Remember Control Data? The Living Computer Museum wants YOU

If you've got a bunch of old computer languages under your belt, the Living Computer Museum in Seattle, Washington, wants you. It's a job that's definitely not for the faint-hearted: as well as being able to handle old IBM, DEC, HP and Control Data Corporation languages, you'd be expected to help create and debug hardware …

[Mark 85]

They don't want much...

That is one hell of a wish list for an employee. They might be better off instead of one person who can do all of it to maybe finding few retirees who can do bits and pieces and possibly have some old software bits and pieces tucked away at home. Since that's not current mainstream stuff and hasn't been for awhile, that's probably what it will take. As the article points out, anyone proficient in PDP is probably better off (money wise and stress wise) in the nuke power industry.

[ecofeco]

Re: They don't want much...

This is the norm these days. For some reason today's moron managers don't know that things are done with teams, not individual supermen and high tech is STILL, high tech, which is arcane and complicated, and not charted accountancy.

[corestore]

Re: They don't want much...

Moron managers at LCM?

I happen to know some of the people there. One of their pdp-10s came from my collection. They're extremely high-calbre folks.

You do NOT know what you're talking about, and I suggest you pull your horns in rather sharpish.

[OldDawg]

Re: They don't want much...

Wow, I didn't know I was a superman. I'm sorry you can't accomplish anything without a support group. And see my post below.

[DaiyuHurst]

Re: They don't want much...

In fact they DO do things with teams.

[DaiyuHurst]

Re: They don't want much...

While most of the folks with that skill set are retired or near it, not all are.

[Anonymous Coward]

I'm a museum piece, myself

In my career I programmed a VAX 11/750, a Xerox Alto, a DECsystem 10, IBM PCs, and a CDC computer (batch only, and I lusted for the Cyber series machines), amongst others. But I'm done and dusted.

[Andrew Commons]

Re: I'm a museum piece, myself

The Cyber series were rather fun, there was a mean Startrek game on them (or at least the ones I got to play with) and an animated Snoopy that alternated between the circular main console displays. Inside they were a birds-nest of wires.

[pierce]

40 years ago I knew the IBM 1130 inside and out, down to the bits and bytes, IO devices and all.

Today? not so much so.

Would I want to dredge all that back up? not for any sane amount of money such a museum job could possibly pay.

[John Gamble]

Port That Job!

Heh. If I recall correctly, we used University of Minnesota languages and Purdue University operating systems on our CDCs. I have very fond memories of programming them in MNF (a Fortran compiler), but that was over forty years ago, and I'd probably need a refresher course on the tricks of programming with 60-bit registers.

Maybe we could fork Node.js to it.

(We need a "not entirely joking" icon now.)

[Andrew Commons]

Re: Port That Job!

Aaaaahhhh 60-bit registers on the Cybers.

Actually there were two sets of registers - A and X - which were 18 and 60 bits respectively (there may have been B as well which were also 18 bits, in fact I'm pretty sure of it). X for data and A for address...and there were 8 each of them. from memory A0 was just a dumb register, A1-A5 were paired with X1-X5 and loading a value into them (A1-A5) fetched that memory location into the corresponding X register. A6 and A7 were the reverse and loading a value into them wrote the contents of X6 or X7 to memory. This resulted in quite a bit of mental gymnastics to make sure that whatever you needed to get back into memory was in one of the two registers at the right time...which is where the use of a couple of (three?) XOR operations to swap values in registers was a lifesaver.

No hardware stack of course :-)

An interesting architecture.

[Tromos]

Re: Port That Job!

Quite right about the A and X registers. There were indeed also B registers of which B0 was actually a set of connections soldered to chassis earth! B0 gave a handy constant of zero and could not be changed. By convention B1 had the value 1, but this was not fixed by the hardware. Woe betide the programmer who forgot to kick off with a SB1 1 instruction.

There was a stack, but not in the conventional sense. It was an instruction stack and if you could code your loops so that they fitted in the stack, they ran a lot faster as no instruction fetches from memory were necessary. The MNF Fortran compiler mentioned earlier did a fair job, but it was the FTN compiler at its higher optimisation levels that did the best job of getting those inner loops slimmed down and in the stack. Took a while longer to compile, but those extra seconds could shave hours off a complex scientific task.

The other interesting part of the architecture was the peripheral processors. The CPU was completely incapable of any I/O, so this was down to the PPs. They had parallel access to the central memory, so a program requiring I/O would write a request to a certain memory location which was monitored by the PPs and they would then perform the required I/O to a central memory buffer pointed to by the request. The program could carry on running while waiting for the request, or could relinquish control to another task and be recalled after completion.

[John Gamble]

Re: Port That Job!

Ah, I'd forgotten about the load/store via the A registers aspect. Thanks for the reminder.

Huh, it occurs to me that I may actually still have a matrix multiplier that I wrote in CDC assembler all those decades ago (I was pleased enough with it that it survived the Great Home Paper Purge of the nineties). I need to look it up.

[Kurt 5]

Re: Port That Job!

Don't forget it was ones complement.

[Tromos]

Unrealistic job requirements

I can only really comment on the Control Data aspect of the job requirements but I think I can safely say that anyone who had knowledge of both central and peripheral processor assembly languages and architectures some 40-odd years ago will definitely not be looking for a museum job. There may be a handful prepared to do some part-time work, but if they are up to the CDC requirements, they will be seriously lacking in IBM/DEC/etc. To a CDC systems man (or even the extremely rare woman), the architectures and languages on these other systems are arcane and illogical and to be kept at bargepole distance. I'm sure the converse is also true.

What the job needs (if the budget will stretch to it) is a keen but inexperienced person and half a dozen part-time consultants to steer him/her in the right direction.

[John Gamble]

Re: Unrealistic job requirements

"...but if they are up to the CDC requirements, they will be seriously lacking in IBM/DEC/etc."

Seeing as my classmates and I worked on CDCs, PDP-11s, and Vaxen (that's the plural of VAX)... yeah, not so much.

Honestly, we're not so inflexible as that.

[DaiyuHurst]

Re: Unrealistic job requirements

Wanna bet?

[Evil Auditor]

"Control Data"

Reading this, I first thought of the album by Mark Stewart. I'm not that old after all.

[Tanuki]

Memories...

First serious computer I ever programmed (batch FORTRAN) was the UMRCC CDC7600, with an ICL1906 (under GEORGE3) as the RJE terminal. As others have mentioned, it used 60-bit words - meaning if you used BCD encoding you could store ten characters per word and then do some evil masking/bit-shuffling.

I developed a hate/hate relationship with the method the thing used to request tape-mounts: often it took rather longer for the ops to retrieve and mount a tape than the OS was prepared to wait before timing out and killing the requesting job. I got used to programming save/restore points throughout large runs.

[Andrew Commons]

Re: Memories...

"I got used to programming save/restore points throughout large runs."

I got used to bribes :-D

I wonder what was more effective? I also had to read several boxes (2000 per box I think) of cards in on the high speed card mangler regularly so being on good terms with the operators when working out which bits of compacted card had been read and which ones you had to try and resurrect via duplication was important.

The human element has really gone out of modern computing :-D

[Pascal Monett]

Re: The human element has really gone out of modern computing

No it hasn't, it has just be locked up in committee meetings filled with people who either know nothing about the project, the constraints or the requirements, or are not the ones taking the decisions.

Not exactly progress, eh ?

[Gordon 11]

I still have a printout of a FORTRAN program I ran on Imperial College and ULCC's CDCs in the '70s

60-bit word lengths...not a byte in sight.

[Tromos]

60-bit word lengths

15 or 30 bit instructions packed into these words, or 10 characters of 6 bits apiece. The peripheral processors used 12-bit words and had all of 4k of these to perform their tasks. Octal was the order of the day, hexadecimal has no place in a 6-12-15-30-60 bit world. And not forgetting 18-bit addresses of course.

[Mike 16]

Dancing Lines

I would love to get a copy/scan of that printout, even one page. All my printouts from the CDC501 seem to have gone missing, and I'd love to have a comparison for when "kids these days" wonder what was such a big deal with the 1403, which only profs and grads with funded research were allowed to use. (email to the printer model mentioned above at nulli.us, please. It's a nonce account for this purpose)

[another_vulture]

Too modern.

I have no experience with the newfangled 3rd gen CDC Cyber or IBM 1130. I do have experience with real computers: 2nd gen CDC 3800 and IBM 7040.

The 3800 was the supercomputer of its time. 48-bit words, discrete logic (one single flip-flop on a small module, so 48 modules per register.) Freon cooled.

[Mike 16]

Re: Too modern.

The 3800 came out a year after the 6600, and the same year(1965) as the IBM1130 and CDC6400. IIRC, both of the 6x00 (and more) were re-badged as "Cyber" a bit later. If we are going all Yorkshiremen, I've used two different computers with tubes/valves, not counting the bottles in the 6x00 console.. But I do recall the 6400 fondly.

[John Smith 19]

You need "A few good men from Univac"

(History of the CDC)

It's in my side pocket.

[OldDawg]

Gee, I didn't have any problems with it

To people commenting that the job requirements are 'unrealistic': I did that job successfully for over five years, until I left for other totally unrelated reasons. I restored DEC, IBM, Xerox, Burroughs/Univac and other equipment.

That's because I didn't learn IBM *or* DEC *or* [insert your favorite model here]. I learned electronics and programming and computer architecture.

And that so-called 'moron manager' is a good man, and I take exception to the ad hominem attack someone posted. I will refrain from commenting about some of the people to whom *he* reports. But I miss working with him.

[Morten Bjoernsvik]

My adventure into Cobol

I once worked on a project setting up batch jobs for data extraction from IBM mainframe (My project was a linux server working with that extracted data). My tools were an ancient EBCDIC terminal and large cobol programs that did extraction for other programs I had to copy and rewrite so it fitted my needs. And to test it I had to call a sysadmin with the program id and he would execute it. There were also another level, when I checked off the program as working. it was then run for some days on the test mainframe verifying it did not fuzz up actual work-loads and then moved to production queue on the production mainframe.

I've replaced some of them with 50 lines of very verbose perl. IBM provides an ancient perl(5.8.X) on their mainframes. It was quite fun to show some perl one-liners to the old cobol maestro. His comment: 'It is impossible to do a proper stacktrace on that, what if it fails". The perl code was never put on the production mainframe. They were afraid what another language would do to their production load. But it still resides on the test mainframe.

[Geschnutz]

Wish come true

Wow. I was just thinking two days ago that I ought to go work at a computer museum, doing almost exactly this kind of thing -- minus the hardware part, though, alas. Software, I could handle. Finding, writing, reverse-engineering, you name it. I'd have to learn these specific machines -- my old-iron experience is mostly VAX/VMS and early personal computers (i.e. pre-IBM-PC).

Too bad this one is in the UK; I'm in the US.

[DaiyuHurst]

Re: Wish come true

Actually, no, this job is in Seattle.