The Mystery of the Lost Computer

You probably already know from Harry Harrison’s auto/biographical notes that during the Second World War he worked with analogue computers, and you’ve probably read elsewhere that he’s currently working on a book about the history of these machines. The fact that these types of computers are often overlooked when people write about the history of computing led HH to write the following article for the UK magazine Personal Computing World in the 1970s.

The Mystery of the Lost Computer
by Harry Harrison (1978)

I am beginning to feel that I once lived in another world, one of those parallel worlds so familiar to readers of science fiction. Perhaps because I write SF I am beginning to believe in one of my own unreal universes. Why even my good friend, the good Doctor Asimov, who knows everything about everything, doesn’t seem to know about my world. In his book Science Past – Science Future he discusses the computer revolution. He jumps in one mighty narrative bound from Aiken’s Mark I, 1937-1944, to ENIAC in 1946. With nothing in between.

Nothing? With quivering hands I picked up the glorious first issue of Personal Computer World and turned to ‘Past Procession’ with its compact boxed history labelled COMPUTER CLOCK. The truth revealed at fast. But unhappily for me the truth outlined there is the same one that Isaac Asimov revealed. Yet I put it to you that both of these learned gentlemen are wrong. There were computers in general use between those dates; hard-working, practical computers that did their job remarkably well.

I should know, since I was a computer mechanic during that period and serviced these machines.

What both Asimov and Kewney have ignored for the moment is the fact that there are more ways of running a computer than by electronics and digital means. What about Mr Babbage? With that clue in mind I challenge all you computer fanatics out there to figure out what kind of computer I am talking about. Cover the next lines with your hands and think – no cheating! Have you thought? Good, the answer is…

… a mechanical analogue computer of course. I have before me, stained and dog-eared, a United States War Department Hand Book on the operation and maintenance of The Sperry K-3 Automatic Computing Sight. Something that computes is a computer, right? And when I worked on these things my technical MO was Power Operated Turret and Computing Gunsight Specialist, Skilled.

Since I once saw one of these gunsights on sale after the war in a surplus shop, I hope I am breaking no laws by quoting from the hand book. (Dated Sept. 1943, I must add.)

“The sight computes the range and automatically compensates for ballistics and prediction … data input is transmitted to the computing mechanism … After the complete solution, the line of sight is offset the proper amounts from the line of bore, in lateral and vertical planes, to compensate in all deflections.”

That last bit, translated from the military technicalese, means the sight points one way, the guns the other. Like leading a bird with a shotgun. When the trigger is pulled the bullet rushes off to the spot where the plane is going to be, not where it is at the moment of firing as with a stationary target. (And let us not be too sure of that stationary bit, either. The Paris gun, with which the Germans shelled the city in World War I, fired a shell that travelled a high arc 92 miles long and took 3 minutes to reach its target. In order to hit the city the rotation of the Earth had to be considered in the computations.)

The only constant the K-3 worked with was the muzzle velocity of the guns and the resultant ballistic orbit. All of the other variables were either fed in by the navigator, the gunner or the instruments; altitude (therefore air pressure), wind velocity, distance to target, relative speed of target, and more. These variables were inserted by the strangest Heath-Robinson collection of gadgetry ever conceived by man. Gear trains, take-up springs, revolving plates with offset ball cages and – pride of the pack – three dimensional contour cams with cam followers. These were beautiful things, machined out of solid brass, and could be rotated about a central axis, or moved back and forth along that axis. They were the most educational device in the world because you could see the analogue actually at work.

But our computer problems weren’t the ones you have today. No glitches or static electricity problems to bother us. Backlash in the geartrains was what we had to worry about! Not that we didn’t have electricity in the K-3 – it was not powered by steam and leather belts as has been hinted – why, we had a constant speed 2,000 RPM motor. And a light bulb for the reticle sight.

The computer worked. These same principles were applied in that other analogue computer, the Norden Bomb Sight. (Someday I must tell you about the complicated device it used to release the bombs.) The B-17’s and B-24’s that carried power turrets with computing gunsights flew in stepped formation so any attacking pursuit plane would have to face the fire from a number of guns. The German fighters, for all their strengths, were very unstable gun platforms. They had open sights and their wing guns were zeroed at a point 200 yards in front of the plane; that is, the cone of fire from the guns met at this point. The maximum range where they might hit what they aimed at.

The K-3 was accurate from 1,000 yards on in.

So the aerial gunner had theoretically 800 yards to fire without being fired back at. Thirty shots a second from his two calibre .50 guns firing armour piercing ammunition. The sights worked and they did the job – we won the war, didn’t we?

So don’t let anyone knock mechanical analogue computers. If the airborne argument doesn’t suit you, why, just think of the Senior Service. All that beautifully accurate fire and great ranges, many times over the horizon of the guns firing, was controlled by ‘fruit machines.’

Read ‘mechanical computer’ for that last and you’ll have a good idea of where the computer action was in the forties. It wasn’t until 1945 that I saw my first electronic computer. This was the Central Fire Control system on the B-29, the plane that dropped the atomic bomb. All thermionic valves. I had no idea of how it worked, I would have had to re-enlist to find out and civvy street’s lure was too strong. I must find out some day – because I notice that the histories of computers don’t mention this one either.

© Harry Harrison, 1978

First published in Personal Computer World, Vol.1 No.2, 1978.