Gordon Welchman Read online

  Welchman particularly enjoyed visiting Quex Park, a unique country park of 250 acres on the Isle of Thanet set within an estate of 1,800 acres around Birchington, Acol and Richborough. The property was an oasis of parkland and trees, all planted in the nineteenth century by John Powell-Powell and his successors. It was also a haven for birdlife with numerous species recorded there. In the 1950s it was the home of the Powell-Cotton family.7 Katharine’s mother and Mrs Powell-Cotton had been good friends for many years, and Katharine and one of her cousins often visited Quex in childhood. Welchman and Mrs Powell-Cotton became friends when she got him interested in dowsing. He would hopefully follow the pipes that conveyed water underground to an ornamental pond in the extensive garden, dowsing stick in hand. As with anything that interested him, he was relentless in pursuing the subject. This personal drive had served him well at BP.

  In 1957 Welchman and his family were on the move yet again, this time repeating the journey they had made more than ten years earlier. They arrived back in New York in May and took a two-year lease on an apartment at 15 East 75th Street. Always one to plan ahead, Welchman had secured a job as manager of a new Electronics Division in Ferranti Electric Inc., a New York subsidiary of the British Ferranti Company. His main task was to build up a manufacturing business in the US based on the achievements of the Ferranti research teams in the UK. He was also able to help develop a US market for British-made Ferranti products starting with a type of magnetostrictive delay line that had been developed for the Ferranti Pegasus computer.8 He subsequently met people at the US Navy, Army and Air Force research labs about possible applications in systems such as radar, sonar, IFF (an identification system designed for command and control) and multipath communications. Welchman’s meetings led to contracts for further development to meet specific requirements and, subsequently, quantity orders.

  Towards the end of 1958, an old MIT associate, Norman Taylor, invited Welchman to help him build up the electronics division of the ITEK Corporation based in Lexington, Massachusetts, where he was vicepresident. Welchman duly joined ITEK in 1959, taking over the management of an Applied Technology department that combined system studies with mechanical, optical and electronic engineering. His team was able to take advantage of ITEK’s considerable research activities in photography, optics, chemistry and information sciences. The core of the business was the storage, retrieval, reproduction and distribution of information recorded on photographic film, particularly documentary information. His department developed a number of new techniques for handling sheets of film for the Patent Office, the Library of Medicine and the USAF.

  The years of moving the family back and forth between the US and Britain had taken their toll on the Welchman marriage. Welchman’s long absences from home during the war had convinced Katharine that he was having an affair. The couple had been living separate lives for several years but had stayed together until their youngest child Ros reached a suitable age. In 1959 they were finally divorced. As a couple they had mixed socially in the artistic New York scene and both would soon find new partners.

  Later the same year Katharine married Francis Bitter, a physics professor at MIT. Bitter’s father, Karl, had been a noted sculptor. Welchman meanwhile had become friendly with Fannie Hillsmith, a fairly prominent New York artist and friend of Katharine. Fannie’s grandfather, Frank Hill Smith, had been a painter and one of the founders of the Boston Museum School, where she had trained for four years. In 1934 she moved to New York, became inspired by vanguard art, and developed an abstract style. Her work had been exhibited in a number of New York galleries. Welchman and Fannie Hillsmith were married on 4 August 1961. Fannie kept her New York apartment but the couple moved to a house that Welchman had purchased on Middle Street in Lexington, the same town he had lived in almost a decade earlier. While the house was a modest one, Fannie filled it with her painting. Welchman’s domain was the garden. He had read a book about the stone statues on Easter Island by Norwegian scientist Thor Heyerdahl and how the ancients were able to move such heavy megaliths. He used the same technique with the help of friends and family to place large stones around the garden. While undertaking this work friends and family would be entertained by classical music which constantly emanated from the house.

  By 1962, it became clear that ITEK management did not share either his or Taylor’s vision for the future of data storage. He therefore left the company to join the MITRE Corporation, based in Bedford, Massachusetts. MITRE had been founded four years earlier and one of its founding members was Bob Everett,9 Welchman’s former colleague on Project Whirlwind.

  As has been explained, MITRE, ERA and Project Whirlwind all had similar origins as examples of the military and private sector working together to deliver technological solutions to military problems. Around the time Welchman left to join ERA, the focus of the Whirlwind project changed and by September 1951 funding was available to support further computer development. The project was also moved from the Servomechanisms Laboratory to become a separate administrative unit under Forrester and Everett’s direction. Six months later the laboratory had become Division 6 of the newly established Lincoln Laboratory based at Hanscom Air Force Base near Bedford, Massachusetts.

  The Valley Committee, formally called the Air Defense System Engineering Committee (ADSEC) was chaired by MIT physics professor, George Valley. He had worked in MIT’s Radiation Laboratory during the Second World War and was greatly concerned about the lack of ‘effective co-ordination between the operational and technical personnel’ and the need for technical guidance in systems research. These were themes which would resonate with Welchman for the rest of his working life. The Air Force was also troubled by the inadequacy of continental air defence as well as a perceived lack within the Air Force of an appropriate research and development effort.

  Most British accounts of early computer developments have ignored Project Whirlwind. Welchman summed up Forrester and Everett’s achievement in a letter to Turing’s biographer, Andrew Hodges, on 25 June 1984:

  Almost all the early computer people had narrow views of mathematical and computational applications. But Forrester and Everett, with no knowledge of Turing’s ideas, were also fighting for the concept of a universal machine. And their objectives, from the outset, included real-time control. Their machines had to be fast enough for the purpose, and could not be allowed to fail. Other early computers were categorised by ‘Mean Time Between Failures’.

  Forrester and Project Whirlwind were unpopular with most of the computer community, which was mainly concerned with mathematical and scientific problems. He was fighting for a fast universal machine that would operate without failures, and this cost a lot of money. That he got away with it was due to the urgent need for an effective defence against nuclear air attack. His project led directly to the SAGE computers, built by IBM to his specifications. And he himself developed the first reliable fast-access storage – his magnetic core memory. This was a remarkable series of achievements.

  In April 1953, the US Air Force accepted a proposal from a small group which included Forrester, Everett and Valley for improvements to be made to the ground environment of the US Air Defense Command. In the summer of 1954 the proposed system was named the Semi-Automatic Ground Environment (SAGE). Division 6 had started design work on the Whirlwind II while IBM was commissioned to build a computer for the SAGE system.10 In 1956, Forrester left Lincoln to return to MIT as a Professor of Management and Everett became head of Division 6.

  Throughout 1957, the Air Force had been unable to find a contractor to take on the integration of the myriad of weapons and sensors into SAGE. It therefore decided to create a new, not-for-profit corporation that would become the Air Force’s centre of technical advice on SAGE and its future development. The core of the new company was Lincoln’s Division 6 and Bob Everett became Technical Director. The new company, called MITRE, had the daunting task of integrating a long-range surface-to-air missile system (Bomarc), a line-of-sight m
issile system (Nike), and the SAGE system. The first SAGE command centre became operational on 1 July 1958

  The work at MITRE represented the beginning of a permanent technical revolution in weapon systems. The US military became ever more dependent upon advanced technology and the engineers who implemented it successfully, much as the Germans had become dependent on the Enigma system and its operators. Who better for MITRE to hire than the man who had led the attack on Germany’s tactical communication system during the Second World War? Welchman joined MITRE in 1962 at the request of his former colleague from Project Whirlwind. Bob Everett wanted Welchman to contribute to tactical programmes in development and to explore some open-ended questions such as the application of digital technology to battlefield communications.

  Welchman had first-hand knowledge of Germany’s innovative use of radio communications to support battle plans. Blitzkrieg, the rapid, disruptive penetration of enemy lines, often achieved by racing around entrenched positions, depended on interconnected radio networks. Commanders in the field could call in air support and maintain constant contact with headquarters.

  Welchman’s first assignment at MITRE in 1962 was a small project related to USAF operations during a nuclear attack. As project leader, he supervised the formulation of a scheme, which was fully approved by NORAD command. The US was worried about the vulnerability of the headquarters of NORAD’s Commander-in-Chief (CINCNORD). The Canadians had already constructed an underground headquarters for the North-Eastern Region, for which they were responsible. A system was sought that would allow this facility to act as an Alternate Command Post (ALCOP) if CINCNORD became inoperative. MITRE was asked to establish Project SNOCAP (Survivable NORAD Capability) to look into the matter.

  In the same year, Welchman opted to become an American citizen. Having decided that his future now lay permanently in the USA, he felt that it was appropriate to pledge allegiance formally to his new country.

  In early 1964 he was transferred to a department studying air operations on a conventional battlefield outside Europe, specifically offensive missions against ground targets. He completed and published an analysis of future limited war environments in the summer of 1964 which became MITRE’s authoritative statement on limited war environments up to 1980. This was followed in August 1965 by a lengthy analysis of the characteristics of offensive air missions against ground targets in various possible limited war situations.

  During the following year he became involved in advanced planning for the overall command and control of tactical air operations and set out to deliver a system concept which would offer improved capability in a number of associated problem areas. From August 1966 to August 1967 his assigned objective was to identify areas of Air Force development that could lead to significant improvements in Joint task Force capabilities in the post-1980 time period. At the end of the year’s study, he had decided that tactical communication stood out as a problem of paramount importance.

  As Welchman wrote in The Hut 6 Story:

  The planning of battlefield communications gradually deteriorated into little more than methods of applying telephone-system thinking and switchboard technology to provide a rigid structure of point-to-point communications.

  Welchman believed that this mode of operation was long since obsolete in light of what modern communications technology could bring to the battlefield. In 1966, he had realized that ‘digital packets’ containing short bursts of information and broadcast over high-capacity common-user radios could provide attackers with a tremendous advantage. By 1968, his ideas had evolved into ‘a general-purpose battlefield communications system that could handle teletype, digital data, digitised voice and digitised pictures’. He illustrated the workings of this system with an inverted ‘U’ or horseshoe to represent an ‘information pipeline’ or bus.11 Eventually refined under an architecture known as ‘time division multiple access’ (TDMA), this bus was divided into small, sequential intervals called time slots. Aircraft and conceivably ground or naval forces would be assigned a discrete time slot in which to feed in position and status information in digital packets. This information would be updated at precisely defined intervals in a continuous cycle – a characteristic that led Welchman to describe the entire system as a ‘cyclic information system’. Welchman’s studies of military engagements across a wide sweep of history reaffirmed a problem labelled by the Prussian military strategist Clausewitz as the ‘fog of war’. MITRE’s Eric Ellingson attributed battlefield communications failures to people with information not knowing who needed it, people needing information not knowing who had it and there being no means for them to find each other.

  The proposed digital radio system appeared to rectify these problems because it relied on broadcast rather than point-to-point communications. Up-to-date information would be available to all subscribers at all times. Welchman labelled this concept, ‘selective access to information’ or SATI.

  The work coincided with other MITRE work on a transportable, long-range navigation system called Loran-D that was designed for tactical operations in Vietnam. This led to a demonstration system known as PLRACTA (Position Location, Reporting and Control of Tactical Aircraft). The system was demonstrated during the winter of 1970/71 and by the end of 1971, MITRE had shown its feasibility. However, the system’s success required the support of other branches of the service and it was resisted by pilots and aircraft programme managers.

  With retirement from MITRE looming on the horizon in June 1971, Welchman attempted to sell the concept in Washington. His 200-page document, M70-97, was prepared under US Air Force contract F19(628) 71-C-0365, Project 603C. A shortened version appeared in the February/March 1971 issues of Signal, the journal of the Armed Forces Communications and Electronics Association in the USA. In The Hut 6 Story, Welchman described the outcome of his attempts to win over senior military authorities in Washington to his proposals:

  The idea of command responsibility for communications was revolutionary, contrary to doctrine, and therefore unacceptable. And that was that!

  However, vindication came in 1972 when PLRACTA became part of a bigger Air Force programme called Seek Bus, which aimed at specifying prototype hardware and software for the installation, test and evaluation of the SATI concept. In 1974, the US Secretary of Defense, James Schlesinger, expanded the programme for use by all branches of the military, merging Seek Bus with a related Navy programme to form the Joint Tactical Information Distribution System (JTIDS). The US Department of Defense formed the Joint Service Program Office to administer the plan.

  In The Hut 6 Story, Welchman gave his final thoughts on these developments:

  The US Army is very interested in JTIDS not only for its own use, but also as a means of achieving more effective co-ordination of ground and air forces in combined arms operations. In view of all this, it is hard to understand why progress has been, and still is, so slow.

  Once again, Welchman’s vision was to be realized. After more than forty years of evolution, JTIDS is now a fully operational command and control system, providing information, distribution, position location and identification capabilities for the Air Force, Army, Navy, Marine Corps and British, French and NATO forces. Its elegant system design allowed for future expansion and adaption and has stood the test of time.12

  One former colleague who still works at MITRE recalled his first day there in 1968. Having joined the organization straight from college, he was given an office next to Welchman who was working on an air defence system. Welchman greeted him with a question: ‘What do we have for accurate time sources. What have you learned?’ He then took the young man under his wing and mentored him through his early days. They also became friends and their families would often spend time together. MITRE seems to have been designed to facilitate this type of interpersonal relationship with its interconnected buildings creating a campus-like environment. Another colleague attended a number of meetings with Welchman and was impressed with his grasp of the problem at h
and – the need for a vast data communications system. It was clear that Welchman was the man to deliver it and his contribution would be significant. Yet MITRE colleagues knew little if anything about Welchman’s wartime work at BP.13

  It wasn’t just his job with MITRE which was coming to an end. His marriage to Fannie was in terminal decline and had been for some time. At the beginning of July he began the painful process of telling his children that the marriage was over. Perhaps Fannie wasn’t cut out to be the sort of wife that Welchman thought she should be. She kept two houses, one for looking after her mother, and the couple were often apart pursuing their own careers. A psychologist advised them they should divorce as the Arts and Science didn’t mix! Given that Welchman’s two marriages, one to a musician and the other to an artist had failed, there may have been some truth in the advice, at least in Welchman’s case.

  After his marriage to Fannie ended, Welchman started to look for somewhere to spend the rest of his life in retirement. He travelled across the country and, after visiting a second cousin once removed in Westchester, Pennsylvania, he considered settling in Sarasota, Florida. On his way back to the home he had shared with Fannie in Lexington, he happened to pass through Newburyport, Massachusetts. Located on the southern bank of the Merrimack River where it reaches the Atlantic Ocean, the city had a long and rich history. The area was originally inhabited by the Pawtucket Indians. It was first settled by Europeans in 1635, prospered and became a city in 1851. Newburyport was best known as a seaport and as a shipbuilding centre in days gone by. Welchman was captivated by its New England charm. He drove through the city and continued on the causeway and drawbridge over the Plum Island River to Plum Island. He knew almost immediately that this was where he wanted to spend the rest of his days.