The Information War
How electronic computers emerged from the chaos of the Second World War
Konrad Zuse stands out as a remarkable individual in the history of computer science. He was the first to develop a true binary computer as well as, arguably, the first programming language. However, he suffered early in his career from being in the wrong place at the wrong time. Living in Nazi Germany during the Second World War, he worked practically alone with no access to the British and Americans who were pioneering the field. His fellow computer scientists around the world might have been hugely impressed by his computer, the Z3, had they known about it. The Nazi government however was not. They clearly had other priorities and so deemed his work “strategically unimportant”, denying him further funding to build an improved successor machine. Zuse continued work independently, having formed his own company, but before long the tide of war turned against Germany. The Nazis committed to a total war economy, effectively cutting off Zuse’s supplies of suitable parts needed to build his new Z4 machine. In the end, he was reduced to scrounging components from downed Allied aeroplanes or taking copper from the overhead cables of streetcars.
But just as the Nazis lost interest in Zuse’s work, the British realised that computers would play an essential role in actually winning the war. The same year Zuse was told to go on his merry way, Britain was the only major power left fighting Germany and suffering terribly for it. The island nation, dependent on imported goods to survive, faced starvation as thousands of tons of precious cargo were being sunk weekly by enemy submarines. It became of paramount importance to counter the U-boats, but it was almost impossible to fight back against a foe who could remain hidden until the very moment they struck. The British clearly needed to know ahead of time where the U-boats would be so they could prepare accordingly and remove the U-boats’ critical surprise advantage. They listened in on the radio messages used to coordinate the German U-boat fleet, but this did no good — the messages were transmitted in coded form. Hardly very sporting behaviour.
The encoding was done by the Enigma cipher machine, itself rather computer-like. Cipher machines allowed the sender to turn normal readable messages (plaintext) into meaningless gibberish (ciphertext) by substituting each letter for a different random one. The cipher-text message could then be transmitted in the knowledge that only people with their own cipher machine configured the exact same way could decode the cipher-text back into the original message. Eavesdroppers would listen in only to hear apparent nonsense. The German cipher machines used some especially cunning tricks, chief among which was that the encoding process was different for each character in the message. Even if you knew the first W in the encoded message decrypted to E, that still wouldn’t tell you which letter every other W decrypted to. This really was most unsporting.
The British set up a special unit at a place in England called Bletchley Park, devoted to breaking the German codes. Among the ranks of mathematicians and engineers there was Alan Turing, who had been drafted in to help. He and the other code-breakers there knew that the key to breaking the ciphers’ codes was to analyse the ciphertext for patterns which would betray the inner workings of the machine. These patterns would allow them to develop a mathematical algorithm that could zero in on the settings used by the cipher machine. The encrypted message could be fed into the algorithm and the plain-text message would be the result.
Although these patterns were eventually found, a fundamental problem still remained. The algorithms were incredibly complicated and took the code-breakers days or even weeks to carry out by hand, by which time the information was often out of date. Not surprisingly, Turing saw great potential in using computer technology in the fight against the ciphers, having (almost literally) written the book on computers and algorithms. Instead of doing the decoding work by hand, a specially programmed computer could execute the algorithms much faster.
As Turing’s team fought to decode the Enigma messages, another team at Bletchley led by Max Newman was busy trying to crack the Germans’ even stronger cipher machine, the Lorenz, which was used for the highest-level army communications. Newman was impressed by Turing’s arguments in favour of technology and had his team construct a machine to run the Lorenz decoding algorithms. The machine that resulted was affectionately known as the Heath Robinson, after the cartoonist known for his fantastically complicated contraptions. (Ironically, being only an early computer at Bletchley Park, the Heath Robinson was actually relatively simple-looking compared to what was yet to come.) Like Konrad Zuse’s machine, the Heath Robinson used relays thus making it a mechanical computer, but it was slow and unreliable. Clearly the German armed forces liked to talk a lot, because the number of messages needing decryption grew and grew, becoming more than Newman’s overworked machine could handle. The decoders at Bletchley Park were in danger of going back to square one: waiting so long for a message to be decoded that it was already out of date when finished. A faster replacement computer was needed to decode all these vital communications.
What Newman needed was something that acted like a switch, but did not operate on the principle of physical movement.
Speed is a key problem with mechanical relays, because their performance depends on physical movement. There comes a limit to just how fast you can keep them moving as an algorithm is executed, which puts upper bounds on the speed of a machine. What Newman needed was an alternative component to an electrical relay. It had to be something that acted like a switch, but did not operate on the principle of physical movement. In the 1940s, the best alternative was the vacuum tube, a device that was electronic rather than mechanical, but had a reputation even worse than the relay’s.
Turing recommended to Newman a Post Office engineer called Tommy Flowers, a man with plenty of experience when it came to vacuum tubes. He designed a new computer for Bletchley Park called Colossus, but many were sceptical of the proposal. Vacuum tubes are like little incandescent light bulbs, but have an extra electrode inside the glass casing in addition to the hot filament. When the vacuum tube is switched on, the filament gets hot, forcing electrons to shoot off its surface which in turn are attracted towards the electrode. This generates a current, just like a relay when it is switched on, only the switching process — dependent on the rather more rapid speed of electrons — is much, much quicker. But the speed was not what deterred the sceptics, it was the reliability. Like an incandescent light bulb, vacuum tubes have a tendency to burn out after a while, particularly when they are being switched rapidly on and off. Flowers proposed the Colossus computer would have no fewer than 1500 of them, far in excess of anything tried before. With this many, and knowing that a single burned out tube would cause the whole computer to break down, the conventional wisdom was that the Colossus would be unable to operate for any length of time to be useful.
But Flowers persisted. “There is a war on, you know.” He knew from experience that vacuum tubes were most likely to burn out when first switched on cold. By keeping the machine running twenty-four hours a day and the filaments permanently heated, the Colossus would perform much more reliably. This was demonstrated to be true when it finally went on-line in 1944 after a year in the making. From then on, the German armed forces could barely say a thing to each other without the Allies hearing every last word. The rest was history: the Germans were defeated and electronic computers proved their worth.
Computers have remained electronic ever since. Your computer is crammed with millions of electronic switches, but ones of a different nature. Transistors have long since replaced vacuum tubes and continue their remarkable trend to shrink in size, allowing us to pack more and more computing power into our computers as time goes on.