By John Vardalas
Copyright © 1994 IEEE . Reprinted from IEEE Annals
of the History of Computing, Vol. 16, No.2, 1994
This material is posted here with permission of the IEEE. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by sending a blank email message to firstname.lastname@example.org. By choosing to view this document, you agree to all provisions of the copyright laws protecting it.
Ferranti Canada's efforts to launch an indigenous computer industry, with itself at the centre, and the eventual collapse of this dream, is a more complex story than the usual versions which speak of great ideas lost to an unresponsive society. The present paper gives an account of Ferranti Canada's involvement in the beginnings of a Canadian electronics industry in the twenty years, or so, following the end of the Second World War against the background of the influence of world events on the Canadian political, military, and industrial scene.
The Ferranti-Packard firm arose from the 1958 merger of Ferranti Electric Limited and Packard Electric Limited. The first company, a wholly owned subsidiary of the British firm Ferranti Limited, was first established in Canada in 1912 and the second, a wholly owned Canadian company, was incorporated by the American Packard brothers in St. Catharines, Ontario, in 1894. Manufacturing electrical power equipment has been their traditional business.1 Throughout this paper, Ferranti Canada will be used to refer to the Canadian subsidiary, whether it be Ferranti Electric or Ferranti-Packard. The term Ferranti U.K. will be used when referring to the British parent firm.
In a 1961 attempt to carve out a profitable niche in a market dominated by IBM, Ferranti Canada set out to develop an innovative mid-sized general purpose computer. The resulting computer system, called the FP-6000, was the culmination of a corporate and technological process that had started in 1949 with a military project called DATAR. Though a subsidiary of Ferranti U.K. the Canadian company steered a remarkably independent technological and corporate course from its British parent firm. The FP-6000 was Canada's first attempt to establish an indigenous computer industry. Designing and building an innovative computer, however, turned out to be easier than marketing it.
In 1963, developments in the United Kingdom abruptly ended Ferranti Canada's fourteen year quest to turn electronic digital computer technology into a profitable business. In an effort to make the British computer industry more competitive in the world market, the British government encouraged consolidation. One important consequence of this policy was International Computers and Tabulator's (ICT) acquisition in 1965 of Ferranti U.K.'s non-military computer operations.2 The transfer of the Canadian FP-6000 design rights to ICT was the central element to this deal. In ICT's hands, the commercially uncertain FP-6000 became the technological take off point for its highly successful ICT 1900 series of computers.
Peace-Time Canadian Military Enterprise and National Technological Development
The memory of Canada's unpreparedness on the eve of World War II still haunted the minds of Canada's military leaders as the nation planned for postwar reconstruction. For example, in 1936 Canada did not possess a single antiaircraft gun, the armaments guarding the Pacific and Atlantic entrances to Canada were defective, and there were no bombs for the twenty-five outdated aircraft comprising Canada's Air Force.3 "Canada," wrote the Chiefs of Staff in 1945, "had not had to pay the awful price of defeat, but still had to pay dearly for lack of preparedness."4 The chiefs of the armed services argued that Canada needed to retain a sizeable peace-time defence force. However peace-time military preparedness demanded far more than retaining high levels of recruitment. Important economic, industrial and technological issues had to be addressed.
On the 10th of February 1947, Lieutenant General Charles Foulkes, the Chief of Staff of the Canadian armed forces, wrote a lengthy confidential memorandum to Brooke Claxton, the Minister of National Defence, in which he argued that the nation's future military preparedness required the existence of a strong national industrial and technological capacity.5 Recalling that it took four years before Canada's industrial mobilization was felt in the active theatres of World War II, Foulkes emphasized to the Minister that "it is highly unlikely that we shall ever again have as long a period as we had before the last war to mobilize the resources of the nation."6
The Chiefs of Staff suggested four principles on which to establish this program. The first asserted that the centrality of science and technology in modern warfare required the integration of technical knowledge into the highest levels military thinking. The second asserted that the financial and professional status of technical experts within the military needed to be raised to prevent the peacetime attrition of its scientific manpower, both military and civilian. The third called for a more coordinated scientific and technological exchange of information between the armed services. The fourth principle called for a new partnership between academic, industrial and, military research personnel.
Unlike the United States and the United Kingdom, Canada had neither the financial resources, industrial infrastructure, nor the scale of military operations to pursue research and development in all facets of modern warfare. To ensure access to the latest weapons technology of its allies, while still promoting a critical mass of indigenous research and development, the General Staff adopted a very Canadian strategy: the pursuit of multilateral technology sharing arrangements with its allies; the quest for comparative technological advantages; the promotion of transnational standardization; and the centralization of Canadian military-oriented research and development.6 One important institutional outcome of this strategy was the creation, by an act of Parliament in 1947, of the Defence Research Board (DRB). This organization was charged with responsibility of ensuring that Canada develop a research and development program consistent with the above strategy. The value of promoting technology within the military was given such importance that DRB was created as a fourth Service within the Department of National Defence. The Chairman of DRB became an equal member in the Chiefs of the General Staff.
Ferranti Canada Enters The Digital Electronic Information Age: DATAR
One of the most pressing strategic issues facing the newly established DRB was the need to encourage more peacetime industrial electronics research. In August 1948 Captain E.G. Cullwick, Director of DRB's Electrical Research Division, tried to solicit industry participation in a military electronics research and development program.7
Ferranti Canada Limited in Toronto was one of the companies to receive Cullwick's letter. The company had no experience in electronics, but its parent firm, Ferranti U.K., had considerable expertise in this area. A copy of the letter was forwarded to Vincent Ziani de Ferranti who was the President, Chairman of the Board and the principle share holder of Ferranti U.K.
De Ferranti's senior managers and engineers convinced him to encourage DRB to expand it's research requirements to include digital electronics and electronic digital computers. Eric Grundy, a member of the Board of Directors of Ferranti U.K., reminded de Ferranti that the "Admiralty Flyplane", which the company had designed and built, embodied electronic computing. He also said that Dr. D.G. Prinz, an eminent research scientist on staff at Ferranti U.K., was studying methods of high-speed data transmission.8 Even more importantly, Ferranti U.K. had just received an estimated £35,000 per annum five-year contract from the British government to do the necessary design and production engineering to turn the Manchester University Mark I computer into a commercial venture.9
However DRB had stipulated that "the value of such work to the problems of defence will of course depend on the degree to which the programme can be carried out with Canadian facilities and personnel."10 Unfortunately the Canadian subsidiary had neither the financial, managerial, nor technical resources to undertake such research. While the parent firm was willing to offer the necessary technology transfer, it was unwilling to underwrite the capital and operating costs.
The meeting on October 18, 1948 between senior officials from DRB, Ferranti Canada and Ferranti U.K. proved to be a big disappointment for de Ferranti. He expected to get assured contracts in return for his offer to set up a Canadian digital electronics team in Toronto. Instead, Dr. Oman Solandt, the chairman of DRB, explained to de Ferranti that DRB had not "yet reached the stage where we could undertake to support a research venture such as the one which you propose which has no direct commercial value but is entirely dependent on government support."11 Ferranti Canada's response to the position of DRB was prudent: "Solandt leaves a small loophole for us to be associated with the Board, provided we will work for nothing", wrote A.B. Cooper, the president of the Canadian subsidiary, to de Ferranti.12 However, Cooper added that it was perhaps in the company's "general interest not to ignore Solandt's suggestion of unpaid co-operation."12
|Part of the 30,000-vacuum-tube DATAR system crammed into the below-decks sections of a Canadian minesweeper (Ferranti-Packard Transformers Ltd.)|
Word of de Ferranti's failed meeting with the Defence Research Board soon reached Lieutenant Jim Belyea, a research officer in the Royal Canadian Navy's electrical laboratories in Ottawa. Belyea had a bold idea of developing a comprehensive naval electronic information processing and communications system, which later became known as DATAR (Digital Automated Tracking and Resolving). The movement of large convoys across the Atlantic during World War II presented monumental logistical and tactical challenges. During storms and under the cover of night, keeping track of relative positions was extremely difficult for the ships of the convoy. Furthermore, when the convoy came under attack, it was impossible to have precise real-time data of enemy movement that could be shared simultaneously by all the ships. Belyea was convinced of the vital need to develop a system that would allow all the ships in a convoy to gather, process, share, and display real-time radar and sonar data. Such a system, for Belyea, would not only have to display the movement of friendly and enemy aircraft, ships, and submarines, but it would also have to present this data to each ship relative to each of their reference frames. Belyea's idea lacked any technical structure; however he was convinced from his experience with naval training simulators that digital rather than analog methods were necessary. A digital paradigm was necessary, but how was the question. In Belyea's mind, the Ferranti group held the answer.
On the basis of a funding scheme proposed by Belyea, de Ferranti agreed to set up the nucleus of a digital electronics R&D group in its Canadian subsidiary in Toronto.[see note 1] Kenyon Taylor, the man who first directed the electronics group in Canada, reflected:
It seemed to our group that what [Belyea] had in mind was very much the proper thing to be doing ... It was a first step in push-button warfare. Lt. Belyea was thinking 15 years ahead of his time and Sir Vincent de Ferranti and the rest of our party were well in tune with him.1
By 1950, the small Canadian team at Ferranti Canada had demonstrated that radar tracking data could be reliably encoded in digital form and communicated over long distances by pulse coded modulation (PCM) radio. But it was the heating up of the Cold War that provided the added incentive to commit substantial resources to DATAR. During the Korean War, Canada launched a massive rearmament program. Naval appropriations, for example, rose from approximately $73 million in 1949-50 to over $332 million in 1953-54.13 With 100 new ships ordered for the RCN in February of 19513, DATAR, if successful, promised to make the Canadian Navy the most advanced in the world from the point of view of electronic information processing.
The RCN spent over $1,900,000 to move DATAR from a technological possibility to a realistic prototype.14 One RCN briefing explained:
DATAR is being designed and developed as an integrated system ... to provide the Command with a complete, concise and up-to-date picture of the tactical and strategical situations, to provide the necessary information to Weapon Control systems to assist in Target Designation, and to direct aircraft and, in some cases, ships.15
Though these tasks could be carried out by any ship, control and coordination was centralized in a special-purpose, hard-wired, electronic digital computer equipped with a high-speed memory drum. The data flow between each ship and the memory drum was carried out automatically and in real-time by PCM radio transmission. Equally important, the information processing took the relative movement of the ships into account. As a result each ship in the DATAR system had a complete picture of everything within the radar and sonar detection range of any every other vessel in the convoy.
Through the Fall of 1953, the DATAR prototype was demonstrated, on Lake Ontario, using a convoy of three ships, two Canadian Bangor class minesweepers and a shore station which played the role of the third ship. The surface and air radar data was real while the sonar data was simulated from a location on shore. The unqualified success of these demonstrations prompted RCN Commodore Lay to bring DATAR to the attention of the members of Permanent Joint Board on Defence (PJBD)[see note 2]. He explained to the PJBD:
...that while certain research and development work on somewhat similar systems has been carried out by both [Britain's] Royal Navy and the United States Navy, it is believed that the Royal Canadian Navy's DATAR system with a capacity of 500 targets is more flexible and has progressed further in its development.16
The failure to sell the Americans and British on DATAR underscored some of the contradictions the Canadian military faced in trying to advance Canada's technological and industrial capacity, albeit in specialized areas. The costs associated with the design, prototype development, production engineering, and deployment of very advanced large-scale defence systems were more than Canadian procurement alone could support. With Canada's relatively small peacetime Armed Forces, foreign sales were essential to underwrite any military flirtation with technological and industrial self-reliance. [see note 3]
The need to give Canadian industry access to the military markets of its allies was one of the driving forces behind Canada's push for transnational standardization - first among its British and American allies, and later within NATO. Canada's pursuit of component standardization and interchangeability across the entire alliance was a reasonable response for a small nation whose historic position has been to sit on the frontier of two empires: Great Britain and the United States. The standardization of screw threads, for example, was a remarkable Canadian-led initiative to ensure, among other things, that Canada's manufacturers could supply both the British and American military forces, without the wasteful duplication of machine tool setups. [see note 4]
|The prototype trackball used in the DATAR system. The device, invented by Tom Cranston and Fred Longstaff sometime in early 1952, used a bowling ball from the Canadian game of five-pin bowling (smaller than the American 10-pin ball). The air bearings were developed by Kenyon Taylor of Ferranti-Packard (Ferranti-Packard Transformers Ltd.)|
In addition to political factors, the vacuum tube also contributed to a general reluctance to proceed further with DATAR. The sheer size and power requirements of large-scale circuits made it increasingly more difficult to fit electronic equipment onto ships and airplanes, for example, the electronics in a destroyer contained some 3,000 tubes while the navigational equipment in a large bomber used 2,500.18 The problem of reliability was even more troubling. As the number of tubes in military electronics grew into the thousands the mean time between failure became dangerously short. In a survey of 1050 U.S. Navy fleet and shore station installations, it was found that one-third of the equipment was not operating properly, due to breakdowns--the U.S. airborne AN/APQ-24 radar bombing and navigation systems experienced, on the average, one failure per 21.5 hour interval.18 In a memorandum to the Chief of the Canadian Air Staff, RCAF Wing Commander W.P. Likeness wrote:
...the rapid rise in the failure rate of electronic equipment coincident with the increasing complexity and miniaturization of that equipment is a cause for alarm ... The perfection of components would reduce failures in a ratio of only two to one whereas reduction in a ratio of fifty to one is required for acceptable reliability."19
With over 30,000 tubes in the DATAR, Ferranti Canada had a heightened awareness of the reliability and design problems associated with the vacuum tube paradigm. During the many DATAR tests and demonstrations it was not unusual to find young Ferranti Canada engineers, armed with cartridge belts of vacuum tubes, racing through the ship's interior replacing faulty tubes.
The Collapse of Military Support and the Search for a Commercial Niche
Rather than a wholesale transfer from the parent firm, the professional staff of the Ferranti Canada electronics group was drawn almost exclusively from recent Canadian university graduates. Thus DATAR allowed a talented but inexperienced young group of engineers to mature into one of Canada's leading industrial research and development centers in digital information processing. The collapse of DATAR, however, suddenly left Ferranti Canada with an electronics team with few contracts to support their ambitions. In a 1954 effort to get funding the group's manager, Kenyon Taylor, tried to convince the Navy and DRB to fund an R&D program to see if DATAR could be reliably implemented in solid state. From the Navy and DRB's perspective, however, there was little merit in pursuing Ferranti Canada's proposal. After the Korean War, the political will to support growing military expenditures had weakened. Furthermore, in 1954, the military was still unsure of the merits of transistors in military electronics [see note 5].
The DATAR experience had convinced the Ferranti Canada digital electronics group that it could build computers as a business. There had already been some talk, in 1951, about the possibility of Ferranti Canada commercializing the University of Toronto's experimental UTEC computer.1 The collapse of the UTEC project, which ironically was provoked, in 1952, by the University of Toronto's purchase of a Ferranti U.K. computer, quickly ended this prospect. However, this same sale had convinced Dr. Arthur Porter, who came from Ferranti U.K. to manage the DATAR project, that the Canadian subsidiary should be in the computer business, even if it meant competing with the parent firm.
In the Summer of 1952, buoyed by the rapid progress being made on DATAR, Porter told the vice chairman of DRB, E.L. Davies, that the Canadian subsidiary "could produce a computing machine as efficient if not better than the present Ferranti equipment [Ferranti Mark I], in approximately twelve months for roughly $150,000."1 Davies offered Porter an opportunity to make good on his statement:
I am suggesting a method by which Ferranti Canada could effectively and cleanly cut the throat of Ferranti England...This depends on whether or not your statement at the very pleasant lunch last Thursday was affected by the liquid refreshments! ... We do not have in Canada, at present, any need for a further computer but we have heard that Dr. Ellis Johnson, Director of Operations Research [in Washington, D.C.] ... is considering purchasing one of the Ferranti machines for roughly $300,000. This is your chance to go to it and earn some U.S. dollars for Canada, our contribution being the know-how we paid for in your development of DATAR.1
It is not known whether Porter followed up Davies's suggestion at all. The fact is that Ferranti Canada never did build this computer.
The year 1956 marked the first important step in Ferranti Canada's perilous journey to create a business in computer technology. In that year, Ferranti Canada designed and built the first computerized mail sorting system for the Canadian post-office [see note 6]. The computer was a hard-wired system in which postal code locations were kept in a look-up table stored on a high speed drum. The speed of the memory drum was important in the real-time response of the system and Ferranti Canada had patented an air-suspension system that allowed its large memory drums to rotate at extremely high speeds. The bar-coded postal code information was optically scanned by a system that was also built by Ferranti Canada. The experimental model, which was fully transistorized, could sort over 36,000 letters an hour. The mail sorting system received international attention: "The U.S. sent a group of congressmen to check on the Canadian system before congress voted five million dollars to fund a research laboratory in Washington D.C. for a similar study."20 The first computerized mail sorting system in the U.S. went into operation four years later in 1960 in Rhode Island; Great Britain followed in 1966.21
With the Canadian government's plan to build much larger and faster sorting systems, the business prospects for the Ferranti Canada computer group looked bright. However, the business fell apart on the rocks of political controversy. Calling the proposed full scale computerized mail sorting system a "million dollar monster", the Conservative opposition attacked the Liberal government's spending habits. With the defeat of the Liberals in 1957, any hopes Ferranti Canada had of getting a contract to expand the prototype into a full scale system vanished. As with DATAR, the collapse of the computerized mail sorting project left the company's electronic digital engineering group with an innovation and no buyers.
In 1959, Ferranti Canada won a pivotal contract from TransCanada Airways to build Canada's first computerized airline reservation system. The resulting system, RESERVEC I, was a real-time, on-line system that allowed ticket agents from anywhere in the country to query a central computer, located in Toronto, about the availability of seats and make reservations. With RESERVEC, Ferranti Canada had finally managed to obtain a contract to build a full-scale information processing system. Ferranti Canada no doubt made money on the $3.25 million RESERVEC contract. However, the company was still faced with the challenge of developing a business around this new technology. Canada had only one national airline. IBM's SABRE system dominated the American market. Rather than promote the proven technology of its Canadian subsidiary, Ferranti U.K. wanted to develop and sell its own computerized airline reservation system to British Airways. Ferranti Canada's failure to get Ferranti U.K.'s co-operation, reflected the parent firm's inability to formulate a coherent corporate wide strategy. Ironically, it was this same inability that permitted the Canadian subsidiary to pursue its own ambitions.
|The "double plot" that displayed all the tracking information from the DATAR computer (note the trackball at the front of the control console) (Ferranti-Packard Transformers Ltd.)|
Though unable to sell any turn-key computer reservation systems, RESERVEC's operational success intensified the feeling within the electronics group that Ferranti Canada should get into the business of building and selling general purpose computers. Technically, the Gemini computer used in RESERVEC offered a sound starting point from which to design a general purpose machine. The logic circuits were proven and, unlike DATAR and the post-office computer, RESERVEC's architecture was considerably more complex and programmable [see the article by Dornian in this issue]. A new sense of urgency emerged, within the electronics group. However, there was a growing feeling of exasperation that upper management had no real interest in building a computer business. Faced with this increasing discontent and impatience, the manager of the company's Electronics Division assured the computer group that he was working on a plan to manufacture the parent firm's newly developed ATLAS computer.
With no indigenous computer manufacturing facilities in Canada, the federal government was receptive to Ferranti Canada's request for financial assistance to set up production facilities. All that remained was to get the approval of Ferranti U.K. and arrange for the necessary technology transfer. Immersed in trying to save its own badly managed computer operations, Ferranti U.K. had little interest and rejected the ATLAS proposal. Angered at the news and the absence of any support from its own upper management, many of the senior computer engineers tended their resignations. After a long meeting, the CEO and chairman of the board of Ferranti Canada, Tom Edmondson, managed to convince them to reconsider their resignations and persevere a little longer.
From Cheque Sorting to the FP-6000
The technical success of the Ferranti Canada's computerized mail sorting system had caught the attention of officials in New York's Federal Reserve Bank. The cheque sorting system installed in 1958 in the New York offices of the Federal Reserve Bank was a direct transfer of the technology used for the mail sorter. Instead of an optically scanned bar-coded postal code, the cheque sorter scanned MICR (magnetic ink character recognition) encoded cheques. While previous cheque sorting techniques worked one digit at a time, Ferranti Canada used their hard-wired computer to scan all eight coding digits and then sort the cheque by reference to a look-up table stored in memory.
Ferranti Canada's design approach was to make all aspects of the computer cheque sorting system modular, in order that customers could expand the capacity of their sorting systems in an economical and flexible manner. A fully expanded cheque sorting system could process up to 290,000 cheques per hour. By 1961 however, technical advances had started to undermine Ferranti Canada's special-purpose computer approach to cheque sorting. As the price-to-performance ratio of general-purpose computers dropped, they became a competitive alternative to special-purpose computers, even in very specialized applications. Furthermore, programmable computers allowed cheque sorting operations to be expanded to include a wider range of data processing tasks as circumstances warranted. Honeywell's efforts to introduce general-purpose computers into the cheque sorting operations of banks made Ferranti Canada realize how vulnerable their business plan was.
In the Fall of 1961, armed with the recent success of RESERVEC, one of Ferranti Canada's system engineers, Paul Dixon, convinced the Federal Reserve Bank to expand its data processing operations by purchasing a general purpose computer from Ferranti Canada. At long last, the company was in the commercial computer business. This new computer would be called the FP-6000.
The problems of its parent firm offered an important object lesson to the Canadian team in the pitfalls to avoid when designing the FP-6000. Although Ferranti U.K. produced high-powered computers for scientific applications, the company had difficulties addressing the various levels of the burgeoning electronic data processing and office automation market. Ferranti U.K. was unable to find the right technical and marketing response to this still untapped market. Preoccupied with technical sophistication, the company lost sight of where the market was heading. This error was reflected in the rising losses Ferranti U.K. suffered in its non-military computer operations from 1958 onward--throughout the period from 1956 to 1964, Ferranti U.K.'s computer operations were, from a profit perspective, marginal at best.1
Ferranti Canada felt that its best chance to survive was to aim at a mid-level market and stay clear of big systems. The challenge then became one of designing a lower priced, mid-sized computer. Another crucial factor that influenced this design choice was the contract deadline. From the time the contract with the Federal Reserve Bank was signed, Ferranti Canada had one year to design, build, test and install the FP-6000. Consequently, Ferranti Canada decided to use the logic circuit design that had proven successful in the RESERVEC computer system. The merits of this choice were reinforced by the experience some members of the design team had, in 1961, during their participation in the ORION II project in England.
Acutely aware of its diminishing share in the computer market, Ferranti U.K. decided to expand its product line downwards by launching a smaller machine called the ORION. In the design of the ORION, Ferranti U.K. decided to gamble on a radically new circuit design called the Neuron. This decision proved to be disastrous. At one point in the project's life, Ferranti U.K. management was fearful that the technical and financial problems might prove to be insurmountable. As result, a project for a backup computer, called ORION II, to use the Canadian RESERVEC's proven transistor circuit techniques, was started. In 1961 engineers from Ferranti Canada had gone to England to participate in the ORION II project. For a time two groups worked in parallel: the first in Manchester building ORION and the second in Bracknell building ORION II [see note 7]. This experience later served to reaffirm the Ferranti Canada's belief that from a circuit point of view RESERVEC was a sound basis for the FP-6000.
To make the FP-6000 a more attractive option for the mid-level user, Ferranti Canada decided to incorporate in it the very advanced feature of multiprogramming. By the end of 1960, only a small number of attempts had been made to implement multiprogramming on very large systems, like IBM's Stretch.22
According to E.F. Codd, the biggest challenge that the development of multiprogramming systems faced in 1962 was the design of flexible storage allocation [see note 8]. The approach taken by Ferranti Canada was dynamic program and data relocatability. Each program was allocated a range, in core memory, specified by a starting address called the "datum" and an upper address called the "limit" [see note 9]. All instruction addresses in the program were assigned relative to each program's "datum". In order to prevent any fragmentation in the available core store, each time a program terminated, the FP-6000's supervisory program would temporarily stop the execution of all the other programs and recopy them to the lowest end of core-store. In this way, at any given moment, all the system's available memory physically occupied one contiguous block at the top end of the core store. The task of dynamic memory allocation was even more complex because the FP-6000's supervisory program, called the Executive, allowed the user to program in such a way that the different parts of the program could multiprogram among themselves.
While the advance of multiprogramming had become more widespread by 1963, it still remained confined to very big expensive machines. The FP-6000 computer demonstrated, according to its designers, that:
...time-sharing facilities normally only available on larger computer systems may be realized on a medium-sized system by an intermarriage of hardware and software for minimum cost yet still retain all the safeguards necessary for an operable systems."23
The customer's ability to expand the system economically and flexibly was another important consideration in the design of the FP-6000. The power and performance of the FP-6000 could be expanded in a modular manner to accommodate a company's changing requirements and budget. The ability of the system's supervisory system to accommodate various core-store and processor options with minimal additional software costs was an essential ingredient in Ferranti Canada's modular strategy.
|Santa helping to deliver the FP-6000 to the Toronto Stock Exchange, December 23, 1963 (Ferranti-Packard Transformers Ltd.)|
In his history of the British firm International Computers and Tabulators, Martin Campbell-Kelly suggests that the design of the FP-6000 originated with Ferranti Canada's parent firm [see note 10]. This conclusion, however, assigns undue primacy to Ferranti U.K. technical expertise in shaping the development of the FP-6000. This computer was not an isolated technical and corporate event in the life of Ferranti Canada. Neither was it a product of the parent firm's technological cast offs. Rather the corporate capacity to design and build the FP-6000 was the culmination of a process that started in 1949 with DATAR.
The transistor circuit techniques used in the FP-6000 were first perfected in the computerized mail sorting system for the Canadian Post Office. By the time Ferranti U.K. had installed its first vacuum tube Pegasus computer in 1956, the Canadian subsidiary had abandoned vacuum tubes and undertaken an independent program to develop transistor-based digital electronic circuit techniques. Throughout the latter part of 1955 and early part of 1956, the digital electronic group at Ferranti Canada worked closely with Philco in the U.S. to explore the digital electronic design implications of the latter's new high speed SB-100 transistor.
RESERVEC's transistor logic circuitry, which formed the basis for the FP-6000, was the culmination of Ferranti Canada's early commitment to solid state electronics. The expertise needed to develop the real-time information processing capabilities grew out of the DATAR experience.
While it is important to recognize the innovations that originated at Ferranti Canada, it would be foolhardy to push this argument too far and dissociate the company from the internationally available pool of ideas and techniques. Ferranti U.K.'s experience no doubt had some influence on how Ferranti Canada approached the task of designing a multiprogramming environment. After all, the first comprehensive research paper on the question of multiprogramming was written by Dr. S. Gill from Ferranti U.K. in 1958.24 Ferranti U.K. had also incorporated multiprogramming features in the ORION. But the subject of multiprogramming had also become a widely discussed topic within the international computer engineering community when Ferranti Canada took on the FP-6000 contract for the Federal Reserve Bank in New York.25 Ferranti Canada engineers were familiar with these discussions, particularly those within North America, as much as they were to Ferranti U.K. ideas [see note 11].
Work on the FP-6000 was essentially completed by the end of 1962, and installed and operational at the Federal Reserve Bank in early 1963. Getting subsequent orders for the FP-6000 proved frustrating. During the course of 1963, Ferranti Canada only managed to sell two other FP-6000's: one to the Defence Research Establishment Atlantic, in Dartmouth, Nova Scotia and the other to the Toronto Stock Exchange. The sales team felt particularly discouraged by what it perceived to be an unwillingness of local and provincial governments to buy Canadian. The City of Toronto, for example, chose the imported UNIVAC over the FP-6000 when it implemented the world's first computerized traffic control system in 1963. During the early 1960s, the Ontario government ran an extensive advertizing campaigns urging people "To Buy Canadian". However, when Ontario's Treasury department went shopping for a computer, it overlooked the Ferranti Canada bid and bought American. The computer group at Ferranti Canada was convinced that Canada's import tariff structure "was a disincentive to manufacture systems in Canada."1 The duty rate on components was higher than on complete systems.
Equally frustrating was the absence of any support from the parent firm to market the FP-6000 in the U.K.. When a delegation from Ferranti Canada went to England in the late Fall of 1961 for approval to enter the general purpose computer business, Ferranti U.K. responded coldly. Not only did Ferranti U.K.'s senior computer engineers disagree with the design philosophy being proposed, but they also turned down requests for a cooperative effort. Nevertheless, the parent firm gave it's consent, however reluctant it was, because it could not deny the existence of a contract with the Federal Reserve Bank. But it was also agreed that if the FP-6000 project went well, Ferranti U.K. would offer it for sale. That promise was never honoured with any sense of commitment. Ferranti Canada, however, was unaware that this lack of support arose from the parent firm's decision to cut its own heavy losses and get out of the commercial computer business.
British Corporate Needs and the Death of Computers at Ferranti Canada
In early 1963, Sir Vincent Ziani de Ferranti approached International Computers and Tabulators Limited (ICT) with an offer to sell Ferranti U.K.'s non-military computer operations. ICT expressed little interest in Sir Vincent's proposal. ICT had its roots in punched-card machines and other office equipment. Ferranti U.K.'s product mix did not match ICT's perception of market needs. Despite ICT's rejection of Sir Vincent's offer, ICT retained a strong interest in Ferranti U.K.'s considerable R&D potential and manufacturing capability.
The appearance of the FP-6000 computer suddenly increased the attractiveness of the Ferranti offer. ICT had already decided to move its product line upwards and develop its own mid-sized computer. The FP-6000 offered a readily available technology that had proven itself. The usefulness of the Canadian computer to ICT's plans was reinforced when a Ferranti U.K. evaluation of the FP-6000 stated that:
Were we to begin designing now a machine in the same price/performance range as the FP6000, we would have in some 18 months's time a system that would not be significantly better - if indeed it were any better - than the FP6000.2
An ICT visit to Canada confirmed the Ferranti U.K. evaluation. In April 1963 ICT's upper management concluded that the FP-6000 should serve as the technological basis for the company's thrust into the mid-sized computer market.
In June 1963 ICT formally acquired Ferranti U.K.'s non-military computer operations. This deal was based on the guarantee that the design rights to the FP-6000 would be included in the deal. John Picken, a former technical director and member of the Ferranti U.K. board, later recalled that "without the FP-6000 we would not have gotten the deal we wanted from ICT. The FP-6000 was the golden brick in the sale of our operations."41 In the words of another Ferranti U.K. board member and director of the company's Edinburgh operations, the FP-6000 was "the jewel in the crown that ICT had bought."27
Though the deal allowed Ferranti Canada to follow up, well into 1964, sales leads that had been started in 1963 (and to sell additional machines), it effectively killed the Canadian company's computer business. The computer group at Ferranti Canada approached ICT with a proposal to design and manufacture ICT's proposed 1905 and 1906 computers in Canada (the FP-6000 itself became the 1904). The group tried to convince ICT that it was not only technically well suited for the task, but that Ferranti Canada also offered the British company a gateway into the North American market. Unfortunately, Ferranti Canada was not part of ICT's corporate strategy. By the Summer of 1964, it became apparent to the computer group at Ferranti Canada that a long journey had come to an end. Most of the hardware development team (Gordon Lang, Fred Longstaff, Don Ritchie, and Ted Strain) resigned. This was quickly followed by the resignations of the entire software development team (David Butler, Brian Daly, Bob Johnston, Ted McDorman, Jim McSherry, Roger Moore, Audrey Sharp, Ian Sharp, and Don Smith). The hardware team formed a digital electronics firm called ESE. When ESE was acquired by Motorola Information Systems, Lang, Longstaff, and Strain moved into the company's upper management. The software team formed I.P. Sharp Associates, which went on to build a profitable on-line global database service, as well as becoming one Canada's largest software houses. In 1987, Reuters acquired I.P. Sharp Associates. Those involved in Ferranti Canada's high speed tape readers and drum memories (Cliff Bernard, Rod Coutts, Lawrie Craig and Al Vandeberg) resigned in 1967 to set up a digital electronic engineering company called Teklogix.
B. Van der Wijst asserted that the FP-6000, which was "the first true time-sharing, multiprogrammable computer on the world market back in 1961-1963", represented an opportunity to develop an indigenous general purpose computer industry that was lost to Canada.28 D.H. Wilde, who had been the Manager of the largest FP-6000 installation at the Saskatchewan Power Corporation, reinforced Van der Wijst's contention. Wilde claimed that the FP-6000 was "certainly ahead of its time" and after eight years in operation the computer "still measures up in many respects to current models."29 In their book Entering The Computer Age, B. Bleackley and J. LaPrairie repeated the argument that the FP-6000's multiprogramming capability "was a significant technological breakthrough [which] preceded similar capabilities later introduced in the equipment of other manufacturers."30 And in more dramatic language, David Thomas contends that the sale of the rights to Canada's FP-6000 to ICT "annihilated Canada's capacity to build major computers".31
Putting aside the errors in the above assessments of the FP-6000 technical achievements, these popular historical accounts of the FP-6000 imply that this computer would have been a Canadian success story had it not been for foreign machinations. No one can deny that the sale of the design rights to the FP-6000 dealt a mortal blow to any ambitions Ferranti Canada may have had in building a computer business. But a closer look reveals that the commercialization of the FP-6000 faced serious problems even before its acquisition by ICT.
Though the design of the FP-6000 was more market than technology driven, Ferranti Canada lacked the infrastructure and experience to sell it to the market it had targeted. Furthermore, the Canadian computer group's approach to sales was part of the corporate culture it had inherited from its parent firm. Computers were sold by engineers to technically sophisticated users more on the basis of design achievements than on an appreciation of the customers operations and needs. The attitude was that technical achievement would somehow sell itself. Fred Longstaff, the FP-6000's chief designer, put the problem quite succinctly: "our marketing was zip. We were poles apart from IBM."1 Whether, with time, the Canadian company could have mounted a more effective marketing effort remains an open question.
Restricted access to capital was another fundamental obstacle to the commercialization of the FP-6000. The cost of establishing marketing and service networks would have required considerable financial resources. But the Ferranti corporate structure made it impossible for Ferranti Canada to raise the necessary capital. From the time Sebastian Ziani de Ferranti regained control of Ferranti U.K. in 1923, he and his descendants were determined to keep it a privately owned company under family control. Tradition played a powerful role in the family's refusal to go public as a way to raise capital. As a result, Ferranti U.K. relied only on profits or the British banks to finance long term expansion. As the economies of scale and corporate concentration in both the electrical and computer manufacturing sectors kept increasing, so did the capital investment needed to compete. Pursuing venture capital was not an option open to Ferranti Canada. The Ferranti family's refusal to go public became an onerous financial handicap that crippled both parent and subsidiary.
The problems of raising capital were compounded by Ferranti Canada's own struggle to survive in the electric power market. During the 1950s, offshore competition and dumping wreaked havoc on the entire Canadian electrical capital goods sector.
Trying to keep its electrical manufacturing facilities in Toronto, St. Catharines, and Trois-Rivieres afloat, in the midst of collapsing industry, was pre-eminent in the minds of the company's senior executive and board of directors. The company devoted much of its managerial energies to expanding beyond its traditional markets. In a bold move, it set out to penetrate the American market. Its Canadian competitors, which were all subsidiaries of large American multinationals, could not sell in the U.S. and compete against their own parent firms. Struggling to survive in the North American electrical capitals goods sector absorbed all the energies of Ferranti Canada's top management.
To underwrite any serious entry into the computer business would have entailed an almost complete reallocation of the subsidiary's own scarce capital away from the company's plan to save its electrical business. The one million dollar compensation that Ferranti Canada accepted from its parent firm, in exchange for the loss of the design rights to the FP-6000, was clear evidence that saw Ferranti Canada as being first and foremost in the business of selling of electrical capital goods to industry and the utilities [see note 12]. This perception was reinforced by upper management's unwillingness to bridge the gulf that separated two different technological cultures that had come to make up the company: the old electrical revolution and the new digital electronic revolution.
It remains unclear whether Ferranti Canada consulted its subsidiary before agreeing to sell off the FP-6000 to ICT. Ferranti Canada management claimed that the deal was contingent on its consent. But no documents have yet to be found that corroborate or refute this claim. Given the pivotal importance of the FP-6000 to the ICT deal, one must ask whether refusing to sell was a real option for the Canadian company. After all Ferranti Canada was wholly owned subsidiary of Ferranti U.K., of which Sir Vincent was the principal owner. Whether Ferranti Canada actually could have acted otherwise is a moot point. In the end, however, Ferranti Canada's preoccupation with saving its troubled electric power business made the company an agreeable participant in the sale of the FP-6000 to ICT. The acceptance of the sale of the FP-6000 rights was also a realistic admission by the Canadian company's executive that it had no idea how to go about turning the computer into a long term profitable business.
Sixteen years later, the Canadian executive was not as compliant when the parent firm suggested that it give up another digital electronic innovation developed in Canada - the flip disc display. By this time, Ferranti Canada had decided to diversify and break its vulnerability to the wide swings that periodically ravaged the electrical capital goods sector. When Ferranti U.K. offered to sell Ferranti Canada to NEI, it wanted to retain control of the flip-disc technology. Ironically, Ferranti U.K. expressed little interest in this technology when it was first being developed in Canada. [see note 13]
|John Vardalas (John Vardalas)|
John Vardalas has a B.S. in physics, an M.Sc. in Mathematical Physics, and an M.A.in Economic Geography. He is currently finishing his doctorate in history at the University of Ottawa under a fellowship provided by the Social Sciences and Humanities Research Council of Canada.