power. But a whole different dynamic applies to stuff like that. Some things can muddle by for a long time without the latest and greatest.

Re:If it ain't broke...

[LWATCDR]

There people sell new PDP-11 Systems. http://www.logical-co.com/dec-replacement-systems/

Re:If it ain't broke...

[hjf]

So what's your point? Want reliability: don't use a PC. Got it.
But that doesn't mean the PDP/11 is the only reliable system ever and "they don't build them like that anymore". You can get modern machines in MILSPEC if you want to pay for them . But you don't need to go that far. INDUSTRIAL-grade is good enough.
And honestly, I trust a good PLC, say, Siemens (I have no experience with american brands), to be more reliable than a 40 year old PDP, no matter how well built it was. Siemens has been making automation controls for a LONG time and their products are really good, and I'll guess most of their bugs have been solved in all these years.

As for PLC programming, remember they're not done in "code", but rather in "Ladder Diagram" so they can be well-understood, and have a graphical representation of the process, which is usually more easy to understand than thousands of lines of code.

Re:If it ain't broke...

[RabidReindeer]

If it ain't broke, then why does it need software maintenance? If it needs software maintenance, then by definition something is broke.

No. A lot of software maintenance is required by external changes, not because the bits wore out one day. Self-contained systems like the ones under discussion are fairly immune to that, but more mundane stuff is a complex mix of social requirements, government regulations, changes in connected technology (including, but vastly exceeding injected network malware), OS upgrades, feature requests, etc., etc., etc., etc.

"If it ain't broke" is false economy when it comes to computers. In most systems, it's not a case of will it break, but when. You can consider an app to be "unbroken" if it is unchanging, but important people will disagree strongly if the world that the software is part of changes and the app (or hardware, or OS, etc.) doesn't change to match.

Re:If it ain't broke...

[Archangel Michael]

I am convinced of the "need". I was convinced in 1998 when I ran an ISP and had to process a huge database. Stick with me for a minute please. When I first started processing the database, it took a few minutes, as time went on, and our business grew it took increasing amount of time, and eventually I was unable to process the database to summarize the data in a timely manner. I upgraded the computer and the data processing went from over 1 hour to three minutes.

That is when I realized that speed doesn't equate to how fast things get done, but rather whether or not you actually can do something. This is key, because if you cannot do something because the computer cannot do it or at least cannot do it in a reasonable amount of time, you end up not doing something, so that something doesn't get done. It is a loss of productivity.

Computers, above all else, don't offer us Word Processing, Spreadsheets, Databases, Presentations or whatever. They offer productivity increases that are almost invisible to casual observers. Pulling data from a database, inserting it dynamically into a Presentation Slide is productive. Giving that presentation using Live data is amazing. In 2000, that was nearly impossible to pull off with Office 2000.

And now, instead of having to prove the data in a static presentation, all you have to do is approve the source, and let the data present itself, dynamically.

Do most people, not use this capability? Probably. Does it mean that the powerful hardware doing multiple things (Presentation, Processing live data simultaneously) isn't necessary, hardly. Do not limit me by what "most people" do or need. I am not "most people" and in fact, "most people" aren't like "most people" in some form or another.

Re:If it ain't broke...

[thegoldenear]

Not if they're not attached to the Internet.

Re:If it ain't broke...

[rochrist]

Probably not a big issue when the thing is running a NUCLEAR POWER PLANT.

Re:If it ain't broke...

[Anonymous Coward]

You want to talk planned obsolescence and closed source? How many times do you want to pay 20K to replace a 5 mb original spec hard disk for that POS? And let's not even start on the tape drive.

Look, I'm all for nostalgia and whatever, but in our lab all it took was a discarded labView card from some other group and a weekend of 'programming', and now we're on cheap commodity hardware. Fuck the pdp-11.

Re:If it ain't broke...

[cangrejoinmortal]

This actually answers a question I was making myself about this sense of reliability of old equipment, In my experience its a nightmare to service and maintain old equipment. One of my first job was in the manufacturing (machining) industry, it was a small operation and it needed old computers with old software that controlled some old (~6 years back then) but very expensive quality control and milling equipment; those were all old PC's, with old ISA buses, old and insanely expensive RAM (that eventually failed), I eventually convinced my boss to replace those things with cheap assembled PC's, we had to buy PCI expansion cards for all the serial and parallel ports that the measuring table seemed to need but it worked out just fine. There was a kind of sentiment that those machines were somewhat magical and no new PC's would ever be able to connect successfully and run the aging software (the equipment was second hand, out of support, software update was in the thousand dollar range), they depended on them so the things had to be reliable, right? If we add regulation to this attachment that management has with old 'working' equipment we get the nightmare that it is to deal with medical equipment and (now I know) nuclear facility equipment.

Re:If it ain't broke...

[satsuke]

There are companies that make a living refurbishing PDP11 machines.

I know of a couple companies that keep them around for their call center ACD controllers (Rockwell Galaxy - PDP11/94).

A lot of these machines still use discrete components .. so fixing them is economically feasible.

ACD integration has a funny way of becoming very very customized. The nominal savings of going to a newer platform is outweighed by the changes required in support and IVR systems that feed into it.

Re:If it ain't broke...

[robthebloke]

I visited a nuclear power station a few years ago, and was horrified to see that the newest Nuclear Power station in the UK is running on computer systems from the late 70's (when the station was designed). Apparently they'd never be replaced, ever, simply because they hadn't considered the possibility when designingthe power station. Attempting to upgrade the computer systems would cause, if not a melt down, certainly a situation serious enough that they simply will not make the change.

Re:If it ain't broke...

[inasity_rules]

Hell no. That has no place in any industrial/control system. Especially not if safety is any concern at all.

Re:If it ain't broke...

[hazeii]

I've never seem a PDP-11 power supply fail (probably worked on around a hundred in total). The units I still have must all be over 30-years old, and all working perfectly (some of them do whine a bit, which seems really weird for linear supplies but there you go). In fact, the only dead DEC I ever experienced was hit by lightning.

PC's, on the other hand....

Re:If it ain't broke...

[Anonymous Coward]

As for PLC programming, remember they're not done in "code", but rather in "Ladder Diagram" so they can be well-understood, and have a graphical representation of the process, which is usually more easy to understand than thousands of lines of code.

The term you're grasping for is "relay ladder logic". Rookie. And it's done that way so that really old folks that worked on actual relay cabinets for process control could take the as-built drawings for those controls and put them directly into a PLC using a "language" they already knew. Most PLCs have higher level code modules (BASIC, C) if you want to use that instead of RLL. Now get off my lawn.

Re:If it ain't broke...

[ahabswhale]

A PDP-11 isn't much more reliable than any other system.

lol...lets see if your $400 Acer laptop is still working 30 years from now.

Re:If it ain't broke...

[jimbo]

And most of those programs are completely dormant at most times awaiting user input or I/O. People get excited about quad core smartphones too but have no apps truly utilizing all those cores.

Re:If it ain't broke...

[Erskin]

Well put, my good sir. And, for the record, that "graphical representation" can easily be just as much a mess of spaghetti as any other type of code.

Re:If it ain't broke...

[fnj]

But will you be able to buy a new 486 in 2050?

Re:If it ain't broke...

[hjf]

Nope. GP is a retard. "Ladder is for old engineers" is a stupid affirmation.

"The Right tool for the job" is what you should be looking for. Are you doing simple industrial control? Are you handling many inputs and outputs AT THE SAME TIME? BASIC is sequential. C is sequential. Are you really going to deal with multithreading, semaphores and all that crap? Ladder is great for that. C? Not so much.

You CAN have modules for things you can think of "sequentially". You have to be a complete tool to do math functions purely in ladder. But sensing switches and turning motors on and off? Easy!

Now: combine both! Have your Machine Vision code in C and have it output stuff to your main ladder proces, then send off to your SCADA system for data logging and display.

Same thing with FPGAs. VHDL/Verilog describe circuits. They don't describe sequences. It's a pain to do sequential operations in those languages (you simply can't: you need to describe a FSM to do that). But you CAN use a CPU module and write a program and run it inside the FPGA. Sometimes, the complexity of the CPU is less than the gates you'll use for your sequence. Sometimes, the time you'll spend implementing and debugging a simple, non-time-critical function (like, for example, USB negotiation) on HDL is more than you'll save by just throwing in a small CPU and some firmware.

Re:If it ain't broke...

[intermodal]

In that specific case, I tend to agree with your need. But you're dealing with a specific case. Much as a programmer who has to have a box that can perform builds in a timely manner, large databases are certainly demanding.

But in the end, once you finish describing situations that do present a compelling cases for newer and more powerful hardware and expansion of software that does what the old systems do except faster or better, you revert again to claiming things to be "needs". Both you and I know that in business, the needs of one and the "needs" of everyone else are often very different beasts.

Your ill-defined "productivity increases" are something I question wholeheartedly. Attack the specific example of Office 2000 all you want, but it was one of countless examples out there. For every one out there who needs what you describe, there's probably 10 that either "need" it or don't need it at all.

Re:If it ain't broke...

[Anonymous Coward]

Maybe not the power supplies, but the fans on the PDP-11/45 and PDP-11/40 I used back in the '70s used to crap out pretty regularly. And the memory would start to flake out any time the computer room (!) got above about 85F.

Re:If it ain't broke...

[hazeii]

True - but then I've always seen fans as limited-life replaceable items

Re:If it ain't broke...

[narcc]

Well said.

Your post puts me in mind of Chuck Moore.

Re:If it ain't broke...

[EETech1]

We have an S7-1200 that has problems with random PID outputs (and the output LED) staying on all the time with no error codes. Reset the PID and all is fine for another day, or week.

Three Siemens factory guys came to look at it, (after the distributor spent a week with it, and sent the code in to the factory) and they all say everything with the program and install is fine. Flashed new firmware, verified everything, swapped the CPU and I/O modules but the results are always the same.

Now we are waiting for a new PID block due mid July, and can't ship the machine without a true fix in place.

The Siemens distributor is ready to put Modbus and (non-Siemens) remote I/O on it to take the PIDs out of the PLC!

Normally PLCs work reliably, but I'm glad this one isn't in a reactor somewhere.

Cheers

Re:If it ain't broke...

[hjf]

Well... be glad you're not using Delta PLCs? =D

Re:If it ain't broke...

[Anonymous Coward]

I trust a good PLC, say, Siemens

That's what people think until they get Stuxnetted.

Re:If it ain't broke...

[Alioth]

The Harwell WITCH (built from 1949 to 1951) runs today.

Re:If it ain't broke...

[metaforest]

Or large format CRTs. Just sayin'

Re:If it ain't broke...

[Anonymous Coward]

One of the main reasons things are behind in those industries is paper trails. Rules and regulations. It takes forever and lots of money to get this gear certified. Once certified, it takes an act of God to change it.

A PDP-11 isn't much more reliable than any other system. It has unreliable old-style linear power supplies, unreliable backplane connectors and all parts do fail eventually. Just because they weigh a hundred times more doesn't mean they are a hundred times more reliable.

I understand some of your comments about unreliable equipment ... but I work on 1980's gear in the same industry and understand the pain of change. We have changed all the caps in the PSU and reliability is way up now. I do not know which device you have but DEC had a design flaw in their PSU for the 11/34 with one rail acting more like an RF oscillator!

I am surprised you say the backplane are unreliable I would still expect them to be OK. The other big issue we have faced is the lack of understanding of impedances .... many systems were connected and 'worked' ..... but get the cables made properly and interface devices properly and my goodness what a difference it makes!

1980s

[Beorytis]

I remember one at my high school (1982-1986). I didn't get to do any assembly programming on it. IIRC we had the timeshare OS, RSTS.

The never ending death of programming languages...

[Anonymous Coward]

Every time I read an article about supposedly dead languages which are never dead and will never be, I wonder if the only reason is because the big companies depending on those languages just want new juniors to focus in the relics of the past in order to avoid paying annoying ciphers to the senior people.

But then again, I may just be a paranoic looking for his foil cap.

Clones

[Anonymous Coward]

Actually in Soviet Russia there was a personal computer named BK-0010, popular in the beginning of 1990s. It has a compatible CPU instruction set, and even came with a FOCAL interpreter, similar to the one you could find on a real PDP-11.

Assembly programmer.

[morto]

Just a small correction. The language is Assembly. Assembler is the tool. Best regards.

Re:Assembly programmer.

[Anonymous Coward]

Good point. It's probably just a detail that was lost on the Indian programmers GE has hired to maintain the assembly for these systems.

Re:Assembly programmer.

[Anonymous Coward]

But but but... Linux is written in compiler!

Re:Assembly programmer.

[Anonymous Coward]

Oh, I though they were looking for people to program a new assembler.

Re:Assembly programmer.

[tgd]

Oh, I though they were looking for people to program a new assembler.

Considering most of the PDP-11 coding I did, my fingers, a pencil and some paper was the only assembler ... there was definitely a time I would've been in support of programming an assembler.

Re:Assembly programmer.

[dpbsmith]

If you want to get technical: the language was MACRO-11. Which is an example of an assembly language. I program in "C#", not in "compiler."

In the Digital world, the name "MACRO" stuck because there were very early assemblers for the PDP-1 that did not have macro capability. So "MACRO" was the name for the assembler that did. In subsequent machine generations, "the" assembler was usually called MACRO even though as far as I know there weren't any assemblers without macro capabilities.

And perhaps I should add: the reason that it's called assembly language is because of drum memory. The usage dates back AT LEAST to the IBM 650 and Symbolic Optimal Assembly Language (SOAP). "Assembly" was short for "optimal assembly." Each instruction contained within it the address of the next instruction--they weren't sequential--and "optimal assembly" was the process of calculating how long each instruction would take so that the next instruction could be placed at the right location on the drum that it would be almost under the head when the last instruction had completed. "Optimal assembly" was the memory placement aspect of it.

The symbolic optimal assembly program added to that the advanced capability of allowing programmers to refer to instruction codes by convenient, easy-to-remember mnemonics like UFA and STA, as well as the capability of giving your very own names to instruction locations.

For some reason, the category name got abbreviated to "assembler" rather than "symbol-" um... symbolizer? Symbolic? OK, maybe THAT reason... and it stuck, even after advanced computers like the IBM 704 started to have random-access memory.

Re:Assembly programmer.

[narcc]

Each instruction contained within it the address of the next instruction--they weren't sequential--and "optimal assembly" was the process of calculating how long each instruction would take so that the next instruction could be placed at the right location on the drum that it would be almost under the head when the last instruction had completed. "Optimal assembly" was the memory placement aspect of it.

Please, a Real Programmer wouldn't even bother with a so-called "optimizing" assembler. Too inefficient. "You never know where it's going to put things, so you'd have to use separate constants."

Re:Assembly programmer.

[Anonymous Coward]

Then you probably know it would have been a lot harder in nearly
any other assembly language. You should probably remember the
opcodes (I still remember half-a-dozen or more), and you can
count the number of instruction words while you're writing it, so
the offsets all work.

DEC

[barista]

Nice to see the Digital logo get used again. While the PDP predates my experiences, several family members worked for DEC during the 80s. I assume Ken Olsen is laughing in his grave at HP's boardroom misfortunes of the past several years.

Re:DEC

[Anonymous Coward]

"If you don't have 36 bits, you aren't dealing with a full DEC." ... :-)

Re:DEC

[Virtucon]

HP? HP threw DEC a lifeline after Compaq drove it into the ground because when they bought DEC out in the late 90s they couldn't absorb the operations. DEC was ripe for acquisition and had a ton of customers but those customers weren't the typical PC folks that Compaq was dealing with. As a result DEC products suffered and lost share. It was a $9 Billion blunder and they took on a lot of debt and responsibility that they weren't ready to handle. That's what eventually led HP to buy them out. To be sure though, DEC products under HP haven't fared well because of the competing lines of hardware, i.,e, HP9000s, 3000s and software (HPUX). It is valid business strategy to buy a competitor and dismantle it or milk it vs. letting it be bought by another competitor.

Re:DEC

[rochrist]

DEC and Prime both followed the same path to willful self-destruction. Probably for the same reason, old like hardware guys who couldn't/wouldn't adapt to the changing condition. Who remembers the DEC Rainbow?

Re:DEC

[Virtucon]

Wait, the Rainbow was something they put together to compete against the PC. Early on it had some features that were very competitive but it was priced badly and was built around CP/M while everbody else went to DOS. The Micro-PDP project still lived on though in the PRO series which were meant to be higher end competitors to the IBM PC. I remember a couple of 8800s I was working with in the 80s that used a PRO-380 as a console subsystem. It ran that bastard Professional Operating System or P/OS and was there to boot the system and to capture console info. Maybe the acronym killed it. It really didn't do much more than a dumb terminal.

Re:DEC

[rochrist]

It ran both DOS and CP/M. Had 8086 and X80 processors, used VT220 display hardware and would only read a proprietary floppy format. You had to buy your floppies from DEC. Compatible with the IBM PC? Not so much. In my job, I'd developed a project management system for the PC, and I had to make sure it ran on the IBM, but also the DEC, Data General, Prime and Wang versions of the PC. Each one was incompatible in it's own special way.

Re:DEC

[Virtucon]

LOL, I had forgotten about DOS on the Rainbow, thanks for pointing that out.

Re:DEC

[david_thornley]

Earlier than that, they put the PDP-11 onto a chip (the LSI-11). The only place I ever saw it for sale was in the Heathkit catalog, as the H-11 computer kit. I really wanted one, but my budget wasn't up to it. This was in the Apple II/TRS-80/Commodore Pet days,

Re:DEC

[Anonymous Coward]

If you do, you might be dealing with an IBM 7094.

You go old timer!

[NormAtHome]

My high school got a DEC PDP-11 in my junior year (like 31 years ago) with dual 8inch floppy's which replaced the PDP-8 with dual DEC tapes.

Glad to see that they're still going but after all these years where do they get parts for them? Didn't Compaq buy DEC and then Compaq merged with HP, does HP still support hardware this old?

Re:You go old timer!

[RobKow]

When you have all of the schematics it isn't too difficult to support the hardware yourself, assuming you're running one of the older TTL CPUs. That, and/or a huge storeroom of spares, and you're set.

Re:You go old timer!

[Anonymous Coward]

As an army technician we learned how to gate chase and use logic analyzers. A valuable experience as to how machines really work indeed.

Re:You go old timer!

[bellers]

I have 3 VAXen at work that are still on 4-hour hardware support contracts. HP has achieved legendary status among my team for being able to show up in less than 4 hours with parts in hand for them, every single time.

Re:You go old timer!

[Gman2725]

You would be amazed at the amount of spares that are collecting dust in closets and back rooms. Companies like to hoard things like this so they can count them as assets, especially if the hardware is old enough for the parts to be valuable again. When my wife was an insurance appraiser, she would often arrive at a company property and find rooms upon rooms of ancient hardware just sitting there for these reasons.

A real file system

[invid]

My first real programming job was on a VAX system. The file system and scripting language were so extensive we were able to use them to create a working configuration control system for our source code. DOS was a toy operating system by comparison.

Re:A real file system

[bellers]

Ugh, you're that guy. Our VAXen at work have such a system and migrating off it to another CM tool while preserving 25 years of revisions is such a pain in the dick its easier for us to keep the VAXen alive.

Security?

[MaxDollarCash]

Am I the only one worried about security of such an ancient system running a nuclear power plant? How long before...

Re:Security?

[thoriumbr]

If the system is running fine for decades, what is the chance that it would suddenly die for no reason next week?
It's a very good hardware platform, made to last for centuries. Is different from your brand new GPU card that will fail and die in 4 years. Mine have not failed yet, but will soon.
Almost all the banking business in the world runs on COBOL, compiled almost 40 years ago, and that keeps running and running. Why replace the core COBOL with Java or .NET, if they are working just fine?

Rest assured, the trusthy PDP-11 will keep the nuclear plant running safe, as it has been done in the past couple decades.

Re:Security?

[AlphaFreak]

Those things are probably not networked. Will Adama would have no trouble having them on board of Galactica.

Re:Security?

[Bucc5062]

Horrible CompSci Movie scene

Evil Guy (EG) to young hacker (YG): Okay, here's the terminal. get to work
YG: What's a terminal, this thing got a usb port for my S6 to access?
EG: This thing is a PDP11. They didn't have USB ports when it was made. Can you get in?
YG: Get in? How the hell to I even log in? This thing was old before my father was born
EG: Get in or Die!
YG: Then it's been a good life for they had actual security on these old systems. Now if it was a PC we'd be done
                No way can I hack into this system. Its just too damn old.
EG: , okay then we just blow it up.

Re:Security?

[scsirob]

With nothing more than RS-232 console lines, if at all, there will be little to worry about. If by chance there is a network connection on it, it will be Thick Ethernet (drilling a hole in the cable to attach a new system), and running a network stack that is nothing close to TCP/IP

Re:Security?

[dywolf]

Ancient != bad/insecure.
If anything it's security through obscurity.

Re:Security?

[Anonymous Coward]

Why should you be worried? The PDP-11 is a fine 32-bit computer, fully capable of real-time operations (with the appropriate operating system). In truth, with some exceptions (64-bits, ARM, VLIW systems such as the defunct Wave Tracer, newer bus architectures, etc) it is up to par with most modern systems.

Re:Security?

[AlphaFreak]

The PDP-11 is a 16 bit computer...

Re:Security?

[Virtucon]

Oh hell no! I'll tell you something, those Real Time systems that were built around a PDP-11 were great technology and were state of the art. A whole bunch of process control, nut just for the nuclear power industry, was driven by PDP-11 technology. It was reliable, parts were readily available and there were more options for controlling other devices with off the wall interfaces. Think about it though, the nuclear power industry has heavy regulation and if you start changing components, like control systems I'm sure the red tape will get heavy and thick quickly. Look at the whole San Onofre debacle and you'll see what I mean. Yes, if you were building a new plant you'd look to newer architecture to support it but not one designed or built back in the 70s, Which in terms of architecture is where the PDP-11 was in it's prime.

Re:Security?

[PPH]

No port of Windows to it, so it will be fine.

Re:Security?

[Anonymous Coward]

it's security through obscurity

One of *the* most understood hacked on systems ever is obscure? Damn I am old...

Re:Security?

[seven of five]

So these guys are gambling they'll never have a hardware failure? Where would they go for replacement parts? Is there a shed in back of HP with a bunch of NOS DEC stuff?

Re:Security?

[Casandro]

They probably have a room full of replacement parts already. PDP-11s are not exactly expensive, typically you'll get them for the price of shipping. It's normal for such institutions to have people hunting around for spare parts.

Re:Security?

[Anonymous Coward]

> If the system is running fine for decades, what is the chance that it would suddenly die for no reason next week?

None at all. I'm sure it will die for a perfectly good reason next time it feels like it.

You should also ask that question again in late January, 2038. Right after half the nuclear plants in the world have melted down.

( http://en.wikipedia.org/wiki/Year_2038_problem )

That is government for you.

[140Mandak262Jamuna]

It is not surprising. Nuclear industry is very heavily regulated. These regulations are not updated regularly. The regulations were laid with whatever was the state of art and they never paid any attention to cost, upkeep or updates. It leads to quite ridiculous situations like maintaining old bugs as is. I don't know why or how. But I hear stories about nuclear customers demanding some buggy behavior to be reproduceable after the software update, even if the update was about that very bug. "Give us a setting/env switch to reproduce the old buggy answer!". Same way the Air Force is still flying B-51 bombers that are older than their pilots, older than their commanding officers, now getting to be older than even their commander-in-chief.

Re:That is government for you.

[kevinT]

The Air Force is flying B-52 bombers. Minor typo, but important to those that care.

Recently I read about a pilot that looked in the older logs of the plane he was flying and noticed his Grandfathers signature as pilot in some of the entries!

Re:That is government for you.

[SJHillman]

Something well made will last for many generations... a fact we've forgotten in modern society with throw-away everything and a new model every year. My neighbor purchased a circular saw in 1946 after he got back from WWII. When he passed away a few years ago, his widow gave it to me and I still use it regularly. I wouldn't be surprised if it still works when it turns 100.

Re:That is government for you.

[dbIII]

For even greater astonishment there are DC-3s flying commercially - to Antarctica! That ski aircraft in the 1952 version of "The Thing From Outer Space" looks very similar to the two that fly down south. They did have a problem section cut out from in front of the wing to be replaced with a longer section, and do have turboprops, but the airframes are from the 30s, 40s or early 1950s. Not being pressurised they don't suffer fatigue everywhere but instead only in a few high stress areas. I think the B52s typically only pressurise the crew compartment so similarly can fly for a lot of hours (or more correctly pressurisation cycles).

Re:That is government for you.

[Alioth]

I own a half share in an Auster Autocrat (light aircraft) that was built in 1945. Still flies today as good as it did in 1945.

DEC

[Lawrence_Bird]

made some amazing equipment. It was sad when they were sold to Compaq. I never worked on a PDP-11 but did work on a DEC2060 during college. This was a 36 bit machine and used an improved version of the KL10 processor originally used in PDP-10s. IIRC there are a few TOPS20 (twenex) enthusiast sites on the net too.

I loved the PDP-11

[richardlr65]

I remember writing code for a bare machine (no OS). I'd set up the code to come in through the serial port and then use the front keypad to key in the instructions in octal to get the code, load it into memory and execute it. It was a little easier with UNIX, even with a dozen people on it.

We had one in our house

[Stele]

I lived with 4 other guys in a big old house in Blacksburg, VA (VA Tech) in 1991. One guy had a knack for finding computer surplus, and brought home *two* PDP-11s, along with a bunch of other "vintage" equipment. One sat out in the garage, while the other adorned a landing in our stairwell.

Re:We had one in our house

[Anonymous Coward]

How did you dispose of the spent nuclear material?

Re:We had one in our house

[hAckz0r]

That's easy. Just like China recycles all our electronics, most Arab speaking countries have a free nuclear materials recycling program.

</sarcasm>

VAXen are still around

[AlphaFreak]

... but the PDP-11 had (has) better interrupt latency so it was preferred for realtime applications. It is not (so) surprising it is still around...

The legacy of the PDP-11

[Anonymous Coward]

The design of the PDP-11 was quite elegant. It should be feasible to implement it in a single chip these days, including memory and I/O... :-)

Re:The legacy of the PDP-11

[AlphaFreak]

Indeed. There are several projects around to implement a full PDP-11 in single FPGA. For example:

http://opencores.org/project,w11

Re:The legacy of the PDP-11

[hazeii]

DEC J-11 was about as integrated as it got.

Re:The legacy of the PDP-11

[emt377]

The design of the PDP-11 was quite elegant. It should be feasible to implement it in a single chip these days, including memory and I/O... :-)

It was indeed elegant, often in very clever ways. For instance, most OS installs came on tape. The boot loader for tape was supremely simple, and consisted essentially of "tell the tape controller to read one block", "jmp .". I forget if it was 2 or 3 instructions, but it took only a few seconds to enter them on the front panel switches. After a full reset the block would the first on the tape, DMA to location 0, caches disabled, I/D space unified, etc. This then caused 256 words (512 bytes) from the first block on the tape to overwrite the code entered, and in particular the "jmp .". And so the secondary boot loader from the tape would start at that location (4 or 6 or something like that).

I had a pdp-11/05, a nice little toy useful for training. It came with a microcode dump, and extender boards permitted single-stepping the microcode. It basically took the source addressing mode, shifted it, jumped to that location in the microcode and computed the source address. The exact same code was then used for the destination address. It would then proceed to the instruction implementation. I forget off hand but I think the entire instruction set fit in 256 words of microcode. It was supremely elegant, clever, neat, and to me as a young kid interesting in all things computer architecture, positively brilliant. Of course, microcode is inherently inefficient. But it was an amazingly elegant exercise in factoring.

Re:The legacy of the PDP-11

[Anonymous Coward]

This kind of thing is also why the executable 'magic number' of
nearly every binary in the world is something near 0407.

On the PDP-11, this is 'branch forward 7 words'. If you forgot to
strip the binary header (8 words) off the boot loader before you
wrote it to disk (and you didn't use the last 8 words), then your
boot would still run (provided you used relative addressing modes
(natural) for everything other than machine registers).

Whatever

[sunking2]

Unless they aren't teaching assembly anywhere anymore there is nothing special about PDP-11 assembly. In the late 80s it's what we learned it on, and I'd venture that it's pretty simple in comparison to anything modern.

Re:Whatever

[fortunatus]

If you compare PDP-11 assembly with 80386 assembly, nothing special - because with the 80386 Intel *FINALLY* got over to the IBM360 style orthogonal instruction set that the PDP-11, VAX-11, Motorola 68000 had implemented for years. If you look at the Intel 8080, 8086, 80186, 80286 instruction sets, you'll see a mucked up bucket of junk. That goes for lots of other micros and minis from the era, too. Back then it was easy to sacrafice the instruction set to save registers and control logic - there was an historical thread of minimal logic CPUs with crappy instruction sets. But there was also a thread of beautiful orthogonal instruction sets comming out of the late 1950's and the 1960's, exemplified finally by the IBM360. The PDP series contributed a lot of orthogonality in this period, too, but the 360 was 32 bits while the other orthogonal PDP's were 18 and 36 bits. The PDP-11 was a power of 2 bit width, half the 360's, and came out looking a heck of a lot like the 360.

DEC made a sample of the mucked up variety as well - check out the PDP-8.

Don't Freak! Prolly A PCI Card Emulator

[cmholm]

This doesn't surprise me. Back in the day, Huge Aircrash had a big investment in PDP-powered test bays, and didn't want to incur the risk/cost of replicating the functions of the assembler software libraries on a new platform/language. So, the PDPs slid out, and rack-mount PCs slid in, featuring a hardware emulator on a PCI card. Minor bonus: a bit more speed. Obviously, Huge wasn't the only customer. Google "PDP hardware emulator", and you'll find a number of vendors.

Re:Don't Freak! Prolly A PCI Card Emulator

[K. S. Kyosuke]

And if you want to learn the assembly dialect, just fetch SIMH - the greastest thing since the bit-slice microprocessor.

Re:Don't Freak! Prolly A PCI Card Emulator

[Anonymous Coward]

Careful, not a lot of people will actually remember Hughes Aircraft. It's been a while now.

Exceptions

[Anonymous Coward]

We've to go to nuclear power plants to get exceptions like this. It's sad.

Most of the IT industry is focused on switching every few years to something different (rarely new).

Well, if you're really that old...

[Pollux]

How come you have such a high /. ID? Were we "too hip" for you, old man?

Re:Well, if you're really that old...

[mrr]

Okay, noob.

Re:Well, if you're really that old...

[Anonymous Coward]

No doubt he was busy working and providing for his family and future retirement and didn't have time to waste with a bunch of pimply fanboys touting the latest re-invented language.

Re:Well, if you're really that old...

[BrokenHalo]

Some of us simply lost our passwords for older accounts and could never be bothered taking the trouble to track them down.

Re:Well, if you're really that old...

[Anonymous Coward]

How come you have such a high /. ID?

Ever tried running a web browser on a PDP-11?

Re:Well, if you're really that old...

[dAzED1]

I've often been tempted to abandon my mixed-case username simply because what I thought was funny 800,000 years ago is not what amuses me now. There's a bit of that sort of thing too for the older group, I think. (I did though have the nick "dazed" for which I forgot the password, but then created this second acount soon enough thereafter to still look like an old timer)

Re:Well, if you're really that old...

[Remus Shepherd]

I have a low ID, but PDP-11s were obsolete by the time I started my career. The only experience I have with a PDP-11 was in translating code from it to use on our newfangled VAX machines.

This entire article is making me feel very old. I wish I had my Atari-400 to cheer me up.

Re:Well, if you're really that old...

[Anne Thwacks]

lynx runs fine

oh, wait ... you are not running Unix on your 11/70?

Re:Well, if you're really that old...

[TheSeatOfMyPants]

Exactly. My first accounts had 6 digits, but now I can't find their passwords, the email account I used is long gone, and one has a now- embarrassing name I wouldn't use anyway (though from what I recall, a lot of users back then had names like that).

Re:Well, if you're really that old...

[TheSeatOfMyPants]

There's a bit of that sort of thing too for the older group, I think

Agreed. All of the names borrowed from fiction/games (Legolas, MotherZero, etc.) would've come into use around the same time, and from what I remember, quite a few ended up using RPG handles that now seem cringe-worthy. Yours doesn't look bad at all compared to most of those.

Re:Well, if you're really that old...

[kermidge]

I'd settle for my old 800, the machine I learned to program on, for my first proprietary program.

Reading the experiences and history from a topic like this are one the big reasons I find /. valuable.

The cost of a reactor meltdown can be high

[Anonymous Coward]

If there is a nuclear meltdown, tens of square miles of real estate will become uninhabitable, and can contaminate groundwater for hundreds of years. ie, very very expensive. So, there is very strong economic incentive to avoid a major failure in a nuclear power plant.

Dave Cutler's work lives on

[cohomology]

Obsolete? Not the ideas.

Dave Cutler designed and wrote much of the popular RSX-11M operating system for the PDP-11. He went on to design the OS for the Vax (VMS). Programmers observed that it was just like RSX-11M, but better. Microsoft hired him to lead a team that designed Windows NT. That kernel lives on in modern versions of Windows.

Re:Dave Cutler's work lives on

[Anonymous Coward]

Microsoft hired him to lead a team that designed Windows NT.

Wow. Way to re-write history. No, I'm afraid that's not how it went down. Dave Cutler developed what ultimately is NT while working for Digital Equipment Corporation. DEC wasn't interested in Dave's creation. Dave unethically and possibly illegally shopped it around. Microsoft was interested. Dave and his entire engineering team left DEC, and went to work for Microsoft, and actually, literally stole DEC's intellectual property and eventually released it as Windows NT. Yes, I am saying that Windows NT is the intellectual property of Digital Equipment Corporation, and Microsoft never paid DEC a red nickel for it.

On a personal note, I am divided about Mr. Cutler. Windows NT might have been the best Windows ever, and NT itself isn't a terrible platform. What Microsoft did to it is unfortunate for users and administrators everywhere, but it essence, NT wasn't terrible. Cutler is an impressive developer... quite amazing... yet it sickens me that what he and Microsoft did was insanely unethical, and no one noticed. Microsoft's main flagship product was STOLEN, and no one noticed, and this is hardly ever acknowledged.

Re:Dave Cutler's work lives on

[Anonymous Coward]

Dave and his entire engineering team left DEC, and went to work for Microsoft, and actually, literally stole DEC's intellectual property

Yeah yeah, there are other examples of that type of allegation, e.g. Fairchild Semiconductor/Intel, DEC/Data General, NCSA/Netscape. And that's just for IT. But did the engineers actually steal literal designs, or just walk away with specialized knowledge that could be used to arm a competitor?

Re:Dave Cutler's work lives on

[Anonymous Coward]

They took the data with them when they left. They began work at Microsoft with everything they had at DEC before they left. They had more than just the knowledge and experience... they had the source. I'd call that stealing. Granted, it was tragic that DEC saw no value in the work, and likely never intended to use it... but it belonged to them, and it was DEC's right to shelve it. I have trouble believing that DEC was entirely hoodwinked here... I think there was probably power at DEC that knew this was happening, some kind of underhanded thing that was off book, so maybe Microsoft did compensate DEC somehow, but if that occurred, I've never heard anything about it (student of History and Philosophy of Science and Technology here).

Re:Dave Cutler's work lives on

[Anonymous Coward]

This was modded up? Seriously?

Sir, where is your reference or citation for such claims. I have read books and interviews about the history of NT and what you wrote doesn't come close.

A) True that DEC wasn't interested in Dave Cutler's work, but there is more to the story. DEC kind of screwed him over and he was not happy about it.

B) He didn't take a finished OS kernel and shopped it around to the highest bidder. That is just false.

C) Dave Cutler had a disdain for Microsoft. In an interview he said that he didn't have respect for the "Microsofties," which were the programmers fresh out of college and working for Microsoft. Basically, inexperienced developers. Smart people, but inexperienced and non-veteran. Ergo, bad design decisions.

D) Microsoft didn't purchase an OS kernel from Cutler. They offered him a job to build a new OS kernel. Why? Because his name had a reputation. And, he initially refused the offer as he thought little of Microsoft. Microsoft, especially Bill Gates, was persistent in trying to hire him and he eventually agreed. He figured that it was an opportunity to build a system to his liking since Gates was promising a scenario where Cutler and his team (colleagues from DEC that he trusted) would be left alone to do their thing. Of course, Gates would regularly meet with Cutler to pester him and try to get the project moving along and delivered, but ultimately Cutler and his team got to do their thing.

E) As explained above, Microsoft heard Cutler was the best and wanted him. It took a few tries to hire him. Does this story sound familiar? If not then read about Anders Hejlsberg.

F) In addition to points B) and D), the development of NT took a while. It took longer than projected, which in a way kind of worked out because NT had steep hardware requirements from the beginning, which Gates didn't like. Gates and Cutler got into arguments about it but Cutler stood his ground. Cutler wanted an OS kernel that was solid and that meant a minimum amount of RAM and hardware to get that level of quality. However, as time passed on NT's steep requirements were starting to become shallow as Moore's Law was in effect. Make no mistake thought, when NT was shipped it had relatively steep requirements. But, in all fairness, it was suited for servers rather than desktop personal computing. But that is all in the past as Windows XP was the OS that finally brought the NT kernel to the masses (yeah, you can say that Win 2K did that but I'd argue that the device driver support wasn't there yet).

Re:Dave Cutler's work lives on

[Anonymous Coward]

Citation please.

Re:Dave Cutler's work lives on

[Anonymous Coward]

Liar!

You're implying that only ONE of Microsoft's "main flagship product"s was stolen.

Re:Dave Cutler's work lives on

[Anonymous Coward]

You're sort of white washing the salient detail. Cutler took his team, and the source code of his work, to Microsoft when he left DEC. Perhaps from Microsoft's perspective all they did was poach an employee of DEC, and Microsoft has deniability regarding DEC's intellectual property eventually resurfacing as Microsoft Windows NT. But Mr. Cutler knew exactly what he was doing. As you say, Cutler was unhappy with DEC. He was a disgruntled employee. But he didn't trash his office when he left... he didn't piss in the coffee maker... he pulled a Jerry Maguire (I'm not gonna freak out... But if anybody else wants to come with me, this moment will be the ground floor of something real and fun and inspiring and true in this godforsaken business and we will do it together! Who's coming with me besides... "Flipper" here?), and he walked out with his source code, the intellectual property of DEC. And the other half of that is he left nothing at DEC, so he not only stole the code (which, granted, he was the author of, but not the legitimate copyright holder), he also very likely sabotaged the work he did at DEC... and deleted it off DEC file shares.

Re:Dave Cutler's work lives on

[softcoder]

I remember it that Cutler left DEC when DEC would not develop a 32 bit real time OS, or a 32 bit PDP-11.
MS hired him partly to rescue NT development, which was in trouble, and partly because they wanted an Alpha ( the next great DEC chip) port of NT.
But I would sure like to know if Anonymous' post has any truth to it.
Sad to say MS's ethics are so bad that one cannot dismiss it out of hand.

Re:Dave Cutler's work lives on

[Anonymous Coward]

I'd like to nominate parent comment for an Academy Award in the category of Best Awesomeness.

Re:Dave Cutler's work lives on

[Anonymous Coward]

The architecture of NT is similar to the Mica kernel for PRISIM he was working on for DEC. But when DEC killed PRISIM he took the ideas but not the code. Mica was not written in C -- it was written in a Pascal like modular language. The NT kernel is written in C (with structured exception handling syntax thrown in).

It's not mechanically translated from some other language. DaveC did what any great developer does who creates mountains of software over a career: You keep building on what you lean and keep the ideas that work well. So that's what they did with NT. You can't copyright an architecture.

Certainly in the last 5 products I've worked on spanning some 15 years you can see similarities in the architecture and coding style. But that doesn't mean I took code from one company and dragged it to another. Code has a flavor from its author like most creative works.

Re:Dave Cutler's work lives on

[Anonymous Coward]

He never worked on Windows NT. Hi did a LOT of work on OS/2 though... maybe it is you who should catch up on your history lessons? (at least try to get some facts straight the next time you post)

Re:Dave Cutler's work lives on

[Anonymous Coward]

There were two money exchanges that happened around the same time.

Round 1: Money settlement to DEC and they committed to porting WinNT to the Alphas. If you ever read "Inside Windows NT", it read like "VMS Internals" except on inferior hardware. One of the reasons VMS was pretty hard to escalate privileges on was that security was built into the hardware. It also had the credo of "death before dishonor". It would crash itself rather than let a mode or memory violation occur. Now you can say that it leaves open a DOS atack, but I would say better that then an un logged memory attack.

Round 2: Intel admitted to copying Alpha arch components. They "bought" an old fab that DEC had (Hudson St.), did a tech xfer to AMD (to make Justice Dept happy), and took over some of the compiler teams.

awesome

[lkcl]

the PDP-11 is awesome. i believe its instruction set was the inspiration for the 6800 (http://en.wikipedia.org/wiki/Motorola_6800#MC6800_microprocessor_design yes it was) which then resulted in the 68000 all the way up to the 68040, processors which both commodore and amiga used to great effect up until the early 90's. at imperial college we didn't write a compiler for 68000 or even x86, we wrote a compiler for the PDP-11 instruction set.

the other thing is: if they're still running PDP-11's in large geometries (.35 micron or even bigger) then chances are it'll be much more robust and less prone to random radiation hits/changes. the kind of thing you really really REALLY want to be still working and under computer control is the "emergency shutdown" procedures in the event of a radiation leak. the LAST thing you want is one of the bits changing a floodgate to "open" instead of "shut" due to a random gamma ray flipping a bit somewhere.

Compare the 68K

[tepples]

The design of the PDP-11 was quite elegant. It should be feasible to implement it in a single chip these days

When you take a PDP-11 and redesign it to be partially 32-bit, you get a 68000.

including memory and I/O... :-)

Like the 68K-driven SOCs in Palm PDAs.

I cut my teeth on PDP/11s and 8s

[Virtucon]

I started out programming on PDP 11s and it still amazing that those systems are still in operation especially in Real Time applications. I remember just before DEC was bought out by Compaq, the PDP 11 business was still doing a billion dollars a year. I guess that much value to Compaq for some reason even in the 90s. They sold the rights off to Mentec. I think Mentec is out of business or out of the PDP-11 business anyway. There are still lots of third party hardware solutions still keeping the architecture alive. It's been awhile though since I needed to find an enclosure or parts for a QBUS system. Maybe I should dust off some of that old code I have and install an emulator.

Re:I cut my teeth on PDP/11s and 8s

[rnturn]

``It's been awhile though since I needed to find an enclosure or parts for a QBUS system.''

Dang! I guess I'll need to keep looking for someone who needs a couple of Q-bus wire-wrap prototype boards. (Freebies I got back in the '80s.)

Re:I cut my teeth on PDP/11s and 8s

[Virtucon]

Every house move I've done has rendered less of my old hardware stash. At one point I had a 3705 front end from a mainframe housing my beer fridge in my garage. Alas, that went to the scrap heap in the mid 90s. All of that old UNIBUS, QBUS, MASSBUS and NI crap I had has long since been consigned to the scrap yards. That and all the old system manuals (big old three ring binders) for VAX, DECSYSTEM 20 etc.

List of great Assembly language programs?

[WillAdams]

One of my favourites was WriteNow, a word-processor for the Mac OS, and later NeXTstep, ~100,000 of assembly language.

Can doesn't mean should

[sjbe]

These days you can probably replace them with Arduinos.

Theoretically true but not necessarily a good idea. The equipment installed is already known to work and whatever issues it has are probably very well understood. Any installation of new hardware is going to bring new bugs and a nuke plant isn't exactly a place you want to beta test things if you don't have to. Plus there are a host of operational certification issues in play. I get why they haven't "upgraded" the hardware.

On the other hand I'm a little bit surprised (only a little) that doing things this way is the most economical method available, even accounting for the risk involved with updating systems.

Re:Can doesn't mean should

[NixieBunny]

Had management been on the ball, it would have begun a lifecycle evaluation in the eighties and replaced the PDP-11 in the nineties with a suitable 8 or 16 bit micro. As it is, they're rather stuck.

Re:Can doesn't mean should

[intermodal]

I'm not sure they see it as "stuck". No more than being "stuck" upgrading from their 1980s or 1990s microcomputer, at any rate. Management seems spot on to me.

Re:Can doesn't mean should

[LWATCDR]

Really which micro? You have the advantage of 20/20 hindsight.
Maybe go with an IBM PC based system? What RTOS would you use? What hardware interface would you use ISA? EISA? MicroChannel?
Then you have the problem of revalidating the software when a new CPU comes out. Remember the Pentium Bug?
Or you could have gone with VAX.... Dead.... Alpha? Dead, 68k? Still kicking around but mostly dead, 88000, PowerPC, MIPS, Spark?
And yes you can get new PDP-11 hardware. http://www.logical-co.com/dec-replacement-systems/

Re:Can doesn't mean should

[LWATCDR]

Not really if you think about it. Let's say that you did move to a more modern system sooner. The logical choice would have been to go to the MicroVax which is also dead. The reason why that would have been the logical choice is that it also used the QBus system. The downside is that the software would have to be ported from the PDP 11.
You could have gone with X86 but you would have been stuck deciding if the hardware should be ISA, EISA, or MicroChannel!. Then you would have to decide what RTOS to use or would you just go bare metal? Then you would have to validate every motherboard, chip set, and CPU that you would use not to mention re-writing the software for the X86.
This is part of the problem with going with COTS hardware in industrial settings. Today there are critical systems running on 486 systems and ISA based hardware running under DOS. If you think about it they are in fact no less outdated than PDP-11 based systems from the same time period. The last PDP-11s were introduced in 1990. There does seem to be a cottage industry building new PDP-11 based computers. They are much smaller, faster, and use modern hard drives and ram but still PDP-11s. Some probably use FPGAs for the CPUs, some use X-86 and are emulated in software, and some may even use ASCs. Just as you can still find people that can rebuild an R-2800 radial engine and by parts for old DC-3s it is possible to keep PDPs plugging along.

PDP-11 is the basis of modern non-x86 CISC

[Trailer Trash]

This CPU is an excellent 16-bit CISC CPU, and it's the forerunner to not only the VAX-11 architecture, but also the Motorola 68000 series and the TMS9900 series. My only gripe with the assembly language was that it required octal instead of hex. The CPU had a lot of addressing modes, 8 registers (6 GP) and even floating point capabilities.

My first actual programming job was in 1988, making minimum wage, working for a physics professor - translating a PDP-11 assembly library that provided a programming interface to a Grinnell graphics processor into VAX-11 assembly. Part of that was turning the various IO calls on the PDP-11 into QIO calls on the VAX.

The Grinnell was incredibly capable for the time. It produced a 512x512 display with the capability for either 8-bit monochrome or 24-bit color. It also had a monochrome camera attached to it. The display had 5 memory "planes", so you could configure red, green, and blue to whichever planes. Writing an image to a plane took a few seconds. Reading an image from a plane took around a minute. It also had hardware 2D graphics commands for lines and squares which were hella-fast for the time.

The professor had just upgraded from a PDP-11 to a MicroVAX II (not sure it was an upgrade) and had 1800 fortran programs that used this library to do various graphics things. A lot of them were throw-aways written by students, but he had some cool stuff for the time to do histogram stretches, change contrast, etc. Yeah, stuff we do with a slider in photoshop now, but then we would run the program and wait for a couple of minutes.

Re:PDP-11 is the basis of modern non-x86 CISC

[Anonymous Coward]

I'm pretty sure "modern non-x86 CISC" is an oxymoron these days. It seems like any instruction set introduced in the past 25 years has been some variation of RISC or VLIW. Not that I'm complaining since I like those models, too. Well done CISC is fun to do, too, though :)

Re:PDP-11 is the basis of modern non-x86 CISC

[Anonymous Coward]

How can you put it down for using octal?!

All the register fields lined up on the octal boundaries, so converting
op-codes to values and vice versa was trivial. E.g., mov src, dst
was 1SSDD, where SS is the source register descriptor and source
register, and DD was the destination register descriptor and destination
register. The descriptors were 3 bits (LSB=indir, top two bits = { 00 =
register, 01 = register increment, 10 = register decrement, 11 = register
indexed), as were the register #s.

I've never before or since seen an assembly code so simple to translate
back and forth to machine code.

Re:PDP-11 is the basis of modern non-x86 CISC

[Trailer Trash]

Good point. But it's a PITA when converting from that to hex, where the machine code is irrelevant and just some immediate values matter. Thankfully most of them were all 1s so it didn't matter.

On another note, the TMS9900 also had an extremely simple instruction format similar to the PDP-11 but with 16 registers, so the register # was a nybble and hex made sense.

Re:PDP-11 is the basis of modern non-x86 CISC

[Darinbob]

I agree, the PDP-11 architecture is pretty nice overall. Orthagonal instruction set, general purpose registers, no I/O instructions, flexible interrupt scheme. It's a straight forward system and easy to learn. Compared to the microcomputers at the time it was amazing, since the 8-bit stuff was amazingly ugly and hacky (by necessity though). I think using Intel in the first PC seriously delayed the advancement of microcomputing so that it took maybe 10 years to catch up to where the PDP-11 was, whereas the 68000 was mostly in the same league after a few tweaks.

Re:PDP-11 is the basis of modern non-x86 CISC

[Trailer Trash]

The 68000 also made the 32-bit leap far more easily - of course that was 10-15 years later.

Compare the 360

[fortunatus]

Note, too, that the IBM 360 instruction set is 32 bit and highly orthogonal, very much as is the PDP-11, and later the Motorola 68000, in fact the 360 instruction set pre-dates the PDP-11 by several years.. Both DEC and IBM were heading in the same direction over some of the same years that way. It's hard to really claim that DEC (Gordon Bell) copied IBM there, but it's also really hard to claim he didn't.

Re:Compare the 360

[fortunatus]

PS - yes, I know the PDP-11 is 16 bit, while the 360 and 68000 are 32 bit - I meant to refer to the "orthogonality" of the instruction sets, and structure of the instruction words, rather than the bit widths when I made the comparisons above...

Re:Compare the 360

[srmalloy]

Back in college, a friend of mine and I wrote an 8080 simulator in MACRO-11. Because the 8080 opcodes weren't structured as cleanly as the PDP-11's, we ultimately wound up handling the step that read the 8080 opcode and passed execution to the interpreter function for that opcode as a 256-way branch -- ASL the 8080 opcode and use it as an offset into a jump table of pointers to the interpreter functions.

Re:Compare the 360

[emt377]

Back in college, a friend of mine and I wrote an 8080 simulator in MACRO-11. Because the 8080 opcodes weren't structured as cleanly as the PDP-11's, we ultimately wound up handling the step that read the 8080 opcode and passed execution to the interpreter function for that opcode as a 256-way branch -- ASL the 8080 opcode and use it as an offset into a jump table of pointers to the interpreter functions.

I designed and wire wrapped an NS16016 (later relabeled NS32016) board but didn't have a toolchain. So I ported xlisp to DECUS C (I had previously ported it to Lattice C for DOS) and wrote a simple assembler in elementary Lisp to run it on a pdp-11/40 I had within easy reach - a PC would have been harder. (And I couldn't afford buying a PC myself.) This was the simplest way for me to get a cross-assembler. Macros were just lisp expressions, and I used lisp token syntax so I could use the stock reader. Having my own assembler I could make it do useful tricks like spit out odd-even EPROM banks (the CPU was strictly 16 bit) in a ready-to-burn Intel Hex (ITH) format. I also tried to make it do branch offset optimizations but ran out of address space, and it wasn't strictly necessary anyway. This got me up and testing the FPU, MMU, etc. The idea was to add a drive interface and port BSD to it, but not being affiliated with any academical institution I couldn't get the source (plus I'd need a proper toolchain at that point), so it ended up in a drawer where it has sat for some 25 years now...

Re:Compare the 360

[david_thornley]

The 360 per se wasn't a 32-bit computer. Instead, it was a set of computers of varying power that all ran the same instruction set, some of which were 32-bit. You could regard 360 machine code as a sort of intermediate language, like the JVM, except that the interpretation was done in microcode.

Amazing longevity

[sjames]

The PDP-11 is amazing in it's long life. Most actual PDP-11s are long gone now, but the platform lives on in the form of PDP -on-a-card products like the Osprey or an emulator.

Guaranteed a Job for 37 More Years...

[Anonymous Coward]

Ha, ha, beat that, you young 'uns! I'll be stepping through code with ODT and living high on the hog 25 years from now when you'll be an expat begging for a job rewriting Java web apps in New Delhi! And livin' in a van down by the river (Yamuna)! Bwahahahaaaaa!!

Only bad side is my ass is sitting atop a nuclear reactor. Note to self: buy lead-lined jockey strap. Or maybe the women like balls that glow in the dark - must check out.

Oh, wow ...

[gstoddart]

Most of the younger /. readers never heard of the PDP-11

Then, seriously, get the hell of my lawn.

My introduction to assembler was PDP-11 assembly on a VAX, and when I got to C the language made a lot of sense since that's the platform C was originally written on.

In the pantheon of Things You Should Know About Computers, the PDP-11 is up there as being hugely important to be aware of.

Where do they buy these PDP-11s?

[Anonymous Coward]

Where does one buy a new PDP-11s?

Did they just buy a pile of them years ago and stash them in a warehouse somewhere?

Re:Where do they buy these PDP-11s?

[Nimey]

You buy them used, of course. Maybe someone's decommissioning one and puts it up for sale, so it gets bought for the value of its parts and stored against future need.

Vax is probably alive and kicking someplaces too

[enigmatic]

The article is about an industry still using the PDP-11,
I am pretty sure we can still find people who use
Vax as well, and plan on continuing.

'cause it just works

[chris.bannan]

The reason the PDP-11 will outlive it's progeny is because it just works. In a time when we hyper-focus on the new and shiny, the PDP-11 harkens back to when engineers designed things to work well. Now, too often, they settle for good enough. But don't blame them, blame their bosses, blame the market and ultimately, blame ourselves.

Real men learned on a PDP-8

[Rambo Tribble]

However, this development, or should I say lack of development, the article details bodes well for those looking for new definitions to the phrase, "a blast from the past", in the event of an accident.

SIMH

[SIGBUS]

The first computer I actually used was a PDP-11/34 running RSTS/E at my high school. A few years back, I decided to throw SIMH on one of my PCs and bring up my very own RSTS system. I laughed my ass off when I discovered RSTS wasn't Y2K compliant, since my high school CS teacher (in 1980!) was warning about the Year 2000 issue even then. I wound up calling the RSTS system "sgtpepper" since I set its clock back exactly 20 years.

SIMH is pretty amazing - but the hilarious thing is that a modern CPU emulates the hardware much faster than the actual machine. I could throw SIMH onto any recent Android phone and run a faster RSTS system with more storage in my pocket than that /34 which filled a full-height rack and had two washing-machine-sized hard drives. Of course, that would really only be worth it for the hack value, not for anything useful, but it would be a fun project for a rainy day.

PDP-11 is the basis of modern 386 CISC

[fortunatus]

Actually, the PDP-11 instruction set *IS* the basis of 80386 CISC - even though PDP-11 style is not a basis of 8086, 80186 CISC. Check it you & you'll notice the 80386 designers flipped head over heels and every which way they could to give the 386 the programmer architecture of the PDP-11. The instruction word structure is not as beautiful - but they tried like hell to achieve an assembly level presentation that matched.

Sad realization

[mooingyak]

Most of the younger /. readers never heard of the PDP-11

I was about to complain about the accuracy of that statement... and then I realized I'm not one of the younger readers any more.
 
/me cries.

Re:The never ending death of programming languages

[swalve]

Maybe a little bit, but smart companies think long term and want to make sure they have talent available when the greybeards wander off.

Nine track 1995 - how about now?

[dbIII]

It's 2013 and my workplace has a small storage shed full of stuff on nine track. In theory it's also on other media and the problem of the clients that sent it years ago - in practice people throw stuff out without transcribing and our stuff ends up being the only copy. We didn't have to get anything read last year but a box full of tapes had to be transcribed in 2011. It's hard to know how much was readable because the file format used can handle missing a few bits anywhere outside the headers and seismic data is a bit noisy anyway - either way every file was read in without modification.
So why is it still on nine track? It's not really our data and the cost of transcribing a few thousand reels of tape is insane, especially since we may need only one or a dozen of them in the future.

PDP-11 racks good for 21st century systems

[k6mfw]

We keep our DEC racks to stack up 19-inch NI and Agilent data acquisition hardware, works great.

MTBF

[fauxscot]

If they are reliable enough, fast enough, and can be coaxed into working, there's nothing wrong with them as controllers. I have programmed them, and am not fond of the machines, but so what? It's not what I'd use for a modern problem, but decades of successful operation is not to be lightly discarded.

What it anything breaks inside?

[aglider]

Like serial lines, 16KB memory banks, power supply or the CPU itself?
Do they have spare parts? Do they know those parts really work now and will work when needed?
I think the plastics around the chips will start degrading much earlier than 2050 ...
I foresee another nuclear incident quite soon... this time triggered by laziness and stupidity.

Re:What it anything breaks inside?

[Anonymous Coward]

There are several companies that manufacture replacement boards/parts etc. It's like the classic car market in a way.
There are even companies that make modern drop in replacements ( http://www.logical-co.com/nupdpq/ for instance )

Re:What it anything breaks inside?

[Anonymous Coward]

I foresee another nuclear incident quite soon... this time triggered by laziness and stupidity.

If only these engineers were as smart as slashdot commenters, right?

Re:What it anything breaks inside?

[aglider]

No, just more reasonable.

Ah! The memories

[AlecC]

I programmed PDP11/23s in Pascal for embedded functions, debugging with a Logic Analyser. I used to be able to read the Logic Analyser screen (aaah... octal), inverse assemble it, and recognise my Pascal code, all in my head. I never got that close to the machine since, despite working on embedded systems most of the time since.

Instruction set (and language) elegance

[jjo]

Having programmed the PDP-8 and then the -11, I can draw two lessons from the experience:

1. You can get stuff done even with an instruction set that looks like a dog's breakfast.
2. Everything is much better with an elegant, well-designed instruction set. You don't truly appreciate this until you've done without it.

These two lessons can be directly extended to higher-level programming languages as well.

Re:Compare the 360 to PDP-6, -10

[Cryptosmith]

I wrote a lot of asm code for the 360/370, PDP 11, and PDP 9, but my undergrad mentor (Uncle Willy Henneman, late of MIT AI and BU, rest his soul) waxed poetic on the PDP 10 instruction set, using the word 'orthogonality.' The -6 was introduced in about 1963 (before the 360) and the -10 was the successful, workhorse version of the architecture till the -20 came out.

Where do they get parts?

[Steve_Ussler]

Amazing...but where do they get replacement parts?

Recruiting PDP-11 people should be easy...

[swb]

..unless MBAs get involved.

Find some CS grads, offer them PDP-11 training and assembler training and a job paying slightly-above-average wages & bennies and tell them the job is guaranteed for the next 30 years.

Right now that sounds pretty good to me --- guaranteed employment on a well-understood platform for 30 more years (although I really only need about 20-25 more years..)

Sure, some guys would rather slave away 80 hours a week to develop iPhone apps, Metro tiles or Web X.0 apps because that's what all the cool guys are doing and it's "the future" (until those jobs are shipped off to the next up-and-coming third world country).

Of course, MBAs would manage to fuck this up by deciding that because it's an "obsolete" technology, you don't need to pay anything.
 

Re:Recruiting PDP-11 people should be easy...

[Arker]

I'll second this only as an old fogey, I admit that I would be tempted by a good offer in the field. Got no PDP-11 experience, and havent programmed at all in years really, I used to enjoy assembler/machine language a lot, but high level languages turned me off so I found more interesting things to do. I'm still sharp and if there was a good chance to get hired doing it I am sure I could learn it quick.

Re:The never ending death of programming languages

[Anonymous Coward]

IBM and Fujitsu both have their own "university" where the next gen are trained and then pimped as high priced consultants, and it's a safe bet that the other big players do something similar.

Disclaimer: AC because I've been pimped by both mega-corps.

IBM and DEC computers are first cousins

[Cryptosmith]

Warning: I'm doing a history geek thing here.

1940s: MIT builds Whirlwind - a beautiful little thing (many KB of RAM) out of thousands of vacuum tubes. They convince the Air Force to use it as the basis for nationwide air defense.

1950s: IBM builds the SAGE air defense system in conjunction with MIT and Lincoln Lab, using people like Ken Olsen. IBM uses Whirlwind and SAGE lessons to build its "scientific" computer line - parallel arithmetic, control store ("microcode"), grand bus architecture, etc.

1960s: Olsen and other SAGE refugees start DEC - PDP-1, PDP-8, PDP-6, PDP-9, etc. No surprise there are clear architectural antecedents going back to SAGE and Whirlwind.

An Observation

[Kozar_The_Malignant]

This story seems to have the lowest average user ID for posters of any story on /. in a long while. Makes me feel like a youngster. I recall the issues of begging for time on the PDP-11 machine in the Engineering Department at my college. As a biology major, I was WAY down the list.

Re:An Observation

[ZahrGnosis]

Is "average user ID" actually calculated? That would be a fantastic metric.

Macro-11 was the language for the first PLC

[Anonymous Coward]

WAY back when Modicon was founded and created the first PLC, I think it used DEC hardware. Later versions, at least some of which used the AMD2901, the 084, 184, 584, and 984 supported "SLM" functions, Software Loadable Modules. In the mid-80's I wrote one to do PID for (admittedly poor) motion control but it was much better than the factory one that solved the loop every (programmable) 5-255 seconds IIRC. This was during the time that Gould owned it.

They've mostly been retired with the tremendous surge by other players, Rockwell (AB), GE, and Siemens presently seeming to have most of the market. I don't believe the present products (owned by Group Schneider's Square D division) have anything in common with those of old.

Fond memories of PDP-11? Hatred for DEC here

[Almost-Retired]

As the now retired CE of a CBS affiliate, my memories of the PDP-11/23 were far less impressive.

CBS bought a few pallets of them, and wrote the control schedule program in pascal that had to be compiled each time it was re booted, for the main network dish, a 7 meter monster from Scientific Atlanta, that was on a turning post Az-Ell mount, used 5 HP Roland jackscrews to move it, very precise and expensive waveguide switches to swap polarities instantly and even had a motor to fine tune the polarity.

The compile requirement was because they could, over the closed caption facility (it has more data that just closed captions) download updated source code to it, and rebuilding it at boot time then installed the new version of the control program.

But it crashed, at first maybe once a week, and the missed satellite switches (no crash alarm so we, unless privy to the schedule) were airing the wrong commercials, pet food in place of toothpaste etc. When the logs showed that we aired the wrong commercial, we of course didn't get paid for that, costing us money.

So we called DEC, who had a couple of recipe followers in the Morgantown WV office and who could usually get around to servicing the maintenance contract in 3 or 4 days. We were precluded from doing anything but reboot it, and the contract said 12 hours, but it was interpreted as 12 business hours, not wall time.

DEC's people replaced everything in that PDP-11 except the frame rail carrying its serial number. Over about 2 years the crashes got worse until it was 4 to 6 times a day.

Our losses got to the point that I asked the guru at CBS if he had a test mule so he could also test hardware the stations sent in as some of it was made in Canada, and customs to ship it back and forth officially didn't have a quarter to call anybody that might give a shit that it was sitting in the border lockup because FEDEX or UPS or us, hadn't crossed a t correctly. 10 grand a day cost to us meant diddly to them. So most of us, who had to send something back for factory repairs, sent it through NYC and CBS, who apparently had the fine art of filling out the many pages of paperwork to get it through customs down pat.

So I called CBS & said this is bull shit, get me a PDP-11 that Just Works(TM). Hugo had DEC move my serial number to his place in NYC, and moved his serial number to us.

His machine did Just Work(TM). The only thing we didn't exchange was the hard drive, a 10 megabyte monster, which because of a paperwork snafu at install time, had a custom satellite location table that because CBS could phone it up, they had helpfully 'fixed'. The second time I called Hugo and got instructions as to how to make that file immutable. The only time I was ever in it as root. Each time was about 2 days putzing to find and mark the locations of all the satellites again because you had to do that sort of thing in off network time. Major PIMA.

But, Hugo then had no test mule as he couldn't even get it to fully boot before it crashed. DEC in NYC was no more help than my local DEC office was, so CBS had no choice but to replace all of them with IBM industrial rated machines, on their nickel at about 10G's a station by the time they'd had much more capable software written. And it, like most IBM stuff, only got rebooted after a power failure from then on.

DEC field engineering, just the phrase running through my mind makes me recall the totally incompetent people they had in their field offices, most doubled as sales force, were hired because they could sell. Totally clueless on a service call, they kept records of course, which is how we finally knew everything but that frame rail (and the outside slip on case) had been changed. But every time they left, saying it should be fixed, the time to crash was cut in half.

Miss DEC? Its like remembering broken bones, I'd druther not.

No Cheers this time, Gene

There are replacement systems

[Anonymous Coward]

There are several companies like this one that offer complete compatability including bus compatability.
http://www.logical-co.com/nupdpq/

Job posting

[Anonymous Coward]

Energetic and self-directed assembler wanted for PDP-11 embedded hardware development. Must have 10-35 years hands-on experience (no emulators) in radiation-hardened development.

(In other words, I don't think anyone alive today who wanted to learn PDP-11 assembler would qualify for the job.)

predecessor to PDP11

[gordona]

The predecessor goes back to the MIT machine in the 60's, the Link-8 which became the PDP-8. It had two DEC tapes, 4KB ram. We programmed it in LAP-4 and had to key in a bootstrap loader on the front panel to load in the OS. At the time we wrote a 64 point FFT and a routine to output the results to a Calcomp plotter. It took 5 minutes to run the FFT. Memory was divided into 2 2KB segments, one for data and one for program. We used one page (512KB) as an executive routine and the other 3 pages to program the FFT which kept the DEC tapes spinning as different segments of the program were shuttled in and out of those 3 pages.

Maybe not

[rickb928]

Around January 19, 2038, they will have either made some changes or changes will be made.

But that's a long time from now.

Re:Maybe not

[manu0601]

For the ones that do not know about it, many systems coded the date as seconds from 1970/1/1, in a 32 bit integer. That will overflow on 2038/01/19.

Re: Maybe not

[rickb928]

Assuming they don't patch the kernel and go 64 bits. That gets them to HDOTU before the next rollover.

Re:Maybe not

[Arker]

That's a common convention on 32bit Unix. These machines are not 32bit and they do not run Unix.

Spacewar!

[Anonymous Coward]

I'm surprised nobody has mentioned Spacewar. I remember running it on a PDP-11 in 1978 or so, that was the very first video game I ever remember running and it was absolutely amazing at the time.

What's the PDP-11 OS here?

[unixisc]

These nuke plants - are they using their PDP-11s w/ Unix, or RSX-11 or something else?

Canonical architecture

[spaceyhackerlady]

The PDP-11 has always been one of those canonical architectures that everybody studies (or should, if they don't), even if they never actually see or use one. Very clean, very orthogonal.

I did an undergrad course in operating systems using XINU on an LSI-11. Great fun. I worked for DEC for a while in the early '90s, but only played with VAXen. It was the beginning of DEC's death spiral, so it wasn't a fun place to be.

I've sometimes thought it would be fun to own a real PDP-11, cool front panel and all. No idea what I'd do with it, but that's another matter. :-)

...laura

I can fix that

[bregmata]

HELLO [7,3]
PIP/DELE *.*;*
BYE

OpenVMS and PDP's relationship...

[ZahrGnosis]

"Not sure about the OpenVMS vs PDP comparison"... Since one is software and one is hardware, the confusion makes sense, but the comparison was basically valid:

The PDP-11 was 16-bit and gave way to the 32-bit VAX-11 (except, apparently, in nuclear power plants). The operating system developed to run the VAX-11 system was VAX-11/VMS (later just VMS). As the OS matured and Digital Equipment Corp (DEC)'s hardware moved to the Alpha CPUs, VMS matured along with it until HP eventually ported it to Itanium as OpenVMS.

While all of those hardware platforms ran multiple systems (early UNIX variants even ran on the PDP-11), VAX/VMS were tightly integrated to anyone that worked with them, and the PDP/VAX lineage was well established. In that way, comparing the lifecycle of PDP-11 to OpenVMS makes a lot of sense (and, honestly, I'm very surprised by which lasted longer!).

Two words:

[davidwr]

Job security.

For someone.

By the way, I still sometimes use a graphite writing stick.

Same with SOLAR16 in france

[lolop]

These systems, built by Telemecanique in 70's, are used in nuclear power-plants, and many are still active (more than 16000 were sold worldwide, in europe they sold similar quantity as PDP11 units).

As nuclear plant control software have been validated for these computers, devices control are interfaces with their bus, they are not easy (and very expensive) to replace (I've seen devices with memory + battery + DAT unit having same IO interface as old cartridge disks they replaced).
They have nice features considering the time they were built: failure detection and correction, high availability, realtime OS. See http://fr.wikipedia.org/wiki/Solar_(ordinateur) (sorry, found no page in english)

PDP11 - heaven!

[Anonymous Coward]

For my dissertation modeling work, I was able to use the PDP11 in another lab as a huge favor. I'd type all day, let it run all night (my allotted time slot) and collect the several inches of printout in the morning. Probably saved me a year of work on the ... wait for it ... SOL.

Apples and Oranges

[Anonymous Coward]

OpenVMS is an OS. The PDP-11 is hardware. There were (and are) any number of OSes that can run on the PDP-11. Some were "official," going back to DEC, others were developed later. There is, in fact, a market for LSI Boards that emulate the PDP-11 hardware in order to keep those old systems going. Look it up.

Coroutines...

[AliasMarlowe]

JSR PC,@(SP)+
We always put a comment at the top of such code saying "only experienced programmers should attempt to modify coroutines". We meant it...

Before the PDP-11 (I used at least the 23, 34, 44, 70, and 73 series, later the Pro series), I used a couple of PDP-8 machines. Now booting those 12-bit wonders was a real lark, involving magic incantations, thumps, and a load of toggled-in instructions. Once you got it going, the paper tape reader worked like a charm (of the malevolent variety).

Almost forgot the IBM-360 and its card-punch and programs spanning several boxes of cards. BTW, although I chase kids off my lawn, I'm not actually retired yet.

Comment removed

[account_deleted]

Comment removed based on user account deletion

re: PDP-11 assembler programmers are hard to find

[Anonymous Coward]

PDP-11 assembler programmers are insanely easy to find. Any competent programmer can pick it up in a few hours. A tight instruction set, with a really good orthogonal design for parameters (see https://en.wikipedia.org/wiki/PDP-11#Innovative_features).

Early hacker

[Anonymous Coward]

I used a PDP-11/34 from 78 to 82 in high school. The fun I had. I loved BSG so my first game was a space shooter. I did an anonymous sex questionnaire, then published the results. I wrote a program to take huge chunks of disk, rendering the system inoperative. I also wrote a login simulator that captured your credentials then logged you in. I wrote a program to do fractions and had to fight the math teacher for a grade. Ever done paper tape or JCL man?

Re:Early hacker

[Anonymous Coward]

Ever use the paper punch w/ 7-track tape?

Used the whole range from the /03 through the /77

[hughk]

For hardware I started with an 11/40, and then went through the range down and up. I got used to pulling and shuffling cards and even removing the wire-wrapped NPR jumper but the software was fun. For the software , we started with a monstrosity called DOS/BATCH, I went on to RSX-11M, RSX-11S (a paired down version of 11M for hard realtime), RT-11 and RSTS/E and ended up on RSX-11M-plus. The latter was a really cool multiprocessing capable O/S. The best thing is that for M and M-plus, they had to give you the kernel source so you could configure it (which was down to conditional assembly and lots of macros).

I was writing drivers and having fun. MACRO-11 was great. I used a heavily modded set of macros to provide a C like structure called SMAC and was using home-brewed structured exception handling by burying stuff in the stack frame, a bit like a VAX does. This gave me the ability to unwind quite gracefully. At one stage I managed to get hold of a copy of Unix, but we were commercial and it was hideously expensive at the time so couldn't use it for anything. The point being that a PDP-11 with EIS/CIS had a really nice instruction set and was easy to hold in the mind so I am fairly certain that a competent assembler programmer could write better code than most of the compilers. The instruction set was truly orthogonal so that all addressing modes worked whether it was real memory or registers.

Simpler is safer...

[3seas]

but for human sake, keep Microsoft out of there.

Re:Vax is probably alive and kicking someplaces to

[iggymanz]

not the same thing, as VMS was ported to Alpha and then Itanium. You can still buy machines from HP that run OpenVMS right now and it is still supported until Decenber 31, 2020 for version 8.4

Yes, I used to be a vaxcluster admin. best microcomputer / minicomputer /workstation OS ever.

Will it survive 2038?

[Anonymous Coward]

When the clock field overflows?

I'm sure there will be PDP-11s running then, and I will be at home, hiding in
my basement that day.

(And the authenticator asks me to translate "incoming" to post this message!)

Yeah, I'll definitely be hiding out in the basement.

Re:Will it survive 2038?

[SnarfQuest]

the 2038 overflow is mostly a Unix thing.

It really depends on what OS you are using. A lot of them don't even care about date or time. Others use a different storage format for time. And some of them failed when the date hit 2000.

SIMH

[Darinbob]

For those who are nostalgic, or curious, there is the SIMH family of PDP & VAX emulators that you can find quickly in a web search. Plus images taken from early Unix tapes, and some web sites with instructions, and you'll be able to get PDP-11 simulation up and running an early Unix version pretty quickly. For earlier PDP versions you can even get Unix version 6 without much effort (the version in the Lion's book), and it is quite interesting to see how much changed between it and V7 which before the whole Unix family tree exploded.

RT-11 SJ and Dungeo?

[bc17101]

I home built (wire wrapped) a PDP 11-05 with a RX01 compatible floppy disk drive back in the 70s and it still works great. I would like my children to see how computer games started out by showing them a working PDP11 running Dungeo (Zork). Since a number of you have clearly some interest in the PDP-11, I wondered if any of you have a RX01 compatible floppy diskette with RT-11 SJ still laying around that I could copy? If you have one with RT-11 and Dungeo that would be even better. Thanks, Bill

Re:RT-11 SJ and Dungeo?

[SnarfQuest]

Look on the simh support sites, and you might be able to find some images for simh. If there isn't one, you can try building your own from the install tapes under simh.

Once you have an image that works under simh, there are procedures that should allow you to transfer the simh image onto an actual floppy through a serial port.

There is a simh mail list that you can ask more specific information from. They usually don't mind talking about real hardware.

They must pay well

[Anonymous Coward]

... since these programmers are probably hard to find. How do I get a job like this?

Re:They must pay well

[RockDoctor]

... since these programmers are probably hard to find. How do I get a job like this?

• (1) Go back to the 1960s.
• (2) Pick a technology with a 100-year lifespan.
• (3) Specialise in it.
• (4) ...
• (5) Profit !

Or you could look into the future.

Re: If it ain't broke...then call in Unit 8200

[Anonymous Coward]

> And honestly, I trust a good PLC, say, Siemens (I have no experience with american brands), to be more reliable than a 40 year old PDP, no matter how well built it was. Siemens has been making automation controls for a LONG time and their products are really good, and I'll guess most of their bugs have been solved in all these years.

A particular benefit of Siemens Simatic Step7 is that it comes with the Stuxnet mil-malware included and will play hatikva.mid on your variable frequency drivers every Yom Kippur. Shalom!

Abstraction layers

[sjbe]

Not really if you think about it.

Oh I've thought about it. My background is as an engineer but I'm also an accountant. I *totally* understand why they did what they did and could probably give you a pretty good model of what the costs look like and the decision tree to make the decisions they did. My point is just that I'm a little surprised (emphasis on little) that they weren't a little more forward thinking.

When I see ancient equipment like this in use long after it is considered obsolete, what it typically tells me is that they didn't do a good enough job defining their interfaces. They spent their effort getting a working system and didn't make the layers of the system sufficiently abstract. Easier said than done I know but really they should be able to plug any computer that can programmed to handle the instruction set for the robot and the underlying architecture shouldn't matter. Likewise the robot shouldn't care that its instructions are coming from a PDP-11. They should have some form of well defined interface layer between the computer and the robot so either can be swapped out without affecting the other half of the system. Again, easy to say, not quite so easy to do. But that is how it ideally should be handled. Then you can swap the hardware or update it or test it and have it make economic sense. Also it permits safety features not otherwise available. You can have several controllers with different architectures and see if they agree (basically what they do on the space shuttle). If you can arrive at the same answer using multiple different methods then odds are good you have a safe system.

MSP430

[Smerta]

For those of you too young to know that you shouldn't be on my lawn...

If you're using (or have used) an MSP430 from Texas Instruments, you've used what is essentially a stripped-down version of the PDP-11.

My entire career has been spent doing embedded systems designs, and the little MSP430 is a great little chip. (Sometimes I wish it had been at the heart of the Arduino, but that's a different discussion!)

Even though most of my work nowadays is on the upper end of the ARM Cortex family, I still love me the occasional MSP430 (or other small embedded processor) design

Big difference in reliability

[Anonymous Coward]

The critical difference between an old PDP-11 and modern computers is that the former's transistor-level electronics will actually last long enough to deliver another 50 years while modern electronics has a far shorter lifetime exactly due to shrinking transistor dimensions. 50 years on a modern processor is laughable unless you never actually turn it on until the last 10 years.

Way back...

[Anonymous Coward]

I remember when the register board went bad in our PDP-11/70 and I had to prove it to the tech via a binary core dump...

RSX-11 and Version 7 UNIX, and BSD 4.2?

[bbsalem]

I wrote DEC-Fortran programs on RSX-11 on an 11/34 in about 1976. Later, I had access to an 11/70 which came shipped with IAS, we, at USGS in Menlo Park Ca, installed AT&T Version 7 UNIX on it in 1979. In fact an important person in the history of minicomputer and PC UNIX, Bill Jolitz, was the person who installed UNIX on our box. It might have been BSD 4.2 that he installed. I remember reading the UNIX articles from AT&T and Berkeley in binders.

Sic fonctionnatit, nil copulatum!

[RockDoctor]

Subject line says it all.

OK, in appallingly bad Dog-Latin.

"If it works, don't fuck with it!"

PDP 11's

[Anonymous Coward]

I started on an IBM 1130 with assembler. A PDP 11 running unix with the Berkeley upgrades would have been great.