Dr. Dobb's | Big Iron Age Man

Big Iron Age Man

When we heard that Dick Loop-Invariant, the final working example of the "old school" of mainframe programmers, was due to be switched off for the last time, we realized that it was imperative to send out Verity Stob to interview him...

August 12, 2000
URL:http://www.drdobbs.com/architecture-and-design/big-iron-age-man/184404376

Verity Stob: Dick, it's great to be with you today. Can you tell us something about the first machine you ever worked on?

Dick Loop-Invariant: Sure. It was in the summer of '46. I was a vacuum tube monkey for a machine called the EDSELIAC MK I. You must remember that things were very primitive at the start. It was many years before the computer industry discovered the, the... um...

VS: The transistor?

DL-I: No, the lower case letter. Everything was in capitals. The EDSELIAC, the AUTOTRAND, the COBLAMATIC 8, the VACUTRON VORTEX. We all had to be issued with aspirin just to be able to read the instructions for the PAPER TAPE READER. I myself was called DICK LOOP-INVARIANT until 1962, when I changed my name by converting myself to EBCIDIC.

VS: Amazing. What was programming like in those days?

DL-I: It was a much more physical, outdoorsy thing than it is now. The machine was big and unreliable, but there was always room for a bottle of beer in the air conditioning unit. I remember when the head programmer was having terrible trouble with a little subroutine to calculate the size of uranium rod required to fuel an atom-powered Buick. She got terribly agitated about it, and in the end became convinced it was a hardware problem. You may have heard of her. Her name was Grace Murray, Grace Murray... um...

VS: Grace Murray Hopper? Grandmother of COBOL?

DL-I: Grace Murray Mint, grandmother of Polo, Spear, Julep and Kevin. So eventually she asked me to climb up into the accumulator gallery and check all the valves. Sure enough, after poking about for a few hours I found a small moth had flown across the contacts and made a short. I gave the fried insect to Grace, and she taped it into her coding journal.

VS: That, I suppose, was how the term "de-insecting a program" came into popular use?

DL-I: Who is being interviewed here?

VS: Sorry. Moving on to the 1950s, then, I believe you were involved with the first commercial computers.

DL-I: Indeedy. I got a job with ChemCorp's scientific division, where the backroom boys were working on a less biodegradable version of DDT. We installed an early magnetic drum machine for them. It was a weird thing: a 16.7 bit address bus, 117 words of delay line memory, you had to toggle in the front panel program by hand before you could switch it on, and all the instruction codes were expressed in base 45.

VS: Despite these difficulties you were able to do useful work? I guess there is where you get to tell us about the miracles of algorithmic ingenuity you managed to pack into a miserable few dozen bytes?

DL-I: Don't be absurd. The machine was entirely unprogrammable. It couldn't even play a decent game of Nim.

VS: Oh. Wasn't this disastrous for you?

DL-I: Nobody seemed to mind very much. For one thing, you could keep a whole crate of beer cool in its air conditioning unit. And for another, it had wonderful tape drives, with huge tape reels that used to twitch round in a rather sinister way. Executives used to love to come and be filmed standing in front of them spouting nonsense about "electronic brains." Or simply watch them moving. Think kitten in front of washing machine. In those days it didn't much matter what the machine did, provided it Gave Good Tape. Just as a modern web site must Give Good Flash.

Talking of I/O and peripherals, I'm reminded of something amusing that happened later on. We used to boot up our System/360 using a whole tray of punched cards. Of course, every time we punched a card, tiny little bits of cardboard would accumulate in a tray beneath the hopper...

VS: I'm afraid I'm going to have to interrupt you there, Dick. I must warn you that I am instructed by the Editor to shoot you if you should mention the word "chad."

DL-I: But the whole anecdote depends on that. Can't I say it just the once?

VS: I'm sorry, but he was extremely clear on this point.

DL-I: As you wish. At about that time, in the 1960s, I was lucky enough to work with one of the most influential figures in software development, Frederick P, Frederick P, um...

VS: [Silence.]

DL-I: A little help?

VS: Oh very well. Frederick P Brooks, author of "The Mythical Man-Month"?

DL-I: No, Frederick P Coredump, author of "Twenty-five amusing things to do with an 8-inch floppy disk at an office party." Actually, our Fred was something of a critic of Brooks' work. He strongly felt that Brooks had skimped on the business of gathering experimental evidence, especially in respect of it supposedly not being possible for nine women, working simultaneously, to produce a baby in one month.

VS: Has Coredump been released from prison yet?

DL-I: I believe he is still paying his debt to society. Then came the 1970s, absolutely the best time to be in programming. Machines as big as a building, in some cases as big as a medium-sized village. Like when I was using a Kray-Twins Super Computer. I was working for the tobacco giant Fold, Spindle and Mutilate at the time, where they were researching a method of increasing the satisfying tarriness of tobacco smoke, to give it more body. They needed the most powerful number crunching facility available to model a cough. Over 27,000 separate vectors, you know.

This K-T was some computer. It ran so hot, they had to cool it by -

VS: - by pumping beer over the circuit boards?

DL-I: Don't be silly. We had to use pure alcohol, or vodka martini at the very least. Those were the days. People had to come and beg, on their knees, if they wanted us to change a program. Software engineers got the respect they deserved.

VS: And then, of course, the microprocessor arrived.

DL-I: The what?