On this page
When you study logic gates in Class 12, an XOR gate looks like a small exercise in truth tables. In 1944 thousands of valves wired into gates like it were helping Bletchley Park break messages between Berlin and German army commanders. The man who designed that machine, Colossus, was not a university scientist but a telephone engineer from London's East End. This post covers who Tommy Flowers was, what Colossus did, a worked example of the logic it used, and why his name stayed hidden for about thirty years.
Who was Tommy Flowers?
Thomas Harold Flowers was born on 22 December 1905 and died on 28 October 1998, according to a biography by Thomas Haigh in the IEEE Annals of the History of Computing (2018). He grew up in the East End of London, where his father was a bricklayer. His mathematical ability won him a place at East Ham Technical School until 16, and he then began an apprenticeship at the Royal Arsenal in Woolwich while studying engineering in the evenings.
He joined the Post Office, which ran Britain's telephone network, in 1926. In 1930 he won a place at its new Post Office Research Station at Dollis Hill in north-west London. Haigh describes it as Britain's main research and development centre for telecommunications from the 1930s to the 1970s.
What was the Lorenz cipher?
Bletchley Park explains that the Lorenz cipher secured messages between Berlin and army commanders in the field. Messages were in teleprinter code, enciphered by the Lorenz machine. Codebreakers could break some by hand, but hand methods "could not keep up with the volume of intercepts". Under the mathematician Max Newman, a section called the Newmanry set out to build machines to speed this up.
What did Flowers build?
The National Museum of Computing (TNMOC), which now holds a working rebuild, records that Flowers spent eleven months designing and building Colossus at Dollis Hill. The first machine reached Bletchley Park in late December 1943 or January 1944 and was working in early February 1944. Cipher text was fed in on paper tape, and Colossus worked out the settings of the Lorenz machine's "chi" wheels. TNMOC calls it the first machine that was digital, programmable and electronic, though its programming was limited.
Tony Sale, who led the rebuild, notes in his account of the Colossus rebuild that the improved Mark 2 had 2,500 valves and ran about five times faster than the Mark 1. It first ran on 1 June 1944, days before D-Day. By the end of the war there were ten Colossus machines, and TNMOC says 63 million characters of high-grade German messages were decrypted by 550 people.
In his own 1983 paper on the design of Colossus, Flowers wrote that the machines had most of the essential features of a modern computer, except that programs were not held in memory. Instead the operators selected and connected hard-wired units with switches.
Worked example: the XOR logic inside Colossus
Sale describes the rebuild team wiring a logic panel of XOR gates and feeding its output to counters. You can see why XOR suits codes with a short example.
XOR gives 1 when its two inputs differ and 0 when they are the same: 0 ⊕ 0 = 0, 0 ⊕ 1 = 1, 1 ⊕ 0 = 1, 1 ⊕ 1 = 0. Teleprinter characters were sent as 5 bits each.
Take a sample 5-bit character 10110 and a sample key 01100, and combine them bit by bit:
Bit 1: 1 ⊕ 0 = 1
Bit 2: 0 ⊕ 1 = 1
Bit 3: 1 ⊕ 1 = 0
Bit 4: 1 ⊕ 0 = 1
Bit 5: 0 ⊕ 0 = 0
The result is 11010, which looks nothing like the original. Now combine 11010 with the same key 01100: you get 10110 again. XOR with the same key twice undoes itself, because any bit XOR itself is 0. That is why a sender and receiver with identical key streams can encipher and decipher with one operation, and why finding the key settings was the whole problem.
A common mistake in board exams is treating XOR like OR. With inputs 1 and 1, OR gives 1 but XOR gives 0. Write out all four rows of the truth table before simplifying any XOR expression. Our guide to semiconductors and logic gates has more practice. If the idea of a gate still feels abstract, the Fermi Concept Library covers every concept in physics, chemistry, biology, mathematics and computing, so you can revise the one idea that is blocking you.
Why did Flowers get so little credit?
Colossus was kept secret. TNMOC records that the first information about it was not declassified until 1975, and that Colossus was left out of histories of computing for decades. Sale writes that the original drawings of the mechanical assemblies were deliberately burnt in 1960. So the story of the computer age was told for years without one of its earliest electronic machines. In June 1996 Flowers attended when the partly rebuilt Colossus was formally switched on at Bletchley Park.
For the theory of computing that came before any of these machines, read our post on Alan Turing and computable numbers. For the valve itself, see John Ambrose Fleming and the thermionic valve.
What can a student learn from how Flowers worked?
Flowers brought practical knowledge from telephone exchanges to a problem that mathematicians had defined. His training came from an apprenticeship and evening classes, not an elite university. He built the first machine in under a year. And he did the work knowing he could not talk about it. Good engineering often means taking a tool you understand well and trusting it on a problem nobody has tried it on.
Try it now: XOR the sample character 11100 with the key 10101, then XOR your answer with the key again and check that you get 11100 back.
