Who Invented the Computer? Tommy Flowers
Who Invented the Computer? This is the seventh installment in our ongoing series.
Scientific advancements are rarely the efforts of a single individual. More often than not, there are others anonymously working behind the scenes providing support, testing and applying tweaks that make the difference. Thomas Harold Flowers was one of these unsung characters without whom the British code-cracking effort of World War II would have been much less successful.
The son of a bricklayer, Flowers was born 1905 in a London working class neighborhood. Lacking the advantages of class and wealth to open career doors, young Tommy learned to rely on pluck and an innate talent for tinkering.
Flowers was fortunate to receive a scholarship permitting him to attend a technical college until age 16. Upon completion he apprenticed as a mechanic with Britain's official armaments manufacturer, the Royal Arsenal in Woolwich, England. His evenings were spent attending classes toward an engineering degree which he completed in 1926.
Flowers joined the Post Office as an electrical engineer and after four years transferred to their research division, where his primary focus was long distance signaling — in particular the problem of transmitting control signals to replace human operators with automatic switching equipment. He also became extremely knowledgeable about employing vacuum tubes to enable long-distance calls to be made without the intervention of an operator.
In early 1941, Alan Turing of Bletchley Park fame contacted Flowers' employer for help in constructing a deciphering machine. Flowers was sent over, but shortly after he arrived the decoder machine idea was abandoned. Impressed with Flowers' abilities, Turing introduced him to Max Newman, leader of a team tasked with solving the German military's newest encryption device, the Lorenz SZ machine.
The Lorenz SZ was used for direct communication between Hitler and his top generals. Its operation was similar to the Enigma Machine, but far more complicated and difficult to crack. Consisting of more cogs than Enigma, the device also utilized a 32-letter Baudot alphabet developed in the 1870s.
The Baudot alphabet added a level of complexity that vastly increased the number of possible cog settings. While Enigma could create 159 trillion possible number combinations, Lorenz's possibility for different combinations was estimated at 5,429,503,678,976 times greater.
Prior to Newman's arrival at Bletchley Park, code-crackers worked out possible cog settings with pencil and paper. In an effort to speed up the decryption process, Newman constructed a machine capable of processing 1,000 characters a second. In a process of elimination, the machine would run through all the possible cog settings until hitting upon the correct one for a particular message. Once the setting was determined, deciphering the message was easy.
While Newman's machine determined cog settings faster than humans, it was still slow enough that a large backlog of undeciphered messages built up. Another flaw with Newman's device was reliance on paper tape for processing data. In operation, tapes would rapidly pass over and around spiked turning wheels. Unfortunately, the tape frequently snapped in half causing the operators to go through a time-consuming process of turning the machine off, identifying the break, then splicing and rethreading the tape.
Flowers was asked to see if he could figure out a way for the machine to operate with fewer breakdowns. With the practical perspective of an engineer, he suggested using vacuum tubes instead of the electromechanical relay switches to process data. He argued that that vacuum tubes with photoelectric sensing could do operate much faster without a need to synchronize paper tapes.
Newman's team dismissed the idea, on the basis that vacuum tubes were notoriously unreliable. Their opinions were based on practical experience working with radio equipment that was often hauled around, roughly handled and constantly switched on and off. The consensus was that vacuum tubes would be unable to stand up to the strain of operation.
Flowers experience with vacuum tubes was entirely different. Working with telephone equipment that relied on vacuum tubes had taught him that if one avoided moving the tubes around frequently and never switched them off, they rarely broke down. Management disagreed and the idea was shelved.
Convinced that vacuum tubes were the solution, Flowers continued working on the project on his own time. After 10 months and more than �1,000 of his own money, Flowers unveiled the Colossus computer. Named because of its size, the computer contained an unheard of 1,500 vacuum tubes. Everyone but Flowers expected Colossus to fail — there were just too many vacuum tubes. To everyone's surprise, the tubes, as long as they remained on, rarely broke.
Colossus was an instant success. It processed 5,000 characters per second, enabling swifter decryption of coded messages — in many instances British codebreakers were able to read messages even before their intended German recipients.
Colossus was also programmable. While there was no operating system, pre-setting switches and sockets enabled the device to work on particular encryptions. Drawing on his practical engineering skills, Flowers created the world's first programmable, digital, electronic, computing device.
Mark II in Action
Newman was impressed and joined with Flowers to build an even more powerful machine, the Colossus Mark II, capable of processing 25,000 characters per second. Before war's end, Flowers would oversee construction of nine more Colossus computers.
The Mark II played an important role in the D-Day invasion. In a briefing with Allied Supreme Commander General Dwight D. Eisenhower, Flowers assured him that decoded German messages showed that reinforcements would not be near the planned landing beaches because Hitler was confident the Allies would land at a different site, further east.
Flowers also shared a decoded message from Field Marshall Erwin Rommel showing that one of the planned drop sites for U.S. paratroopers was actually a depot for a German tank division. Based on the information, the drop site was moved to a different location.
Upon the war's end, Winston Churchill, fearing that plans for Colossus might fall into Soviet hands, ordered the destruction of all 10 machines. Eight of the machines were dismantled and destroyed, including some that were actually dumped down out-of-service coal mines. The remaining two ended up with British Intelligence. Told to destroy all documentation relating to the Colossus computers, Flowers placed all the plans and drawings into a furnace and watched them burn up.
Flowers returned to the Post Office's research division and over time became chief engineer. British government officials granted him £1,000 for his work — it didn't cover his cost for the Colossus, and he shared the grant with staff members who helped build and test his machines.
Due to Bletchley Park's top-secret role, Flowers and others were prohibited from speaking of their wartime work. Their accomplishments were kept under wraps until 1977, when government ministers officially recognized their contributions to victory. Flowers was awarded an honorary degree by New Castle University in 1977. He died at his home in 1998.
Ironically, at one point after the war, Flowers had the idea of manufacturing a general purpose computer. He was turned down for a loan because bank officials 'didn't believe such a machine was possible.' For national security reasons, Flowers couldn't tell them that he'd already built 10 such machines.