Gordon Welchman Page 17
In early 1949 MIT was awarded a contract to begin work on the project. Success would depend on the collaboration of engineers, mathematicians, physicists and air controllers. After nine months’ work, the team had published detailed descriptions of how to program, step by step, a computer of the Whirlwind design to control the positions of up to 100 aircraft more than ten miles away by means of a private-line communications system and a rotating antenna.3 A basic introduction to the available ‘machine language’ techniques was also provided. In January 1950, Welchman and his team were asked to consider air problems of a military nature that would be suitable for computer application. In other words, as Welchman noted in his routine bi-weekly report to project members: ‘We were asked to begin to look at the problem of tracking while scanning.’
To Forrester it was clear that they would soon be looking at problems of a greater military value. By the end of March 1950 a higher priority Air Force programme had pre-empted the Air Traffic Control Program. The goal would no longer be to control friendly aircraft, but instead to detect and, if necessary, destroy hostile aircraft. This didn’t change the mathematics or programming activities as both goals had similar requirements. In either case, for the first time, a digital computer, in this case the Whirlwind, would be connected to radar equipment. Radar pulses would be converted to telephone-line signals and sent to the computer for processing.
While Welchman was engaged at MIT he gave the first lecture courses on computing in the Department of Electrical Engineering. He did this in parallel with his work on Whirlwind which continued to be wide and varied. He was also a member of the Valley Committee and took part in the early studies that led to the creation of the Semi-Automatic Ground Environment (SAGE), an air-defence control system, for the US Air Force. During the early and middle 1950s, the Whirlwind computer became a fully operating machine but, by then, Welchman had moved on.
While Welchman was busy at MIT Katharine started to make friends and had ventured into dog breeding. She had located the owner of an American Pit Bull Terrier, with pedigree papers issued by the United Kennel Club, and mated the animal with Lucy, the bull terrier she had brought with the family from Britain. Unfortunately, the American Kennel Club was the more prestigious dog breeding organization and the resulting offspring were not deemed to be pedigree animals.
As an English family abroad, the Welchmans sought salt water beaches and they soon found several along the coast north and east of Boston, around Gloucester and Ipswich. In 1950, the family moved to Lexington, Massachusetts, about ten miles from MIT They bought a house at 26 Hancock Street for $12,000. Welchman had also found time to complete the book on algebraic geometry that he had been writing before the war. In 1950, Cambridge University Press published Introduction to Algebraic Geometry.
*
The following summer, the British government organized the Festival of Britain. It was a national exhibition with the purpose of giving Britons a feeling of recovery in the aftermath of war and to promote the British contribution to science, technology, industrial design, architecture and the arts. Welchman took his family back to England for much of the summer, travelling the northern route from Boston to Liverpool in a ship that carried cargo as well as passengers, stopping in Halifax, Nova Scotia, and St Johns, Newfoundland. Nick remembered walking with his father up to the headland above St Johns, site of the Marconi short-wave wireless station. Over a few weeks they managed to visit almost all their English relations, as well as the main Festival grounds in London. The family returned to America by air, a new experience for the children.
Back in Lexington, the children were all in good public schools and the girls were able to renew their friendship with two sisters whom they had known in Cambridge and who now lived in Lexington. Katharine started teaching music at Dana Hall School in Wellesley, another suburb of Boston. She also played the organ and directed the choir at the Unitarian church of Lexington. In preparation for the following summer, Welchman and his son began to learn to fish, with the help of the Unitarian minister of Lexington, who happened to hail from northern Vermont. The Reverend Floyd Taylor taught them the rudiments of fly fishing, and they all happily enjoyed their summer holidays in Vermont. The Reverend Taylor had been a very successful Fuller Brush salesman and was quite a potent speaker. His sermons were widely admired and published. Although Welchman was the first man in his family for some generations not to become a clergyman (apart from his brother who was killed in action in the First World War) and churchgoing was routine for him, he did find meaning in church ritual and was interested in discussing theological issues with clergymen. He had once been called upon to say grace at Cambridge in Latin because of his church background. He proceeded solemnly to recite several verses of Ovid, a Roman poet best known as the author of the three major collections of erotic poetry. Not surprisingly, his recital was greeted with stony silence. At one time, Welchman and his son would go to the Episcopal Church in Lexington and then rush over to the Unitarian Church where Katharine, Ros and Sue were working with the choir (Katharine was choir mistress and the girls sang). He was a devout reader of the Bible and claimed to have read it through from cover to cover three times.
*
In 1951, Welchman decided to leave MIT as he had had an offer to join Howard Engstrom at his pioneering new company Engineering Research Associates (ERA). The story of ERA is a fascinating one and perhaps best encapsulates Welchman’s views on opportunities in post-war Britain and the USA.
During the Second World War, a group functioning under the title of Communications Supplementary Activity – Washington (CSAW) had been part of the US Navy. Like a number of people at BP,4 they had been engaged in early computer developments for cryptanalytic purposes. The staff had a mix of skills familiar to BP: cryptanalysts, chess and bridge masters, mathematicians, physicists and engineers. Two key members of this group had been Howard Engstrom, a professor of mathematics from Yale and Bill Norris, a sales engineer from Westinghouse, the large electronics manufacturing company. As the war neared its end, Engstrom and Norris, who had become good friends, started to think about their return to civilian life. Their counterparts at BP had of course faced the same situation. The more radical recommendations made by Welchman and a few colleagues had been rejected. In the US, the Navy was keen to keep CSAW intact and offered the staff civil servant appointments. This was as unappealing to Engstrom and Norris as a return to their former employment, a sentiment shared by Welchman in the UK.
Engstrom suggested to Norris that they continue their CSAW work by setting up a private company. Since they had no capital or business experience, this seemed fanciful but, much to their surprise, they found that a number of key civilian members of CSAW were interested in their ideas. The Navy was also interested because, while keen to keep CSAW intact, it had found former technical staff members reluctant to join the civil service. Senior officers responsible for the work of CSAW recommended to the Secretary of the Navy, James Forrestal, that he make every effort to keep the group and its expertise together, even if it meant organizing parts of it outside of government. With the go-ahead from Forrestal, Engstrom and Norris began the process of creating their company in the autumn of 1945.
Suggestions that the new venture be run within a public sector body such as a university or research foundation were quickly rejected. While the Navy continued to offer civil service appointments, it unofficially endorsed a private venture, even though it could not officially help to set it up or guarantee that it would receive government business. By late 1945 funding still had not been secured and the venture risked losing its main asset, the skilled CSAW personnel. Finally John E. Parker, a wealthy entrepreneur, agreed to finance the new company. Parker had taken over a failed aircraft company in late 1941. He merged the company with Northwest Airlines to form the Northwestern Aeronautical Corporation (NAC). The company began building wooden gliders for the Army Air Corps. After the war ended, he was encouraged by high-ranking naval officers to fin
ance Engstrom and Norris’s venture and run it in parallel with NAC. In January 1946, Engineering Research Associates, Inc. was born, half owned by Engstrom, Norris and another colleague and half by an investor group headed by Parker.
Remarkably, in June 1946 the Navy Department issued contracts to both ERA and NAC without making them go through a competitive bidding process. By having the same management and headquarters, they overcame the problem of ERA as a new company not yet being eligible to receive government contracts. The US Government had in effect acted in an entrepreneurial manner to establish ERA and retain the services of over forty of CSAW’s most skilled staff.
ERA’s initial work was in the development of special digital machines and ongoing research and development on both data-handling and datastorage technology. Its researchers recognized that reliable data storage was essential for the evolving computer developments. In August 1947, ERA was funded to design a general purpose stored program computer. The design was approved in March 1948 and completed in the autumn of 1950. It was called ‘Atlas’5 and described by ERA as ‘a large scale 24-bit parallel magnetic drum, selective sequence calculator’. The machine contained 2,700 valves, 200 more than the Colossus Mark II which was operating at BP in June 1944, but without any memory. At the end of 1951, ERA announced the development of what it hoped would be its first commercial computer, the 1101. It was also starting to develop special purpose computers for customers with applications such as one of the first electronic inventory systems for a wholesale mail order house. Another system delivered in 1953 offered the storage, searching and processing of aircraft flight plans.
Welchman joined ERA in 1951, and became a personal assistant to Engstrom, who had become Vice-President for Planning. Welchman worked on computer applications for a wide range of systems such as wind tunnels, missile control, airborne fire control, war games and inventory control during the Christmas rush for the mail order house. After ERA had become a division of Remington Rand in 1952, he took part in joint planning activities covering both the ERA and the UNIVAC divisions. Remington Rand had already acquired the Eckert-Mauchly Computer Corporation, founded by the makers of the first American electronic computer ENIAC, in 1950. So by 1952 Remington Rand had become one of the biggest computer companies in the USA. It was acquired by the Sperry Corporation in 1955 to form a company then known as Sperry Rand, subsequently shortened to Sperry.6
Welchman had joined ERA’s Washington branch at the time that a triangular war had developed between ERA, Eckert-Mauchly, and Remington Rand’s own research laboratories, all with a wealth of punched-card systems experience. He worked hard to find out what each group had to offer and how their efforts could be co-ordinated for their common good. He was impressed with what the UNIVAC people were doing, but remained an ERA man until he left in 1954. It did mean, however, that he had to commute along a gruelling triangular track, going from Boston to Minneapolis/St Paul, then to Washington and back to Boston.
ERA did little original work except for the physical design of reliable equipment. Because of the combination of security and priority of the work to be performed by the systems after their installation in Washington, ERA was supplied with reports from almost every concurrent development effort everywhere. That this kind of technical espionage was taking place was a carefully guarded part of the whole ERA effort. It resulted in much copying of the better ideas from innumerable sources, and the benefits of ‘second guessing’.
The American military had shown itself to be quite forward-thinking as its Navy sought to preserve continuity in its development capability for cryptologic equipment and it encouraged the formation of ERA to deliver this continuity. In Britain, many of the brilliant engineers who had designed and built an electronic and programmable computer two years before the Americans would play no further part in the fledgling British computing industry which started to emerge at the end of the war.
In 1952, the Welchman family was on the move again, this time to Washington, DC. They bought a house at 3401 Porter Street in January for $27,000, which created financial problems as they had not been able to sell their house in Lexington before the move. It was not until June that they finally managed to sell it for $18,000, which at least yielded a 50 per cent profit on their original purchase. That summer the family decided to go on a road trip in their new Plymouth station wagon. They set off from Washington going south through Virginia to North Carolina, where they stopped for a while at Cape Hatteras. They then went west to the Blue Mountains, on to Kentucky, then up to Illinois and Wisconsin, where they reached the Sperry Rand offices. At this point Welchman left the family to do some work and then rejoined them in New England a few weeks later. Throughout the journey they pitched tents and camped out in state parks. Some of them were very scenic and had good facilities, but it placed a great strain on Katharine, particularly when she was left on her own with the children. They crossed into Canada and then back across the border into Vermont where Welchman rejoined them at the home of Floyd Taylor and his family. Finally, they returned home to Washington. Nick had come to his last year at Shady Hill, and he was planning to go to Philips Exeter Academy, a very good and reputable boarding school in Exeter, New Hampshire.
By 1954, Katharine Welchman had decided that she wanted to return to England. This was a common experience for a number of expatriate families, wherein they would find it difficult to enjoy living in America and would run out of patience with it after five years. However, the danger was that they would then go back to their old country, only to discover that it no longer suited them, either. Nick’s second year at Exeter had gone well and while he generally supported his mother’s wish to go back home, he was less fervent about it. Welchman came up to Exeter for a couple of days and stayed with Nick in his dormitory room. He lived in the school and met his son’s friends and teachers. They cycled together and took out a canoe on the river. Welchman was impressed by what he saw, and was interested to learn that Exeter was half a century older than Marlborough, his old school in England.
Welchman finally agreed to repatriate his family and find employment in England. In the spring of 1954 Welchman returned to Britain and, in April, he bought Milford House in the village of Chobham, Surrey for £3,285. The village was within each reach of London by train; Katharine had lived there when she was single and had a number of close friends in the village. Schools were soon found for the children. Sue was to go to St Mary’s School in Calne, Wiltshire, not far from Marlborough. Ros would go to Hanford, a little school near Blandford Forum run by an elderly friend of Welchman’s, and Nick to Welchman’s old alma mater, Marlborough.
Before the start of the new school terms, Katharine and the children went to a co-educational summer camp in New Mexico, a few miles north of Taos. The camp was run by Eleanor and Sandy Orr, school teachers in Washington, DC, and its campers and staff assembled and proceeded in loose convoys to the site of the camp over five days or so. It would prove to be a busy summer with much outdoor life and scenery. While the family was on holiday, Welchman took up a post as personal assistant to the Manager of the Computer Department of Ferranti Ltd. At the time, Ferranti was one of the pioneering manufacturers of electronic computers in Europe. Ferranti and Powers Samas Accounting Machines Ltd were trying to work out how they could combine their resources to cut down the lead time involved in getting a British commercial computer on the market. Welchman worked closely with the research engineers of both companies and led the liaison for applications studies and equipment specifications. From working with the Commercial Research Branch of Powers Samas and taking part in negotiations with its sales organization, he learned a good deal about the methodology of accounting and tabulating machines. His final task for Ferranti in the UK was to start a computing service for them in London.
The Welchman family never lived in Chobham. It seems that not only did Katharine not care for the house her husband had chosen for them, she no longer wished to live in England either. For years she had been scor
nful of the various American dialects and behaviours, although never in front of American friends. Now that same scorn was directed at hidebound English attitudes and stilted diction, a relic of the Victorian era in Britain. So, after landing at Heathrow and dropping the children off at their respective boarding schools, Katharine returned to New York. Welchman took up lodgings with his old BP colleague and friend Houston Wallace and his sister Hope, in Canonbury, north London.
Nick left Marlborough in the winter of 1955 and a room was found for him in South Kensington with a landlady who provided meals while Welchman continued to live with the Wallaces, in another part of London. Father and son later lived in a pied à terre flat just off the Embankment, and when Katharine finally came back to live with them, Nick moved to an upstairs room. By 1956, Welchman had rented a flat at 44 St Petersburgh Place, W2 to provide a base for the entire family when Sue and Ros were home from school. The family were also able to enjoy the pleasures of a small house, built at the end of a row of coastguard cottages at Westgate-on-Sea. Katharine’s mother had lived there for a while near the end of the war and for few years afterwards. Westgate was an unpretentious, old fashioned holiday place with many attractions for children, including long seaside walks. In the summers the family sometimes visited Dreamland, an amusement park in Margate, with a pier and a small fishing harbour. At Christmas time there were pantomimes in Margate and impressive services at Canterbury Cathedral. Westgate itself was the inspiration of several of John Betjeman’s poems about church bells. He had been at Marlborough at the same time as Welchman and both were commemorated by the school as distinguished old boys.