DATAR - First Digital Computer/Communications System for Anti-Submarine Warfare
From GHN(Difference between revisions)
m (Text replace - "[[Category:Computers_and_information_processing" to "[[Category:Computing and electronics")
|Line 146:||Line 146:|
Revision as of 16:14, 22 July 2014
World War II Antisubmarine Warfare Sets the Stage
In World War II, Germany reasoned that if it could choke-off all the transatlantic re-supply lines to Great Britain, from Canada and the United States, then Great Britain's demise would only be a matter of time. The failure of Germany's surface fleet to sever Great Britain's life-line to North America, led to the promotion of the submarine as Germany's principal form of naval warfare. Unless they were one of the very fast luxury passenger liners, like the Queen Mary, sending solitary supply ships cross the Atlantic was sheer folly. Their slow speeds made them perfect prey for German submarines. To assign a naval vessel to escort each supply ship was also utterly impractical.
The Allies concluded very early on that there was safety in numbers. Large convoys lost proportionately fewer ships. Despite this advantage, protecting slow moving convoys that extended over many square miles proved extremely difficult. By the end of 1942, the German submarine "wolf-packs" were exacting a devastating toll on Allied shipping. In November 1942 alone, 720,000 tons of supplies were sunk by German submarines. With the rate of shipping losses exceeding the rate of production, the Allied leaders gave the submarine problem top priority at their January 1943 meeting in Casablanca. "If the menace [from submarines] could not be conquered", explains historian Gerhard Weinberg, "the steady diminution of Allied tonnage would immobilize the Western Allies." By 1943, the battle to control the shipping lanes had become World War II's pivotal "battlefield". For Hitler, the submarine campaign had assumed, next to the Eastern front, the most important role in Germany's war strategy.
Canada devoted all its naval resources to the role of protecting the transatlantic convoys. By the end of 1942, Canada provided 48 percent of all the convoy escorts. And yet, despite this large contribution, Canada had no say in the strategic use of its considerable anti-submarine resources. The dismissive attitude taken by British and American naval authorities to Canada's views on the disposition of its own naval resources frustrated and angered senior RCN officers. "The British Admiralty," concluded Canada's Admiral Brodeur, "still looked upon the RCN as the naval child to be seen and not heard when no outsider [the U.S. Navy] looked on or listened in". U.S. naval authority was no different added Brodeur. Japan's entry into the war provided the RCN with unexpected bargaining leverage to pry the command of Atlantic convoy escort operations away from the U.S. because it forced the Americans to pour all its naval resources into the Pacific, leaving little for the Atlantic. The time was now ripe for the RCN to win important concessions from its patronizing allies.
Post World War II Realities: Need for New Antisubmarine Warfare Technology
Determined to preserve its hard-won special status in the post-war era, the RCN wanted to cast itself in the role of the Western alliance's anti-submarine and escort navy. By 1947, the RCN realized that advances in submarine technology would soon render its World War II anti-submarine (A/S) equipment obsolete. Submarines would be faster and armed with homing torpedoes. If the RCN wanted to be taken seriously in its efforts to be recognized as an A/S force, then it would have to embark on its own research and development program. Like the other Services, the Navy understood that the intensive exploitation of science and technology had become the sine qua non of military competency in the post-war era. But 1946, the RCN R&D ambitions had yet to be articulated. By the end, of 1948, however, this agenda started to take shape. The most daring item in this agenda was to move naval warfare into the electronic digital universe. The impetus for this R&D came from the technical mid-level officer corps within the Electrical Engineer-in-Chief's Directorate (EECD). Yet it was a receptive senior officer corps who provided the essential political and financial support. These senior officers, who had fought so hard to win military sovereignty in the Battle of the North Atlantic, were ready to finance any bold technological undertaking that bolstered the RCN's claim to be the backbone of the Alliances' anti-submarine efforts.
The inability to capture, extract, display, communicate and share accurate tactical information in a timely manner had been a central limitation of war-time A/S operations. The movement of large convoys across the Atlantic during World War II had presented monumental logistical and tactical challenges. Keeping track of positions of all the ships in a convoy that stretched over many square miles proved problematic. In a battle situation, the difficulty was compounded by the need to monitor the movement of all the enemy submarines. The long human chain needed to convert asdic, radar, and other tactical data into useful information for command-and-control was slow and often unreliable. While asdic signalled the convoy to the presence of attacking submarines, radar tracked enemy aircraft. With a new post-war generations submarines and aircraft, the RCN realized that the slow, human-intensive chain needed to convert input data to a coordinated tactical response had become a serious weakness in anti-submarine operations.
While the flow of tactical information within the anti-submarine ship was slow, precious little flowed between escort ships. The absence of inter-ship, real-time, tactical data exchange severely limited the ability of the escort fleet to respond as one unit. The RCN came to believe that there was an "urgent operational requirement" for new systems that would allow "for closely coordinated tactics by convoy escort and hunter-killer anti-submarine groups." The Development Section of the Electrical Engineer-in-Chief's Directorate advocated an integrated and automated information system that could simultaneously provide the Command of each anti-submarine escort with a complete, concise and up-to-date picture of the tactical situation, provide the necessary information to Weapon Control systems to assist in Target Designation, and that could also incorporate tactical information on all aircraft into the battle picture. Unless the RCN could automate the production, exchange, and use of tactical data, reasoned EECD's technical people, Canada's A/S escort would proved ineffective against the coordinated attacks of a new generation of submarines and aircraft. But how could one achieve this technological breakthrough? To the technically minded mid-level officers in the EECD's Development Section, the ENIAC computer offered the answer.
A top secret project during World War II, ENIAC had been built in order to accelerate the calculation of ballistic and bombing tables. Still the only fully electronic digital computer in the world in 1948, ENIAC was an obscure technology known and understood by a very small circle of people. But the members of the Development Section had followed the military reports on ENIAC with great interest. The incredible speed and precision of electronic digital computation made a dramatic impact on their technical imaginations. They saw more than just a radical advance in calculating technology. In these officers' minds, the new electronic digital paradigm offered a revolutionary way to unify the collection, interpretation, communication, and representation of tactical information into one automatic, interactive and decentralized network linking all the escort ships of a convoy. They called this first comprehensive digital perspective of naval warfare Digital Automated Tracking And Resolving, or DATAR for short.
Who Will Design and Build DATAR?
By the Fall of 1948, DATAR was still only an idea. Eager to prove its technical feasibility, the Development Section of EECD grappled with the question of how to do the research and development. The RCN had neither the electronic laboratories nor technical manpower to tackle DATAR single-handedly. In fact in 1948, few in the world had any knowledge or experience with the digital electronics, communications, and computer issues raised by DATAR. In Canada there was no one, as the pioneering UTEC project at the University of Toronto had not yet started. The search for an industrial partner for DATAR led the RCN to Ferranti Ltd. and its Canadian subsidiary in Toronto, Ferranti Electric Ltd.
The Canadian subsidiary, Ferranti Electric Ltd., had started in 1908 as a simple distribution agency, run by the prominent Royce family of West Toronto Junction (who also owned the Toronto Suburban [Electric] Railway) to sell the watt-hour meters of the British electrical manufacturer, Ferranti Limited. The rapid and widespread adoption of hydro-electric power that accompanied Canada's rapid economic expansion during the 1896-1912 boom period created an enormous potential for electrical capital goods. At the peak of this buoyant electrical market, in 1912, the agency created by the Royce family was succeeded by a full-fledged trading company, Ferranti Electrical Co. of Canada Ltd., which will be referred to as Ferranti Canada for the purposes of this article. Over the next 80 years, three factors shaped the growth of the Ferranti Canada: the need to survive in a North American electrical capital goods market dominated by the large American multinationals General Electric and Westinghouse; the Canadian subsidiary's continual pursuit of greater manufacturing and design autonomy; and the transfer of the parent firm's technology-driven corporate culture to the subsidiary. As of 1948, other than an X-ray Department, which produced machines for war-time industrial inspection and small portable units for the field hospitals, Ferranti Canada had no experience in electronics. But its parent firm had made considerable inroads in this area and was interested in getting a foothold in the burgeoning North American defence market.
E.G. Cullwick, Defence Research Board's Director of Electrical Research and a former RCN officer, was aware of the EECD's search for an industrial R&D base from which to build DATAR. In October of 1948, Cullwick called Lieutenant Jim Belyea, who was one of the leading advocates of the DATAR concept within the EECD, and informed him of the Ferranti Ltd. Delegations visit. Ferranti U.K.'s experience in digital R&D dovetailed well with technical challenges raised by DATAR. Ferranti U.K. had supplied 30 percent of the British Army's war-time requirements in servo-control equipment, which is an integral part of the automated fire control technology. After the war, Ferranti U.K. took up the "design of naval fire control equipment involving electronic computers, regenerative tracking & automatic following" for the Royal Navy. Ferranti U.K.'s work on the "Admiralty Flyplane" embodied electronic computing. Dr. Prinz, an eminent research scientist on staff at Ferranti UK, was studying methods of high speed data transmission. Finally, Ferranti UK had just received a 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. After hearing Cullwick's profile of Ferranti U.K., Belyea seized the moment and visited the delegation at the Chateau Laurier.
It is not known whether Belyea went to the see the Ferranti group with the clear plan to enlist its help. But during the meeting, Belyea became convinced that the Ferranti organization was the company to develop DATAR. At the same time the RCN's project appealed instantly to the technology-driven Ferranti group. Not only did the work fit well into the company's current R&D activities, but nothing in England came close to the scope of DATAR. "It seemed to our group," reflected Kenyon Taylor, who was a member of that delegation, " 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." With RCN support, Ferranti was ready to assemble Canada's first leading-edge industrial R&D group in digital electronic computers and communications.
The Development Section understood that unless it could prove that tactical information could indeed be exchanged reliably between two distant points through digital encoded radio communication, there was little hope that an understandably sceptical RCN hierarchy would embrace this radical concept. Belyea offered Ferranti Canada the challenge to develop the completely novel digital electronic circuitry based on a concept first proposed by H.A. Reeves in 1939 but never really explored. The idea behind Reeves’s Pulse Coded Modulation or PCM was to convert analog signals into a series of binary values for the purpose of communication. Today PCM is an ubiquitous element in all communications technology; but in 1948 it was unexplored territory.
Too late to obtain a formal budget, Belyea funded the first year's work by siphoning off money from other projects. What should have been one relatively large contract was instead broken down into smaller contracts. By keeping each contract under $5,000, and thus within the realm of his discretionary spending authority, Belyea bypassed the need to get formal spending approval from outside the EECD. In a span of three weeks, the Development Section issued four contracts. The first dealt with the study of digital transmission methods and devices; the second, with the design and construction of binary digital components for transmission systems; the third, dealt with the design and construction of PCM components and devices; and the fourth contract dealt with the design of an experimental PCM transmission system.
Though Belyea's contracts were quite modest, they did demonstrate RCN's good faith to Vincent Ziani de Ferranti. In January of 1949, Ferranti sent Kenyon Taylor, his most imaginative inventor, to set up an electronics R&D team in Canada. Taylor had started working for Ferranti in 1931 as a "lab boy" when textile machinery still occupied an important place in the pantheon of Ferranti technological interests. The company's founder Sebastian Ziani de Ferranti had devoted considerable effort to designing high-speed textile machinery. Taylor's success in applying electronic techniques to improving the operation of Ferranti textile machines not only led to Taylor’s first patent, but it also brought his inventive talents to Sir Vincent’s attention. Over the years, they had become close friends. By the time Taylor set sail for Canada, he had acquired 55 patents that covered both consumer and military electronics. As the leader of the new Ferranti Canada electronics group charged with the responsibility of turning the RCN DATAR concept into concrete engineering, Taylor offered the invaluable combination of great inventive talent and a wealth of practical, hands-on experience.
The money from Belyea's four small initial contracts to Ferranti Canada quickly ran out. The Development Section squeezed out an additional $15,000 in May of 1949, but considerably more money was needed to demonstrate the technical feasibility of PCM. To release more funds, the Electrical Engineer-in-Chief's Directorate now had to seek more formal approval of the project from higher levels within the RCN. The series of small contracts to Ferranti Canada had already given DATAR a momentum of its own that made it difficult for senior RCN officials to refuse the additional funds needed to complete the demonstration. On 7 October 1949, Chiefs of Naval Staff approved this initial phase of DATAR. Three weeks later, an additional $50,000 was awarded to complete the final design and construction of the experimental PCM transmission equipment. However, any further financial support for DATAR hinged critically on the success of the proposed communications demonstration.
The idea of the demonstration was to transmit simulated tracking data reliably from Ferranti Canada's laboratories, via radio PCM, in order display the targets' movements in the RCN's Ottawa laboratory. Computing Devices of Canada was to develop the CRT display equipment for the experiment. Founded in 1949 by George Glinski, a professor of electrical engineering at the University of Ottawa, and P.E Mahoney, Computing Devices of Canada became the other strategic industrial element in the RCN's efforts to enter the digital age. The demonstration held in February 1950 made a vivid impression on the Navy's senior staff. A small Canadian team had shown the technical feasibility of digitally sharing target information in real-time, among a fleet of A/S escorts. Euphoric over the success of the small PCM demonstration, the DATAR group decided to take the giant leap into developing a large-scale, naval weapons system. On 5 April 1950, Captain Roger, the Electrical Engineer-in-Chief, went to the RCN's Research Control Committee with a request for a $1.5 million, 30 month program, to design, develop, build, and test a DATAR system prototype.
The Road From Concept to Prototype: A Journey That Took Too Long
Six weeks after his first submission, Roger returned with humbler and more methodologically incremental proposal which the Research Control Committee quickly approved. Instead of jumping directly into full scale prototype, Roger's Directorate was given $279,000 to pay for a thirteen month, detailed engineering study. Most of the money went to underwrite Ferranti Canada's further experimental exploration of all the digital circuit design issues raised by DATAR. The new proposal also responded to the Committee's call on a co-ordinated and comprehensive analysis of the RCN's future DATAR requirements. As an automated information processing system, the DATAR concept required the close harmonization of needs and practices across several RCN command jurisdictions. And yet surprisingly, Roger's original submission made no reference to any intra-service coordination. At the Research Control Committee's insistence, the elaboration of the RCN's specific DATAR requirements was to be a collaborative effort between the Electrical Engineering, Telecommunications, Weapons and Tactics Directorates.
The cornerstone of Canada's post-war military self-reliance was to avoid head-on competition with either the U.S.A or the U.K. in the research, development, testing and production of new defence systems. Within this policy framework, Canada sought to spin an R&D thread that both suited national needs and complemented the work of its allies. Despite the more modest nature of Roger's proposal, by its very nature DATAR promised, in the long-term, to be a large, complex and costly undertaking that was to challenge British and American R&D programs. However, a proposal of this magnitude required ministerial buy-in.
The RCN assured the minister that DATAR fit very well within these constraint of Government policy on military R&D. Not only was it a pivotal technological component of a specifically Canadian responsibility, anti-submarine warfare and convoy escort, but it was also a uniquely Canadian concept in its scope. As the Deputy Minister of Defence explained to Claxton, no one in either the United States or Great Britain had yet to embark on anything resembling DATAR. Canada stood apart, continued the Deputy Minister, in its vision and achievements in the area of digital naval tactical information systems. But the Deputy Minister reassured the Minister that the RCN's program to develop an automated tactical information system would nevertheless be carried out in close cooperation with its allies, to ensure that no wasteful reinvention of the wheel took place. However, underneath this commitment to cooperate was the RCN's firm conviction that because of DATAR's superior technical vision, any transnational harmonization of R&D in this area should be on Canadian terms.
By the time of the successful PCM transmission in February 1950, Ferranti Canada had already recruited a few bright young Canadian electrical engineers. But the company advised the RCN that any further efforts to build the kind of development engineering staff required by DATAR called for a more substantial and long-term commitment from the military. Going from small contract to small contract made it very difficult for the company to nurture a stable design team. Ferranti Canada was demanding a minimum of $250,000 worth of contracts annually over three years. Instead, Roger's proposal only assured them $234,000 for one year only. However, by October of the following year, Ferranti Canada was able to extract an additional $200,000 from the RCN to continue its engineering study and circuit experimentation. One factor that contributed to the contract increases was the unexpected extension of Ferranti Canada's technical responsibilities.
Originally, the Electrical Engineer-in-Chief's Directorate had wanted Ferranti Canada to explore all aspects of DATAR but one - the computer unit. Since Computing Devices of Canada had already been assigned the contract to design and build a digital computerized battle simulator, it only seemed natural that it design DATAR's computer. The RCN allocated $45,000 to Computing Devices of Canada to do a nine month preliminary engineering study of DATAR's computer units. The simulator turned out to be a very complex and all-consuming undertaking for the small Ottawa-based company. Fearful that Computing Devices of Canada could not handle both projects, the RCN asked Ferranti Canada, in early 1951, to build a small 12-bit experimental computer as part of its DATAR work.
This new role had a dramatic effect on Ferranti Canada's subsequent technological and corporate objectives. Suddenly, the design and manufacture of electronic computers became a powerful and pervasive ambition among Ferranti Canada's electronics engineers. Parallels were made between Ferranti Canada and its parent firm in the U.K.. In 1951, [Sir] Vivian Bowden from Ferranti UK, while visiting Canada's atomic energy project, asked his friend W.B. Lewis, "if there was any possibility that Ferranti, Toronto, might be given a contract to build a computer on the same sort of terms as Ferranti UK had." With a contract from the British government, Ferranti U.K. had transformed the prototype computer developed at Manchester University into a commercial product. Could Ferranti Canada do the same with the research going on at the University of Toronto? But the subsequent collapse of the UTEC project abruptly ended this prospect.
A year later, buoyed by the rapid progress being made on DATAR, Arthur Porter boasted to 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." Davies offered Porter an opportunity to make good on his boast. "I am suggesting a method by which Ferranti Canada could effectively and cleanly cut the throat of Ferranti England." He went on to elaborate:
Although the Electrical Engineer-in-Chief's Directorate had succeeded in clearly articulating a consensus view of what Canada's antisubmarine DATAR requirements should be, and Ferranti Canada had made considerable progress in its development work, the RCN seemed content to precede in a very slow and conservative manner. Then in the Fall of 1951, a new sense of urgency invaded the DATAR project. The RCN's Electrical Engineer-in-Chief's Directorate learned that the British Royal Navy's Automatic Surface Plot (A.S.P.) system, which was a rival to DATAR, would be ready for testing as early as 1953. The RCN's schedule for DATAR called for field tests in 1954-55. Roger asked the Research Control Committee for substantial funding increases in order to speed up the program in order to field test DATAR at the same time the British planned their tests.
The Committee balked at Roger's request and only increased spending by $100,000. Roger reminded the Committee what was at stake. Britain was pushing hard to get its allies to accept the A.S.P. system. If Canada did not field its own demonstration prototype to coincide with the British field trials, the DATAR program could well collapse. Unless Canada sold the DATAR technology to its allies, the RCN had little hope of underwriting the costs of DATAR through its own procurement. By allowing the British field trials to go unchallenged, ran Roger's argument, the A.S.P. system would become the de facto standard. What galled the RCN's DATAR proponents was the prospect of having to buy a technology which they considered to be intrinsically inferior in versatility, capacity, accuracy, and operational performance to the Canadian system. Roger's arguments eventually found resonance within the RCN's nationalist senior officer corps and the race between the Canadian and British systems commenced.
The Electrical Engineering Directorate planned to test an experimental version of DATAR on three escort ships. With the new pressure of a June 1953 deadline, the pace accelerated and the costs climbed. By January of 1953, the cost of designing and building the demonstration DATAR equipment had climbed to $1.57 million. Despite the RCN's technological nationalism, this kind of expenditure would not have been possible had it not been for escalating political tensions and the outbreak of the Korean War. Following the outbreak of the Korean War, Canada committed itself to an accelerated rearmament program. From $387 million in 1949-50, the Department of National Defence's budget more than doubled to $784 million in 1950-51. The following year, 1951-52, the military expenditures doubled again to $1.6 billion. In the 1949-50 fiscal year, Navy appropriations stood at $73 million. By 1953-54, they had swelled to $332 million. In this huge swell of defence spending, the RCN had considerable room to fund its single most daring weapon system development.
Even with the deeper military pockets of the early 1950s, the continual procession of cost increases to the DATAR contracts raised eyebrows within the RCN. Stan Knights, the project's technical officer, questioned Ferranti Canada's ability to manage the large project. Ferranti Canada's costs were excessive, Knights told the Research Control Committee, because the company "had been lax in the preparation of its estimates and has not exercised due economy in its methods." Yet, committed to getting a demonstration ready by the Summer of 1953, the RCN felt it had little choice but to pay for what it considered inefficient management.
The demonstrations of DATAR took place on Lake Ontario, from September into November. Originally three Canadian Bangor class minesweepers were to be used in the demonstration. Pulling three ships out of service in the height of the Korean War proved impossible. Instead, the demonstration outfitted a shore station to simulate the third ship. All asdic and radar data, which came in as analogue signals, was immediately filtered and converted to digital information. Sophisticated CRT displays on each ship depicted aircraft and submarines as distinctly different graphical icons. Information about any of these targets could be called up instantly by placing a cursor over its image on the screen. All, or any portion, of the information stored in the computer was shared, via pulse coded modulation transmission techniques, to all or any selected ships of the escort. By taking into account the relative motions of all the escort ships, DATAR presented the complete picture of the tactical situation appropriate to each ship's own reference frame, even though the asdic or radar data displayed on one ship may have been collected by another. If the target was masked by such things as sea turbulence, rain, clouds, or radar windows, DATAR's computer supplied clues as to the probable location of the target.
Because of the rush to get a DATAR demonstration ready quickly, compromises were made in the design of the electronic digital computer. The Ferranti Canada design team chose a magnetic drum, over more elaborate memory technologies, as the computer's main memory. Given the overall complexity of the project, drum memory offered the most reliable option for the demonstration. The reliability problems of electrostatic memory systems would have unnecessarily risked the overall success of the demonstration. Even so, both the RCN and Ferranti Canada knew that the final DATAR prototype had to find a better solution to memory. Drum memory was too slow and electrostatic memory too bulky. Both of these memories were also quite vulnerable to a ship's vibrations and motion. Ferranti Canada's desire had been to pursue Forrester's idea of ferrite cores. But this idea still required considerable development work and the race to get a demonstration ready did not permit this luxury.
The Battle of the Systems: Canada's Effort to Set Standardization
As one would expect, the Canadians and British had different interpretations of the results of the DATAR demonstration. The RCN was elated. Though the demonstration equipment was small in scale and experimental, it nevertheless proved that DATAR technology had the ability to coordinate, with remarkable speed and accuracy, the collective actions of escorts ships and aircraft in an anti-submarine operation. Commodore Lay, who had played a key role in Canada's quest for military autonomy in the Battle of the North Atlantic, announced to his fellow members on Permanent Joint Board on Defence (PJBD) that DATAR had proven the superiority of the Canadian naval digital tactical information system. On the other hand, while impressed with the technical advances in data handling made by DATAR project, representatives from the British Admiralty concluded that the U.K.'s A.S.P. system was still cheaper, simpler and more effective than DATAR.
The British criticism of DATAR reflected a fundamental disagreement with Canada over the role of air defence. The Royal Canadian Navy insisted that any naval tactical information system would have to integrate air defence and anti-submarine data seamlessly. The modern convoy escort, the RCN felt, had to defend against both aircraft and submarines. If the escort ships' weapons could not directly attack enemy aircraft, DATAR could at least coordinate the use of friendly aircraft to fend off the attack. The Royal Navy, on the other hand, saw the two as different problems requiring different systems. While the Canadian Navy was committed to a digital system, the Royal Navy's still pushed the more conventional analogue approach to asdic technology. A.S.P. was effective against submarines to the extent that if kept track of a relatively small number of targets. But the A.S.P. analogue approach became overly complex and unwieldy when it came to handling a larger number of targets, as would be the case with air defence. The digital approach not only allowed the easy merger of air and submarine tactical data, but the use of a computer gave the system a flexibility the British A.S.P. could never have. Under DATAR, each ship's computer could extract, from a common pool of data, the information appropriate to its needs in a given context. The A.S.P. system, on the other hand, presented information in a pre-set and rigid manner.
The Royal Navy also argued that because of Canada's insistence on a comprehensive air/antisubmarine system, DATAR had become too large to fit on anything but a large battle cruiser. But many escort vessels fell into much smaller classes. Canada's rebuttal was that DATAR's design was based on a modular concept. Different sizes of DATAR could be installed on different ships. In today's terms, DATAR used a kind of open system philosophy. The key was a standardized digital protocol for information exchange. A key feature that distinguished DATAR from A.S.P. was the former's decentralized network, in which each ship used its own criteria for information extraction. Sinking one ship, along with its computer, did not disable DATAR. The British A.S.P. system, however, offered a very centralized approach to coordinating the protection of a convoy during attack.
Despite the superiority of DATAR, there was little likelihood that the British Navy could escape from its colonial paternalism to adapt the Canadian system. But for Canada what mattered most was the position of the United States Navy on the standardization issue. As the kingpin of the alliance, the U.S. could make or break any proposal to standardize naval tactical information systems. Because the U.S.N. had done very little to develop it own tactical information handling technology, Canada worried that the British would be able to persuade the U.S.N. to adapt the A.S.P. system. After the success of the DATAR demonstration, there was little likelihood of this. The pivotal issue for the RCN then became the intentions of the U.S.N.. Would the U.S.N., which still lagged considerably behind Canada in the area of automated tactical information handling technology, embrace DATAR? Canada lobbied intensely to bring the U.S. onside. On the 9th of November, the Minister of Defence, Brooke Claxton, three Rear Admirals, the heads of various RCN directorates, the RCN's principal scientific advisers and the RCN's officers in charge of DATAR all gathered to sell the Canadian system to a high ranking U.S. Navy delegation, headed by Rear Admirals W.G. Schindler and M.E. Miles, and Dr. E. Piore, the Deputy Chief of the Office of Naval Research.
After a demonstration of DATAR, the Canadians outlined the operational capacity of the full-scale version of DATAR they intended to build. DATAR was to be the state-of-the-art in automated naval warfare, with each ship equipped with high power air search radar, high surface search radar, high definition navigational radar, an airborne early warning system. attack and scanning asdic, radio direction finding equipment, and UHF radio communications. In a sixteen ship escort fleet, DATAR would allow the simultaneous sharing of the input from all the above devices on all the ships. Each ships computer was to have the speed and power to keep track of 128 targets simultaneously; process an additional 1,000 possible messages for each target; compute automatically, on demand, the course and speed of any target; and produce specialized plots for tactical control and target designation. To overcome the performance, size, and reliability limitations arising from dependence on a memory drum widespread use of vacuum tubes, the memory in the final version of DATAR was going to all ferrite core.
The U.S. delegation was impressed by the Canadian achievement. It was particularly the Canadian claim that the equipment cost-per-ship for DATAR was going to be $400,000 that caught Dr. Piore's attention. He had expected it to be closer to $1 million per ship. Canada had a lot riding on the American reaction to DATAR’s techno-economic merits. If the Americans did not standardize on DATAR, Canada would have a very difficult time financing the further development and production engineering, and the cost of procurement. Even more disastrous, if the U.S.N did not agree to the Canadian data exchange protocol, the RCN would be forced to drop DATAR, regardless of whether it could pay for it or not. The RCN realized that unless it retained the initiative and organized international talks on standardizing the contents and protocol of naval data exchange, DATAR technology would be useless to Canada regardless of its superiority. Commodore Lay pressed the Americans to embark on a collaborative undertaking. Though U.S. Rear Admiral Schindler endorsed the idea of drawing up common requirements, it was still too early to know how close to DATAR the U.S.N. was willing to move its weapons systems.
Faced with a nationalist Pentagon reluctant to buy a foreign weapons system, particularly one as critical as automated electronic naval warfare, and a powerful U.S. defence industry lobby anxious to block foreign equipment programs, the only possible hope Canada had to sell the Americans on DATAR was to get a full scale version built and tested before the U.S.N. had a chance to initiate its own serious program. And yet without some form of standardization assurances from the U.S.N., investing heavily in a large scale production version of DATAR could prove impoverishing. Caught in a chicken-and-egg dilemma, the RCN fell back again into its slow and cautious approach.
Time was of the essence. Yet nearly a year elapsed after Canada's high level meeting with U.S.N. representatives before the Electrical Engineer-in-Chief's Directorate felt it was ready to issue contracts for the construction of the first full-scale prototype of DATAR. Through most of 1954, the RCN's DATAR group re-examined, in great detail, the scope of the operational specifications it had already defined a in 1952, and then again in 1953. In part, this process was driven, as are many military development projects, by the tendency to seek every possible strategic advantage by pushing the state-of -the-art to its limits. Increasing DATAR's marketability was another equally, if not more, important factor behind the expanded operational specifications. In 1953, the RCN had told the U.S.N. delegation that DATAR's eventually would be able to handle 128 targets simultaneously. However, by November 1954, the specifications for DATAR had risen to include tracking 500 aircraft, plus other ships and submarines. This inclusion of air defence into the original DATAR concept was premised on the argument that defence against aircraft played an important role in convoy escort and anti-submarine duty. But no escort role could justify a 500 target air defence information system. Only a major battle group could. Thus the increased air defence capacity represented the RCN's desire to make DATAR more attractive to Canada's Royal Canadian Air Force and the United States Navy.
The RCN wanted the RCAF to adopt DATAR technology for its air defence system. In the event that the United States Navy did not buy into the DATAR development, then the burden of the high development and procurement costs could be shared with the RCAF. When the RCAF proposed to develop its own automated data processing, transmission and display system, the RCN objected. Despite inter-Service differences in tactical information requirements, the RCN felt there was sufficient overlap to warrant close collaboration along the framework of DATAR. To coax the RCAF closer to DATAR, the RCN offered it the equipment used in the 1953 demonstration. To what extent the RCAF was willing to cooperate with the RCN is unclear. In 1954, the RCAF did contract Ferranti Canada to do a preliminary study on how high-speed digital data-processing techniques could be applied to Canadian air defence. Ferranti Canada's study proposed a decentralized system and stressed the importance of rapid automatic dissemination of all data rather than the automation of the control and guidance systems. Though called the Canadian Air Defence Automatic Reporting (CADAR) system, Ferranti Canada had actually offered up a version of DATAR. But USAF plans to press ahead with the full scale production of its own system forced the RCAF to contemplate purchasing U.S. equipment. In the end, trans-national standardization with the USAF proved more of an imperative than national harmonization of R&D with the RCN.
An awareness of U.S.N. interest also explains the expansion of the initial 1951 specifications of DATAR into a generic naval tactical information system. How else can one explain the jump from 128 aircraft tracking capacity to 500? Although the integration of air defence and anti-submarine tactical information seemed reasonable for a convoy escort navy, the need to track 500 aircraft made no sense, unless of course, one envisioned DATAR for an American aircraft carrier fleet.
The early success of the DATAR project permitted Canada to play an important role in defining trans-alliance standardization. The RCN managed to convince the U.S.N that binary, rather than decimal, was the best format for naval data exchange systems. Canada's insistence on error detection in the data communication process, as well as the use of the UHF bandwidth for greater throughput, was slowly being accepted by its allies. Canada's espousal of an open, decentralized, network architecture also appealed to the U.S.N.. The British approach, which was now known as the Comprehensive Display System (CDS), called for a centralized architecture in which all the processing was done on one ship and the required tactical information then broadcast to each ship. In DATAR, although each ship contributed to common pool of tactical data, information extraction was a local process determined by each ship's own computational capacity and needs. While the British approach reinforced the military tradition of hierarchical and centralized systems, the Canadian approach highlighted the advantages of a fluid, flat and less vulnerable organizational structure. By empowering each ship, the DATAR philosophy maximized the survival of the entire system. If any one ship was destroyed, the ability of other ships to share and extract information would not be jeopardized, as would be the case in the centralized British system. When the U.S. had finally committed itself to the Canadian position that air defence and anti-submarine concerns form one integrated information system, the RCN knew it had finally triumphed over the RN. Members of the RCN's Research Control Committee were no doubt pleased with the news that "the U.S.N.'s wanted to adopt a modified version of DATAR for the next phase of its own development program" because it was convinced that the Canadian approach was the right solution for the modern warfare of the next decade (1960s).
The RCN's victory, however, proved to be an empty one. The slow pace of DATAR's development not only jeopardized any possible export sales, but it also made the project more vulnerable to the flagging Canadian political will to pay for costly defence R&D programs. Nearly eighteen months had passed since the experimental version of DATAR had been successfully demonstrated to the world before contracts were approved to start the design, development and manufacture of the large-scale prototype. The RCN had divided the DATAR prototype program into two contracts: data processing and display; and the radio data link subsystem. When the specifications for the full scale version of DATAR had finally been set in November of 1954, the plan was to have the prototype ready in two years. Over a year had come and gone and the basic components of the system had not even been designed, never mind built and tested. A small part of the blame no doubt rests with bureaucratic inertia. However, DATAR project's slow pace owed more to fundamental technical limitations than to bureaucracy lethargy.
As with most of the early attempts to build and maintain large digital systems, issues of design complexity turned the vacuum tube into a critical "reverse salient". The escalating scale and complexity of military electronic design had turned. The sheer size and power requirements of large-scale circuits made it increasingly more difficult to fit vacuum tube-based electronic equipment onto ships and airplanes. Even more serious was the growing problem of reliability. As the number of tubes in military electronics grew into the thousands the mean time between failure (MTBF) became unacceptably short. The Ferranti Canada team had experienced first hand the urgency to miniaturize large-scale electronic digital systems and still improve reliability. About 15,000 thousand vacuum tubes comprised the DATAR system tested on Lake Ontario. Whenever DATAR was being demonstrated, it was not unusual to find Ferranti Canada's young engineers, stripped down to the waist because of the heat and armed with cartridge belts of vacuum tubes, racing within the ship's interior locating and replacing faulty tubes. Even though the full scale version of DATAR was to have proportionately fewer tubes, the questions of reliability, space, and power consumption haunted the Electrical Engineer-in-Chief's Directorate.
The solution to the complexity problem resided in finding radically new electronic switching devices, and not in incremental improvements to existing components. It was during this crisis of the vacuum tube that the military looked to the transistor as a potential replacement for the vacuum tube. Still early in its development, the transistor was itself plagued with serious performance and quality control problems. Its relatively slow switching speeds also limited the transistor's utility for data processing and communications technologies. Furthermore, the absence of tightly controlled transistor characteristics made it difficult to design high performance digital circuits. Despite these problems, the transistor's potential for far greater miniaturization offered the only hope of getting through the complexity impasse. The RCN was convinced that unless DATAR could be transistorized, it would never go into service.
Philco's 1954 surface barrier innovation seemed to offer the RCN the answer to DATAR's miniaturization imperative. The SB-100 transistor offered reliable, high-speed switching. Over the next few years, Philco used its SB-100 transistor as the basis of the world's first all solid state computer. As soon as news of the SB-100 became public, the RCN immediately arranged for a team of Ferranti Canada engineers to visit the Philco plant in Philadelphia with the mandate to explore the utility of the SB-100 transistor for DATAR. Ferranti Canada immediately began experimenting with the transistors and figuring out how to recast DATAR's circuitry into solid-state. While this course of action seemed straight forward, its execution proved difficult and slow.
Time had now become the DATAR project's biggest enemy. For nearly seven years, Canada had held an unquestioned lead in the formulation and technical pursuit of automated naval tactical information systems. But by 1956, a shift in the Canadian political will to finance military self-reliance and the sudden awakening of the U.S.N. to the "urgency" of automated tactical information systems were about to ravage this lead. These two factors combined to bring about the collapse of the RCN's DATAR project.
Canada’s small economy could only support modest levels of military procurement. In such a limited internal market for military equipment, a balance had to be struck between spending to nurture strategic indigenous design and manufacturing capacity, and importing less costly off-the-shelf foreign weapons systems. The Korean War encouraged much higher defence spending than otherwise would have been the case. In the process, the balance tilted more towards military and technological self-reliance. With the end of the Korean War, the political will to maintain high levels of defence spending weakened. Thereafter, all military expenditures were put under the microscope in order to reduce any duplication of Canada’s development programs with those of its allies. The RCN had long justified the costs of the DATAR program on the grounds that neither the U.S. nor the U.K. had any comparable system to respond to Canada's unique anti-submarine role within the alliance. But recent U.S.N. decisions had put this argument on very slippery footing.
Until 1955, the U.S.N. had done relatively very little work of its own on developing an automated tactical information system. That year, the U.S.N. concluded that this technology was of paramount importance to its future effectiveness. Driven by a sense of urgency, the U.S.N. reached down into the deep pockets of American military spending and committed $10 million to a two year crash program to build its own supercharged version of DATAR, to be called the Naval Tactical Data System (NTDS). The Chief of the U.S. Naval Operations saw NTDS "as absolutely essential for the survival of a task force in the post-1960 era." American naval officials took the attitude that "in the future, if a ship did not have NTDS, then it should not leave the port."
Having the benefit of the RCN's seven years of pioneering work and backed with considerable financial and industrial R&D resources, there was little doubt that the U.S.N.'s proposed NTDS would triumph over DATAR. While the RCN had allocated $750,00 to Ferranti Canada to design and build a transistorized information processing and display system, the U.S.N. gave Remington Rand, whose UNIVAC division was a recognized world leader in computer technology, $4 million and Hughes Aircraft $3 million to undertake essentially the same challenge. Given the scope and magnitude of NTDS, the advocates of DATAR knew that their program was extremely vulnerable to the cost-cutting imperative to eliminate duplication. If NTDS subsumed DATAR, or if it was at least congruent with DATAR, then why throw away any more money on a large-scale prototype when the Americans were already working on one? The only strategy open to the Electrical Engineer-in-Chief's Directorate was to try to weave DATAR into the NTDS project in a complementary manner.
From the outset, NTDS was designed to serve the needs of a large carrier-based task force. Protecting destroyers, large battle cruisers, and aircraft carriers from enemy fighter planes was NTDS prime function. In designing NTDS, the U.S.N. had little interest in anti-submarine duty. Accepting NTDS as a fait accompli, the RCN argued that the U.S.N. system, as it was being developed, was not appropriate to Canada's anti-submarine mandate. The DATAR Sub-Committee recommended that the RCN use its expertise to embark on a development program to adapt NTDS to the needs of an anti-submarine navy.
With Cabinet Defence Committee pushing harder for more cost-saving cooperative ventures, the RCN tried to keep DATAR alive by proposing a joint Canada-U.S.A. collaborative effort to design an anti-submarine version of NTDS. The U.S.N. was not interested: all its resources were devoted to NTDS with nothing to spare for the RCN proposal. Offering Canada complete access to all the technical information on NTDS as it developed, the U.S.N. encouraged the RCN to continue on DATAR.
Faced with spending restrictions and little political support for an independent program, the steam ran out of DATAR. The Government's decision in 1956 to retreat from its earlier commitment to build the fourteen destroyer escorts also contributed to the collapse of DATAR. These proposed destroyer escorts were to have been the state-of-the-art in anti-submarine warfare. Not only was the design of the destroyer itself a dramatic improvement in speed and manoeuvrability, but these ships were to be equipped with the most advanced sonar, radar, weapons, and fire-control. Individually, and collectively, the proposed anti-submarine destroyer escorts were truly a large complex technological system. DATAR was supposed to be the electronic nervous system and brain that made all the parts of the system work together harmoniously. With the collapse of the program to build new fleet advanced destroyer escorts, enthusiasm for any further costly work on an automated electronic naval battle system also waned. Canada did propose a three year, $450,000 study to come up with suggestions for a Tactical Data System based on whatever NTDS developed. After seven years of daring pioneering work in real-time digital information processing, communication and display, DATAR was abandoned. Unlike the Canada’s AVRO Arrow which went out with a bang, DATAR died with not even a whimper, and along with it went the RCN's flirtation with military self-reliance.
This article is entirely based on a chapter from the work “The Computer Revolution in Canada: Building National Technological Competence”, by Dr. John Vardalas. For the readers’ convenience, the relevant references used in that chapter are included below.
James Lamb, The Corvette Navy, (Toronto: Macmillan, 1977)
W.G.D. Lund, The Royal Canadian Navy's Quest for Autonomy in the North West Atlantic, 1941-43, in James Boutilier (ed.) The RCN in Retrospect, 1910-1968, (Vancouver: University of British Columbia, 1982), pp. 138-57, p.139.
Marc Milner, North Atlantic Run: The Royal Canadian Navy and the Battle for the Convoys, (Toronto: University of Toronto Press, 1985)
Marc Milner, The U-Boat Hunters: The Royal Canadian Navy and the Offensive against Germany's Submarines, (Toronto: University of Toronto Press, 1994).
Norman Ball and John Vardalas, Ferranti-Packard: Pioneers in Canadian Electrical Manufacturing, (McGill-Queen’s University Press, 1994)
Lavington, Early British Computers, (Manchester: University of Manchester Press, 1980).
W.H. Pugsley, Sailor Remembers, (Toronto: Collins, 1948); G.N. Tucker, The Naval Service of Canada, vol. 2, (Ottawa: Kings Printer, 1952)
John Swettenham, Canada's Atlantic War, (Toronto: Samuel-Stevens, 1979)
John Vardalas, The Computer Revolution in Canada: Building National Technological Competence, (Cambridge, MA: The MIT Press, 2001)
Gerhard Weinberg, A World At Arms: A Global History of World War II, (Cambridge, U.K.: Cambridge University Press, 1994)
Primary Sources (By year)
(Note: DND refers to the Department of National Defence records located at the National Archives of Canada)
Minutes of the 24th Meeting of the Research Control Committee, Item 24-6, The Future of A/S Research, Item 24-6, 5 September 1946. DND, Accession 83-84/167, Box 139, File NSS-1279-33, pt. 2.
Minutes of the 24th Meeting of the Research Control Committee, Item 24-6, The Future of A/S Research, Item 24-6, 5 September 1946. DND, Accession 83-84/167, Box 139, File NSS-1279-33, pt. 2
Present Government Development Work Coming Within The Scope Of The Canadian Defence Research Board's Requirements, September 1948. Ferranti Corporate Archives, Manchester, England.
Eric Grundy, Memorandum to Sir Vincent de Ferranti: Research in Electronics Industry, Canada, September 20, 1948, Ferranti plc Archives, Manchester, England
Minutes of the Eleventh Meeting of the Electronics Advisory Committee, 23 November 1948, p. 6, DND, Box 4233, File DRBS 3-640-43
Arthur Porter, Preliminary Report, October 15, 1949. Ferranti Corporate Archives, Manchester England.
Captain W.H.G. Roger, Electrical Engineer-in-Chief, Memorandum to Chief of Naval Tactical Services, 9 January 1950. DND, Accession 83-84/167, Box 3108, File NSS-7401-382-1
Captain W.H.G. Roger, Electrical Engineer-in-Chief, Memorandum to Chief of Naval Tactical Services, 5 April 1950. DND, Accession 83-84/167, Box 3108, File NSS-7401-382-1
Minutes of the Research Control Committee, 17 June 1950. DND, Accession 83-84/167, Box 139, File NSS-1279-33, pt. 2.
DATAR Development Contracts, Minutes of the Research Control Committee, 14 April 1950. DND, Accession 83-84/167, Box 139, File NSS-1279-33, pt. 2
Deputy Minister of Defence, Memorandum to the Minister of Defence, June 1950, DND, Accession 83-84/167, Box 3108, File NSS-7401-382-1.
Chairman, Defence Research Board, Memorandum to the Minister of Defence Production, 25 July 1951, DND, Accession 83-84/167, Box 3108, File NSS-7401-382-1.
Minutes of the Research Control Committee, 24 October 1951. DND, Accession;83-84/167, Box 3169, File NSS-7407-382, pt. 1
W.H. Roger, Memorandum to the Director of Scientific Services, 13 December 1951, DND, Accession;83-84/167, Box 3108, File NSS-7401-382-1. W.H. Roger, Memorandum to the Director of Weapons and Tactics, 5 April 1952 DND, Accession;83-84/167, Box 3169, File NSS-7407-382, pt. 1
Chairman, Defence Research Board, Letter to R.B. Bryce, Secretary of the Treasury Board, 18 December 1952, DND, Accession 83-84/167,Box;3108, File NSS-7401-382-1
Director of Scientific Services, Memorandum to Scientific Advisor, Chiefs of Naval Staff, 6 December 1952. DND, Accession;83-84/167, Box 3169, File NSS-7407-382, pt. 1
S.F. Knights, DATAR Contract F.E. 109,069 P33-40-881 With Ferranti Electric, 19 January 1953. DND, Accession 83-84/167, Box 3108, File NSS-7401-382-1
Minutes of the Research Control Committee, 24 July 1953. DND, Accession 83-84/167,Box 39, File NSS-1279-33 pt. 2
Commander John Charles, Director of Naval Communications, Memorandum: Report on Visit to DATAR Demonstration, 10-11 September, 15 September 1953
Lieutenant Commander W.R. Howard, Report of Trip to DATAR Project in Toronto, 23 September. DND, Accession;83-84/167, Box 3169, File NSS-7407-382, pt. 2.
Commander A.R. Hewitt, Director of Supplementary Radio Activities, Report of Visit to Scarboro Bluffs, NRC Establishment, 30 September 1953. DND, Accession;83-84/167, Box 3108 File NSS-7401-382-9.
Commodore Lay, Assistant Chief of Naval Staff (Warfare), Conference of Demonstration of DATAR, 9 November 1953: Record of Proceedings, 13 November 1953. DND, Accession;83-84/167, Box 3169, File NSS-7407-382, pt. 2
B. Bromily, Memorandum : DATAR Tests and Discussions at Defence Research Medical Laboratories, 25, January 1954, DND, Accession 83-84/167, Box 3108, File NSS-7401-382-6
Department of National Defence-Canada Research and Development Project, 26 January 1954, DND, Accession 83-84/167, Box 3108, File NSS-7401-382
C.I. Soucy, A Survey of Electronics Equipment Failures Caused by Failures of Replacement Parts and Tubes and Improper Operation and Maintenance, 5 March 1954. Internal Royal Canadian Air Force report. DND, Volume 4185, File 260-640-43
Royal Canadian Navy Research and Development Estimates, 1955-56, 28 September 1954, DND, 83-84/167, Box 139, File NSS-1279-33, pt. 2
Naval Secretary, Memorandum to the Naval Member of the Joint Staff (Washington), 24 November 1954. DND, Accession;83-84/167, Box 3169, File NSS-7407-382, pt. 1
W.H. Roger, Memorandum to the Director of Scientific Services: Brief on DATAR, 12 January 1955, DND, Accession 83-84/167, Box 139, File NSS-1279-33, pt. 2
RCN Proposal For International Automatic Radio Message, 7 February 1955. DND, Accession 83-84/167, Box 3108, File NSS-7401-382-6
F.H. Sanders, RCN DATAR Sub-Committee, Report of the Sub-Committee Deputed to Prepare Recommendations to the Research Control Committee on DATAR, 26 October 1956, DND, Accession 83-84/167 Box 139, File NSS-1279-33 pt. 4
Minutes of the Research Control Committee, 20 December 1956, p. 3, DND, Accession 83-84/167 Box 139, File NSS-1279-33 pt. 4
Kenyon Taylor, Autobiographical Notes, February 15, 1970. Ferranti-Packard Archives. Provincial Archives of Ontario