> ## Content Index > Fetch the complete content index at: https://huizhou92.com/llms.txt > Use this file to discover other available public pages before exploring further. # He Built Computers Two Years Before ENIAC — Then Had to Stand Beside It as a Polite Outsider - URL: https://huizhou92.com/tommy-flowers-colossus-secret-computer-en/ - Published: 2026-10-24T12:25:41.000Z - Updated: 2026-10-24T12:25:41.000Z - Description: Too useful to scrap, too secret to exist - Author: huizhou92 - Tags: #lang-en, Technology, History, Computing, Cryptography, World War II ### Too useful to scrap, too secret to exist The ENIAC is widely remembered as the world's first electronic digital computer. Yet in 1944, two full years before ENIAC powered on in Philadelphia, a fleet of room-sized digital computers built from thousands of thermionic valves was already humming around the clock in an underground bunker in Britain. The engineer who designed them was Tommy Flowers, a thirty-eight-year-old telephone switching specialist at the British General Post Office (GPO). Over the course of World War II, he built ten of these machines, codenamed **Colossus**. Running day and night continuously, they deciphered top-secret strategic teleprinter traffic between Adolf Hitler and the German High Command—confirming that Wehrmacht armor remained pinned down by the Allied deception at Pas-de-Calais, and securing the crucial operational window for the Normandy landings. ### 2,400 Valves on the Eve of D-Day In the early hours of June 1, 1944, the invasion of Normandy was just five days away. Inside an equipment room at Bletchley Park, the newly assembled Colossus 2—the first production machine—ground to a dead halt. The entire system comprised nearly 2,400 thermionic valves. During pulse-switching tests, stray capacitances and wiring inductances across the circuit stages triggered severe **parasitic oscillation**. The oscillation frequency shot completely off the scale of their field oscilloscopes, flipping the internal logic gates randomly and rendering the machine incapable of decoding a single character. The military deadline given to the Dollis Hill Post Office Research Station allowed zero leeway: the machine had to be operational by 08:00 on June 1\. Across the English Channel, hundreds of thousands of Allied soldiers were staging for the crossing, desperately waiting for decrypts of the German High Command's Lorenz SZ42 strategic teleprinter communications. Colossus 2 represented an aggressive architectural leap. Compared to Colossus 1, which had arrived months earlier, it added nearly a thousand valves and introduced a five-stage shift register for parallel processing. In effect, the 5,000-character-per-second optical paper tape reading speed was multiplied in the electronic logic stages to an equivalent of **25,000 characters per second**. But in 1944, no engineer on Earth had ever debugged a pulse digital system of this scale. W. W. Chandler, one of the Post Office engineers on site that night, later recalled how the team worked through the darkness, troubleshooting decoupling circuits and filtering stages to suppress the RF oscillations. At 08:00 sharp, Colossus 2 was formally switched into active wartime service. Bletchley Park's classified post-war summary, the *General Report on Tunny*, recorded the moment in a single dry entry: > Colossus 2 came into action on 1st June at 0800. Operating at five times the throughput of the prototype, it ran around the clock intercepting Wehrmacht communications. Days later, its decrypts confirmed the vital intelligence: Hitler and Field Marshal Rommel remained convinced that Allied landings would hit Pas-de-Calais, leaving the Normandy coastline without reinforcement. Yet even as these pioneering machines delivered decisive wartime breakthroughs, their fate had already been sealed by the British state: after the war, eight would be broken up for scrap and two locked away in secret intelligence vaults, while their designer, Tommy Flowers, was bound by the Official Secrets Act to thirty years of silence. ![Wartime Colossus at Bletchley Park](https://upload.wikimedia.org/wikipedia/commons/thumb/4/4b/Colossus.jpg/1280px-Colossus.jpg) *Wartime Colossus at Bletchley Park, operated by Dorothy Du Boisson (left) and Elsie Booker (right). Source: The National Archives (UK), FO 850/234, via Wikimedia Commons, Public Domain (PD-UKGov)* ### A 1,500-Valve Gamble The real story had begun sixteen months earlier. In early 1943, the thirty-seven-year-old Flowers led the switching group at the Dollis Hill research station. Born into a working-class family in Poplar, in London's East End, he started as an apprentice at the Woolwich Arsenal at sixteen. He earned his engineering degree taking evening classes at the University of London. After joining the General Post Office in 1926, he spent his career working on electrical switching in telephone networks. At the time, Bletchley Park attacked the Lorenz teleprinter cipher with an electromechanical contraption called "Heath Robinson." It relied on two perforated paper tapes spinning in synchronization at blistering speeds: one carrying intercepted ciphertext, the other generating cipher keys. The tapes rubbed violently against pulleys, tearing frequently, and mechanical read speed became the physical bottleneck of the codebreaking operation. ![Lorenz SZ42 cipher machine](https://upload.wikimedia.org/wikipedia/commons/4/4d/Lorenz-SZ42-2.jpg) *Lorenz SZ42 cipher machine on display at Bletchley Park. Source: Matt Crypto / Wikimedia Commons, Public Domain (PD-user)* Flowers proposed what seemed at the time like an absurd solution: discard the fragile key tape entirely, and emulate the German cipher machine's wheel logic and correlation algorithms purely in electronic valve circuits—using 1,500 vacuum tubes. In an era when radar units treated several dozen valves as complex systems, 1,500 tubes appeared to guarantee catastrophic failure rates. Gordon Welchman, who led Enigma codebreaking, protested in writing against "the reckless use of valves by Mr. Flowers," demanding his immediate removal from the project. Yet Flowers had counterintuitive engineering insight. Experimenting with automated telephone exchanges before the war, he had learned how valves actually behaved: nearly all valve failures occurred during power-on thermal shock, when surging currents shattered brittle filaments. If an array of valves was kept powered continuously, even hundreds or thousands operating in parallel could run reliably for months. As historian Brian Randell recorded from his 1975 interviews: > He knew from his prewar experience that most valve failures occurred when, or shortly after, power was switched on Mathematician Max Newman recognized the promise of this electronic approach and backed it against organizational skepticism. Flowers secured authorization from Dollis Hill director W. G. Radley, and with Winston Churchill issuing orders to provide "every facility in the way of staff and materials," Flowers pushed the build through to completion at Dollis Hill. On January 18, 1944, Colossus 1 was delivered to Bletchley Park. Codebreakers were initially skeptical of the sprawling, wire-covered machine. Within ten minutes, however, it produced results that electromechanical tape readers had struggled for days to compute. The *General Report on Tunny* recorded that the arrival of Colossus immediately increased intelligence output "to more than twice its previous level." The military constructed a dedicated brick building, Block H, to house subsequent machines. By the cessation of hostilities in May 1945, ten Colossus machines hummed continuously in Block H, with Mk 2 models scaling to 2,400 valves each. Flowers's continuous-power principle had passed its trial. ![Colossus 10 in Block H, 1945](https://upload.wikimedia.org/wikipedia/commons/0/08/Wartime_photo_of_Colossus_10.png) *Colossus 10 in Block H at Bletchley Park, 1945\. Source: General Report on Tunny, The National Archives (UK), HW 25/4–5, via Wikimedia Commons, Public Domain (PD-UKGov)* ### Decommissioning the Fleet **The moment the war ended, special-purpose machines designed for a single algorithm reached the natural conclusion of their lifecycle.** On June 3, 1945, Max Newman submitted a memorandum titled "Disposal of Machinery." Regarding Colossus, his verdict was concise and unsparing: > There will be no further use for Colossus, but some specimens will be wanted for further investigations. With the wartime cipher traffic gone, the dedicated machines had completed their mission. There was no theatrical drama of Churchill ordering them smashed into pieces "no bigger than a man's hand": eight of the ten machines were dismantled, with their counter chassis salvaged for new computing laboratories like Manchester University, while blueprints were ordered burned or impounded. Special-purpose hardware was dismantled when its problem disappeared—a matter of pure engineering logic. Yet the technology did not disappear entirely. It was too capable for the intelligence services to discard completely. The two surviving machines were relocated into deep operational secrecy. They followed British intelligence to Cheltenham (the new headquarters of GCHQ), where they were modified for general intelligence analysis in the early Cold War. GCHQ publicly acknowledged in 2024 that these two wartime computers remained in secret operational use well into the early 1960s. While those machines ran in the shadows, however, the outside world wrote a very different history. In September 1945, barely four months after the European armistice, Flowers traveled to the United States on a six-week tour, visiting the Moore School of Electrical Engineering at the University of Pennsylvania. There, he viewed the nearly completed ENIAC—which months later would be celebrated across the globe as the world's first electronic computer, boasting 18,000 vacuum tubes. It created one of the most poignant moments in the history of technology: Flowers had already engineered and deployed ten room-sized electronic computers, yet under the strictures of the Official Secrets Act, speaking a single word would be treason. He had to stand by like a polite outsider, listening in silence as American counterparts described their pioneering vision of high-speed electronic counting. After the war, Flowers was barred from working on electronic computing and returned to the Post Office as an ordinary engineer. Public history placed the crown on ENIAC, while the true pioneers were denied even a footnote in the textbooks. ![COLOSSUS part presented by Director GCHQ to Director NSA in 1986](https://upload.wikimedia.org/wikipedia/commons/thumb/4/4e/COLOSSUS%2C_part_of_the_machine%2C_presented_by_Director_GCHQ_to_Director_NSA_in_1986_-_National_Cryptologic_Museum_-_DSC07890.JPG/1280px-COLOSSUS%2C_part_of_the_machine%2C_presented_by_Director_GCHQ_to_Director_NSA_in_1986_-_National_Cryptologic_Museum_-_DSC07890.JPG) *A Colossus component presented by the Director of GCHQ to the Director of the NSA in 1986, now on display at the National Cryptologic Museum. Source: Daderot / Wikimedia Commons, CC0* ### Thirty-Two Years of Silence, Forty-Six Seconds of Irony In October 1975, the British government quietly released a set of captioned photographs at the Public Record Office (under file code FO 850/234). After thirty-two years of official silence, a crack had opened in the secrecy wall. Brian Randell, a computing historian at Newcastle University, followed the clues in those photographs, writing in his 1976 technical report: > In October 1975, after an official silence lasting thirty-two years, the British Government made a set of captioned photographs of COLOSSUS available at the Public Record Office Randell quickly tracked down Flowers. Approaching seventy, the engineer was finally permitted to discuss the vacuum tubes and paper tapes of his youth. But only in 2000, when the *General Report on Tunny* was declassified in full, did historians grasp the full extent of his engineering design. The British state's material reward for Flowers had been modest: an Award to Inventors of £1,000\. Colleague Allen Coombs later observed: "If it had been £10,000 or £100,000 it still wouldn't have been too much." His collaborator W. W. Chandler, placed on a junior civil service grade, received nothing at all. On June 6, 1996, a working replica of Colossus—painstakingly rebuilt by volunteers over years—switched on at Bletchley Park. An elderly, white-haired Flowers sat in the audience, looking at the towering metal racks reconstructed from just eight surviving photographs and scattered engineer notes. Two years later, he passed away peacefully at his London home at ninety-two. ![Tony Sale with the Colossus rebuild, 2006](https://upload.wikimedia.org/wikipedia/commons/thumb/e/e0/ColossusRebuild_11.jpg/1280px-ColossusRebuild_11.jpg) *Tony Sale with the rebuilt Colossus in 2006\. Source: MaltaGC / Wikimedia Commons, CC BY-SA 3.0* During a 2007 Cipher Challenge organized by the National Museum of Computing, the rebuilt Colossus took three and a half hours to break an authentic wartime intercept; a German programmer running custom software on an ordinary home laptop completed the task in forty-six seconds. Historians continue to debate whether Colossus qualifies as the "first computer." But the reflection Flowers left in his obituary remains the purest summary of his work: > At the time I had no thought or knowledge of computers in the modern sense. He had no idea he was building modern computing. He knew only that urgent wartime traffic needed deciphering before dawn, with hundreds of thousands of lives waiting on the output.