Tag Archives: Colossus

The birth centenary of Alan Turing

Tomorrow is the birth centenary of Alan Turing the mathematician, logician, cryptanalyst and computer scientist who was highly influential in the development of computers and artificial intelligence.

A stamp from the Britons of Distinction issue, 23 February 2012. 1st Class – Alan Turing.

A stamp from the Britons of Distinction issue, 23 February 2012. 1st Class – Alan Turing.

Turing is perhaps most famous for his work during World War 2 at the code breaking centre in Bletchley Park. There he and others broke a number of German codes, including that of the Enigma machine.

At Bletchley Park Turing worked with a number of engineers seconded from the General Post Office’s engineering department, including Gordon Radley and Tommy Flowers. Radley and Flowers were both involved in the development of Colossus, the world’s first programmable electronic computer, which broke the Nazi’s Lorenz codes and convinced General Eisenhower to go ahead with D-Day. While Alan Turing was not directly involved in the development of Colossus his work fed in to the thinking behind it.

After the war Gordon Radley returned to the Post Office where he was involved in the development of the first transatlantic submarine cable, the invention the hearing aid, and projects to mechanise post sorting which led to the development of the postcode. He eventually rose to become Director General (Secretary to the Post Office), the first engineer to do so.

Tommy Flowers also returned to the Post Office after his time as a code breaker, where he was involved in developing the pioneering electronic telephone exchange at Highgate Wood, and ERNIE, the random number generator used by Premium Bonds.

Alan Turing’s post-war work and legacy are even more significant. Until his death in 1954 Turing undertook pioneering work in computer development and programming, mathematical biology and morphogenesis. He also developed the “Turing Test” for artificial intelligence, which states that a machine can only be said to be intelligent if its behaviours are indistinguishable from that of a human being.

A stamp from The Inventors' Tale issue, 12 January 1999. 63p – Computer inside Human Head (Alan Turing's work on computers).

A stamp from The Inventors’ Tale issue, 12 January 1999. 63p – Computer inside Human Head (Alan Turing’s work on computers).

For this and his many achievements Alan Turing is often labelled a “genius”. A stamp from 1999, part of The Inventor’s Tale issue, is testament to this: it features E Paolozzi’s artwork Computers, portraying a computer inside a human head. It is one of many of Paolozzi’s artworks inspired by Alan Turing.

A stamp released earlier this year (pictured above) as part of the Britons of Distinction issue commemorates Turing’s work as a mathematician, computer scientist and code breaker. The stamp shows Turing’s “Bombe” code breaking machine at Bletchley Park.

2012 is Alan Turing Year, celebrating the life and work of Alan Turing.

Designing the postcode: sorting machines, psychology and Sir Gordon Radley

In this series of blog posts, Peter Sutton, PhD candidate researching the post-war history of technology and industrial relations in the British Postal Service, offers some reflections on the history of the modern postcode. The significance of the postcode and its origins in the post-1945 era are considered followed by some archival examples tracing different aspects of its design along its journey from a specialised engineering concept to a universally recognised geographical referencing tool.

As mentioned in previous posts, the GPO’s design of the modern postcode became, after the Second World War, a systematic, long-term project which was part of wider attempts at designing sorting machines for a future national sorting system based on new technology. The vision outlined by leading engineers and postal experts in 1946, when sorting was predominantly done by hand at the sorting frame, was of a future system in which the national flow of mail would pass through large, mechanised sorting offices containing machines capable of “reading” each address. Their ultimate goal was to have human labour largely replaced within a generation by the new system, in which the public would cooperate by adding a code when addressing their letters to enable both interpretation and handling to be processed by machines.

As we have seen, as both postmen’s wages and the national volume of mail continued to rise over this period, there was good incentive to pursue these aims and to make what was termed the “code-sort concept” a reality. In the 1950s, the formative years of postal coding theory and practice, new lines of research pursued by GPO engineers helped lock particular characteristics into today’s familiar alpha-numeric postcode, as records held by the BPMA make clear. In particular the files of the Mechanical Aids Committee (MAC) and its subcommittees, which guided GPO policy on letter mechanisation and postcoding, show that choosing from many possible alternative code designs was no simple business.

A colour print of Mount Pleasant sorting office. The print is done in a comical style, portraying the 'un-pleasant' nature of the sorting office with its electronic sorting machine. Post, post bags and postmen are being flung through the office while two managers in suits stand in the middle. Outside is a post van with 'ER' on the side.

A colour print of Mount Pleasant sorting office. The print is done in a comical style, portraying the 'un-pleasant' nature of the sorting office with its electronic sorting machine. Post, post bags and postmen are being flung through the office while two managers in suits stand in the middle. Outside is a post van with 'ER' on the side. (Circa 1965)

One aspect of 1950s postal engineering involved trying to understand the psychological processes at work during the act of sorting letters when operating the new experimental machines then being trialled. This was an important step in both the development of the postcode and also in the GPO’s understanding of the so-called “man-machine interface”. Although at the outset it was hoped optical character recognition technology (OCR) might develop to the point of being able to reliably “read” handwritten postcodes, it was known early on that this was decades away at best. Machines for preparing and stacking the mail were then in development and, through an intricate arrangement of rollers, belts, diverters and memory storage, there were excellent prospects for an economical and reliable automatic sorting machine.

The problem was uniting these two phases of sorting – initial preparation of mail entering a sorting office into uniform piles followed by automatic sorting of that mail for particular destinations. In the absence of OCR, people stationed at keyboards (the man-machine interface) would still be needed to read the postcodes on letters and mark each with the correct machine-readable binary imprint. In any national code-sort system this would mean many millions of keystrokes every working day which placed a great financial premium on finding the most efficient arrangement of characters when designing the optimal postcode. A format acceptable to the public and major commercial mail users and also capable of accommodating the required address information was needed. But of no less importance was arriving at a format agreeable to the postman at his keyboard. An optimal format needed to encourage the most efficient possible coordination of hand, eye and brain.

Sir Gordon Radley

Sir Gordon Radley

A key figure in this regard was Sir Gordon Radley who took over the Mechanical Aids Committee (MAC) in 1955. This coincided with his appointment as Director General (Secretary to the Post Office), the first engineer to hold this office. Born in Birmingham in 1898, Radley gained his Doctorate in Engineering and served with the Royal Engineers during WWI. He joined the GPO in 1920 where he gained successive promotions through the Engineering Department. During WWII he and fellow telephone engineer Tommy Flowers made significant contributions to code-breaking at Bletchley Park and the design of “Colossus”. He spent ten years as Controller of Dollis Hill Research Station, the GPO centre for technological research, where he earned international fame for leading the development of the first transatlantic submarine cable and helping to invent the hearing aid.

When Radley chaired his first MAC meeting on 16 November 1955, he spoke of injecting more efficiency and pro-activity into the mechanisation field, issuing tighter deadlines and drafting in more senior staff. He oversaw the continuation of various long-term lines of technical enquiry and wished to see quicker progress in the development of postcodes, identifying the coding of letters by fluorescent marks as a priority project. Here, he encouraged Mr. A. Crisswell, a Deputy Regional Postal Director who headed the coding study group, to take a more liberal approach and worry less about the financial prospects of new ideas. Crisswell was directed to increase collaboration with operational experts in the Postal Services Department and to specify the desired characteristics of a code and to consult mathematicians specialising in coding theory.

Radley’s personal intervention into developing codes came as a series of questions posed in April 1956. A series of field trials with the Single Position Letter Sorting Machine (SPLSM) were then underway at Bath. There, postmen were trained to operate the machine’s keyboard to which letters were presented for processing, in which a memorised list of two-letter codes (denoting counties and towns) were added to the letters based on their addresses at between 32 and 52 items per minute. Their movements were filmed and then timed and analysed, prompting a series of questions on the MAC about how to deal with this data, with Radley leading the way. “Which kinds of codes are easiest to learn and use?” he asked. “What is the relationship between number of codes in a series and learning time?”, “What is the relationship between coding time and the number of possible codes to choose from?”, “How long does plural coding take relative to simple coding?” and “What is the best training method?”.

Single Position Letter Sorting Machine

Single Position Letter Sorting Machine (1956-1986)

Such precise questions were used to direct new lines of research on the MAC as they sought to evolve better and better machines for letter sorting, of which the coding position was a vital “link”. The Bath trials and the manner in which their results were put to use by GPO engineers is just one early episode in a long history of ergonomic design for coding desks and added to a body of research into the physical and mental demands made upon postmen who sorted the mail. Radley’s questions might also be seen as illustrating the diversity of factors at play in choosing the most suitable code for the British postal service.

Colossus and D-Day

65 years ago today General Dwight D. Eisenhower and his staff met to discuss the Normandy landings, or D-Day. The landings had been planned for some time and their success depended on good weather for the crossing and landing, and minimal resistance from German troops so that the Allies could gain a foothold.

Weather conditions had been too poor for a landing in early June 1944, but chief meteorologist James Martin Stagg forecast an improvement on 6th June. This weather forecast is usually cited as the deciding factor in Eisenhower’s decision to set D-Day for 6th June. However, Eisenhower is said to have received another piece of information during that meeting which was just as crucial, and he had the skill and inventiveness of the Post Office Research Station, Dollis Hill to thank for it.

Post Office engineers re-wire a telephone exchange after an air raid. Post Office telephone engineers developed the first programmable electronic computer during the 2nd World War.

Post Office engineers re-wire a telephone exchange after an air raid. Post Office telephone engineers developed the first programmable electronic computer during the 2nd World War.

Before the war Post Office engineer Tommy Flowers and his team at Dollis Hill had worked in switching electronics, exploring the possibilities for electronic telephone exchanges. But by the early 1940s they were helping the British code-breaking team at Bletchley Park. Colossus, later recognised as the world’s first programmable electronic computer, was their greatest achievement.

Colossus was primarily developed to decipher the Nazi Lorenz codes, high-level encryptions used by senior personnel, rather than the more famous Enigma codes used by field units. Computer technology was in its infancy in the 1940s and when in early 1943 Flowers proposed the machine, which would run on 1800 valves (vacuum tubes), there was great scepticism that it would work as until that point the most complicated electronic device had used about 150 valves.

But by December 1943 Colossus Mark 1 was working and it was soon moved to Bletchley Park, where it was able to break German codes within hours. An improved version, Colossus Mark 2, using 2400 valves, was unveiled on 1st June 1944, four days before Eisenhower made his decision about D-Day.

An essay by Flowers published in Colossus: The Secrets of Bletchley Park’s Code-breaking Computers describes the crucial meeting between General Eisenhower and his staff held on 5th June 1944. During that meeting a note summarising a recent Colossus decryption was handed to Eisenhower. It confirmed that Hitler was aware of troop build-ups in southern England, but would not be sending extra troops to Normandy as he was certain that Allied preparations were a hoax. This information was said to have convinced Eisenhower that the Normandy landings should take place the next day.

But whether it was the weather forecast or the Colossus decryption which tipped the balance in favour of 6th June, Flowers and the Post Office Research Station team made a remarkable advance in computer technology. By the end of the war 10 Colossus Mark 2 computers were in use at Bletchley Park, providing vital information to Allies forces, which certainly reduced the length of the war. After the war Flowers and his team returned to their work in switching, later pioneering all-electronic telephone exchanges. Their ingenuity was only recognised in the 1970s when restrictions on the Colossus project under the Official Secrets Act were lifted.