Secret English Team Develops Colossus

Date December, 1943

A secret team of British mathematicians, cryptanalysts, and electronic engineers developed Colossus, the first all-electronic calculating device, in response to the need to decipher German military codes during World War II.

Also known as Mark I; Mark II

Locale Dollis Hill and Bletchley Park, near London, England

Key Figures

• Thomas H. Flowers (1905-1998), Post Office electronics engineer who led the team that designed and built Colossus
• Max H. A. Newman (1897-1984), mathematician in charge of the Newmanry at Bletchley Park
• William Tutte (1917-2002), cryptanalyst at Bletchley Park who deciphered the Lorenz “Fish” code
• Alan Mathison Turing (1912-1954), mathematician and logician who conceived of a general computing machine, or “Turing machine”
• C. E. Wynn-Williams (fl. mid-twentieth century), engineer at the Telecommunications Research Establishment who designed the Heath Robinson, predecessor to Colossus

Summary of Event

In 1939, during World War II, a team of scientists, mathematicians, and engineers met at Bletchley Park, outside London, to discuss the development of machines that would break the secret code used in Nazi military communications. The Germans were using a machine called Enigma to communicate in code between headquarters and field units. The Enigma used a substitution code whereby a set of letters were substituted for the ones that normally made up the words. This in itself was not new; however, the Enigma enciphered only one letter and then shifted to a new position so that each letter of every word had a different key. The senders and receivers of the codes knew which rotor was being used for the substitution. The machine-made code used several rotors and so had vast substitution possibilities, which made the code extremely difficult to decode. The Enigma was portable, was easy to use, and generated seemingly unbreakable codes.

Polish scientists had earlier examined a German Enigma and had broken the codes used between 1928 and 1938 by employing electromechanical machines called Bombas. In 1938, however, the Germans made the Enigma more complicated, and the Poles were no longer able to break the codes. In 1939, the Polish machines and code-breaking knowledge passed to the British.

Alan Mathison Turing was one of the mathematicians gathered at Bletchley Park to work on code-breaking machines. Turing was one of the first people to conceive of the “universal” use of digital computers—that is, their use as general computing machines that could be adapted to more than one task. He first mentioned the “Turing machine” in an article, “On Computable Numbers,” published in May 28, 1936, in the Proceedings of the London Mathematical Society. The Turing machine was a hypothetical device for solving any problem dependent on mathematical computation and was not restricted to one task. Turing’s original and innovative contributions made him essential to the team working on code-breaking machines. Turing suggested an improvement to the Bletchley code-breaking machine (the Bombe, the British version of the Polish Bomba) which increased the computing power of the machine. The Bombe was an electromechanical relay machine that was similar to the Enigma. The Bombe did not decode messages itself but worked out the position of the Enigma rotors. Once the position of the Enigma rotors was known, the message could be decoded by specialists. The code-breaking machines replaced the tedious method of decoding by hand, which, in addition to being slow, was ineffective in dealing with very complicated encryptions that were changed daily.

The Bombe was useful until the Germans resorted to a cipher known as “Fish” code, produced by the Lorenz SZ40/42, a machine that was more sophisticated and less portable than the Enigma machines, which were used in the field. Fish codes were used for high-level communications and were binary (they used base 2, in which the only numerals are 0 and 1); decrypting the code was done by hand, based on the decryption worked out by William Tutte in 1941 after the Germans had violated their own rules by transmitted the same encrypted message twice with the same key settings. Max H. A. Newman, the mathematician who was placed in charge of the subunit at Bletchley Park responsible for decrypting the Fish (eventually known as the “Newmanry”), realized that hand decoding was far too slow to keep up with the Lorenz machine and believed that it was possible to design an automated device that could break these codes more quickly.

Working with C. E. Wynn-Williams of the Telecommunications Research Establishment (TRE) and engineers from the Post Office Research Station at Dollis Hill in northwest London, Newman oversaw production of the Heath Robinson machine, which could read two thousand characters per second. Operational by June, 1943, the Heath Robinson was a major advance but had problems with reliability, mainly because it was designed to compare two synchronized paper tapes, one with the encrypted message and one with the Lorenz machine’s encryption patterns. This mechanical action stretched the tapes, which soon were out of synch with each other when the machine ran at rates of more than one thousand characters per second, leading to unreliable readings. Moreover, there was a need to read characters more rapidly than the machine performed even at top speeds.

Tommy Flowers, an electrical engineer at Dollis Hill, suggested an idea that transformed the Heath Robinson into the Colossus: storing the information from one tape—the tape that held the Lorenz patterns—electronically. That would eliminate the need to synchronize two mechnical actions (two paper tapes running in synchronization). With only one paper tape traveling through the decoder, being compared with an electronic “tape” that held the Lorenz patterns, it would be possible to decode at rates much greater than that of the Heath Robinson.

Such a machine would require the use of what then was considered an enormous number of thermionic valves(vacuum tubes). Because these tubes were prone to breakdown and often needed to be replaced, skeptics questioned the idea. Flowers persisted, noting that the breakdowns occurred in the presence of power surges, when the valves were powered up. If the machine were left on at all times, the valves were unlikely to break down. Flowers succeeded in securing Newman’s support but failed to gain official support from Bletchley Park. He therefore developed Colossus by bypassing the Bletchley hierarch, gaining only the approval of Dollis Hill director Gordon Radley. Working with Sydney Broadhurst and William Chandler, Flowers designed and built the Colossus in an astoundingly brief eleven months at Dollis Hill. Although Turing and Wynn-Williams were not directly involved with the design of the Colossus, their previous work on the Heath Robinson was crucial, since the first Colossus was based on the Heath Robinson.

In December, 1943, the first Colossus, the Mark I, was assembled at Bletchley Park after being shipped there in parts from Dollis Hill; it was operational by January, 1944. It used fifteen hundred valves and could read five thousand characters per second from a paper tape. Faster speeds were possible, but at twice the speed the paper tape would disintegrate, so the five-thousand-character rate was determined to be acceptable for normal operations. Flowers had made arrangements for the manufacture of the time-consuming components in case other machines were required. The second generation, the Mark II, was in operation on June 1, 1944. It used twenty-four hundred valves and was extensively redesigned: It was capable of parallel operations (it carried out several different operations at once, instead of one at a time), and it had a short-term memory. Nine Mark II’s were constructed, and the original Mark I was converted to a Mark II. An eleventh machine was constructed at the end of the war.

After World War II ended in 1945, the secrecy of the Bletchley operations led Prime MinisterWinston Churchill to order the machines dismantled, although two of the Mark II’s were moved to the Government Communications Headquarters (GCHQ) at Eastcote, North London, and then Cheltenham. They were eventually dismantled as well. The approximately ten thousand men and women who worked at Bletchley—who had been sworn to secrecy and had worked under a “need to know” policy (being told only the information needed to complete an assigned task) had honored their commitment so effectively that few outside the project knew about the code-breaking work at Bletchley Park for more than thirty years after World War II ended.

Information began to emerge in 1976, when the term of Britain’s Official Secrets Act ended. In the early 1990’s, Bletchley’s museums director, Tony Sale, began to investigate the history of Colossus through some rare photographs, circuit diagrams, and interviews with Flowers and others; by 1996, he had succeeded in reconstructing a working replica of Colossus.

Significance

The Colossus machines gave Britain the best code breakers during World War II and provided information that was crucial for the Allied victory. Equally important, however, is their place in the history of computer science.

The Colossus machines were special-purpose, program-controlled electronic digital computers, the only known electronic programmable computers in existence in 1944. The later work of several of the people involved with the Bletchley Park projects was important in computer development after the war. Newman went to Manchester University shortly after the war. He was interested in the impact of computers on mathematics and received a grant from the Royal Society in 1946 to establish a laboratory for calculating machines at Manchester. Several other members of the Bletchley Park team joined Newman at Manchester, including Turing in 1948. Before going to Manchester University, however, Turing joined Britain’s National Physical Laboratory (NPL). At NPL, Turing worked on an advanced computer known as the Pilot Automatic Computing Engine (Pilot ACE). While at NPL, Turing proposed the concept of a stored program, which was a controversial but extremely important idea in computing. A “stored” program is in residence inside the computer and then a particular program and data are fed through an input device simultaneously. Turing was among the first to explain the stored program concept in print. He was also among the first to imagine the use of subroutines: smaller programs within a large program which pause the large program to conduct separate tasks and then return command to the large program. Turing also contributed to the computer output facilities, worked on the art of programming, and wrote the first Manchester programming manual.

Contrary to popular belief, the earliest binary, electronic, semiprogrammable computers were arguably the Colossi, not the ENIAC developed by John Presper Eckert and John William Mauchly—although the nearly simultaneous development of ENIAC, coupled with the secrecy under which the Colossi were cloaked, leaves open to debate the issue of which was “first.” Unlike the ENIAC, the Colossi used binary code; like ENIAC, they were partially programmable, by rewiring. The ENIAC, however, was “Turing complete”; that is, ENIAC was designed for general computing. What is beyond dispute, however, is that some of the most important figures at the vanguard of computing—Turing, Newman, Flowers, and others—were also part of the Bletchley experience and emerged from that experience to make key contributions to the development of computers and computer programs long after the war ended.

Bibliography

Anderson, David. Was the Manchester Baby Conceived at Bletchley Park? Report UoP-HC-2006-001. Portsmouth, England: University of Portsmouth, 2006. Uses primary source documents and sound clips to reexamine the history of computing in the light of key figures such as Turing, Newman, Flowers, and their connection with the innovations at Bletchley Park as well as later developments.

Coombs, Allen. “The Making of Colossus.” Annals of the History of Computing 5 (July, 1983): 253-259. Coombs provides an account of how the Colossus was built and put into operation. As one of the people who actually worked on the machine, Coombs offers an entertaining and well-informed report of the various problems and concerns involved with the operation of the Colossus.

Flowers, Thomas. “The Design of Colossus.” Annals of the History of Computing 5 (July, 1983): 239-252. Flowers, who designed the Colossus, describes how the machine evolved from the earlier Heath Robinson machine. Flowers provides background information on the Heath Robinson and the modifications that had to be made in order to produce the first Colossus. The article is well written and a valuable firsthand account.

Gannon, Paul. Colossus: Bletchley Park’s Greatest Secret. London: Atlantic, 2006. Journalist Gannon makes use of documents finally declassified after more than sixty years to tell the story of the code breakers: the cryptography, the military strategy, the significance of Colossus to the history of computing, and the personal heroism of the inventors and mathematicians.

Good, Irving J. “Pioneering Work on Computers at Bletchley.” In A History of Computing in the Twentieth Century, edited by N. Metropolis, J. Howlett, and Gian-Carlo Rota. New York: Academic Press, 1980. This chapter by one of the men who worked on Colossus provides interesting, firsthand observations about the development of the Heath Robinson and Colossus machines. The account is admittedly incomplete, since Good operated on the “need to know” basis and because some of the information was still classified.

Hodges, Andrew. Alan Turing: The Enigma. New York: Simon & Schuster, 1983. The Colossus is only one of several projects that Hodges covers in his focus on Alan Turing, his life, his profession, brilliance, and the times he lived in. This is definitely a book to read if one is interested in the history of computing, rather than only one project.

Randell, Brian. “The Colossus.” In A History of Computing in the Twentieth Century, edited by N. Metropolis, J. Howlett, and Gian-Carlo Rota. New York: Academic Press, 1980. Randell offers one of the most complete amounts of the Colossus and the work done at Bletchley Park. He has gathered information from numerous sources, including official government releases, interviews with those directly involved with the project, and material already in the public dominion. The result is a thoughtful, informative work. Illustrated, with extensive references.

Ritchie, David. The Computer Pioneers: The Making of the Modern Computers. New York: Simon & Schuster, 1986. Ritchie’s book is a well-written, entertaining account of the people involved with early computer developments. He relies on information from other published sources for the chapters covering the Bletchley Park project, but presents the information in a refreshing manner. Ritchie provides a brief but helpful summary of the machines discussed in his book, as well as a glossary of terms, notes on sources, and recommendations for further reading.

Watkins, Gwen. Cracking the Luftwaffe Codes: The Secrets of Bletchley Park. St. Paul, Minn.: Greenhill, 2006. An inside account of the day-to-day work at the German Air Section, the code-breaking staff at Bletchley Park, of which the author was a member and to which she affectionately refers as the “biggest lunatic asylum in Britain.”