4 - DEFENCE INDUSTRY
by BERNARD A HODSON
Moving to Canada I became a 'systems engineer' with Avro Aircraft , which had designed a fighter aircraft called the CF100. It had also designed one of the first jetliners, called the CF102 (but generally going by the name of Avro jetliner), and was busily designing a supersonic plane designated the CF105, known as the 'Arrow'.
The job I had was initially with three other individuals. We worked behind locked doors so that nobody could see inadvertently the work we were doing, which was, initially, to evaluate different missile systems for the Arrow. We had access to all sorts of documents labelled secret. One of the studies I was involved with had to do with proximity fuses on warheads and some of the documents were from before the first world war, still labelled secret.
The designated targets were, of course, Soviet planes and we had to become familiar with the different types. These were given code names such as the Badger and Bison, and we needed to know their speeds, what manoeuvres they were capable of, their armament and so forth. In turn we had to become familiar with the characteristics of the different missiles we might use on the Arrow, the two main contenders being the Falcon, built by Hughes, and the Sparrow, built by Douglas, but we also became familiar with several other missile systems.
The characteristics (speed, manoeuvrability etc.) of the plane and the missile determine what is known as the 'probability of kill', assuming that everything works, which is by no means certain and has to be taken in to account. The radar systems in the missiles had a forward 'window' of sight so that the plane had to be within that window to be shot down. The situation would change from minute to minute, as the plane took evasive measures and the missile responded to the manoeuvre. Our job was to determine the 'probability of kill' when the Arrow had different missile systems, against different target aircraft. The propaganda of the day, of course, was that missiles would obsolete the bomber but in point of fact, even if everything on the missile worked (the probability of this happening was considerably less than 100%) there was still only a relatively low probability that it would shoot down the target.
Our first approach was to buy models of the target from the local toy store, and construct a gadget that enabled us to place the missiles, with their appropriate forward window, in different locations relative to the bomber, then assess whether the attack would be successful. In a sense we were 'simulating' a variety of missile/aircraft engagements to determine the probability of any one missile shooting down any one target, assuming initially that everything worked, but later introducing a 'doubt' factor based on test experiences with the missile in question. The simulation was crude at best.
The Aerodynamics Department, headed by Chamberlin, who later handled the Gemini project in the United States, was using a recently acquired computer called a CRC102A, developed by a firm in Minneapolis. This machine had a paper tape telex system for input of computer instructions and data, and for output of results. It worked at a few characters per second in both input and output. The memory of the computer was limited but adequate for many calculations, using a cylindrical drum containing the coded data and instructions. By today's standards it was very, very slow. With the early drum systems one had to know the speed of rotation of the drum and place instructions for the computer around the drum, not in physical sequence but one which enabled the computer to pick up an instruction without having to do an extra drum revolution. It meant knowing the time for processing each instruction (add is quicker than multiply, etc.).
I suggested that the computer might be useful in our simulation work but Chamberlin could not spare anyone to do the training, at which time I asked if I could use it, if I learned how to do it myself, to which he agreed. Borrowing the manual I learned how to program that particular computer. Most computer manuals today are very poorly written and leave a lot to the understanding but this particular manual was written so that you could actually learn how to program the machine. Having mastered the technique I then wrote a program that would simulate an aircraft/missile engagement under various conditions and determine the probability of kill for different types of missiles with the Arrow attacking different targets. This was the largest program that I have ever written, even to this day and, as far as I know, was the first ever use of a computer for simulation of any kind.
Having no graphical capability at this time (it took several years for this to develop) all results had to be printed and then plotted by hand. Nevertheless it saved us hours and hours of manual work. I then taught my colleagues how to program the computer and our section was in business. The program, however, took several hours to run, simulating a whole variety of situations, so I obtained permission to run it overnight, starting the program at evening quitting time. The odds of any stranger knowing what was on the output tapes was virtually nil, and overnight staff had security clearances.
We then had the installation of an IBM 704 series computer. Northrop Aircraft, in the meantime, had developed what they called a FORmula TRANslator or FORTRAN. This enabled us to write applications in a more scientifically oriented notation, provided that the established rules for the language were kept. You could then write readable statements such as A = B + C. This notation had to use what were called 'compilers' which would go through several steps before generating code that could be used by the computer.
Reliability was a problem in the early days when most computers were built with a series of vacuum tubes or 'valves', and later with transistors. (Today unreliability is the plague of software rather than hardware). I remember visiting the USA Rand organisation (a strategic think tank) in California where they had two large IBM computers linked together. A demonstration was cancelled because one of the console typewriters had broken down, disabling both computer systems.
During thistime I was sent to a course at a University in Cleveland. There, three professors, Churchman, Arnoff and Ackoff, had written a book 'Introduction to Operations Research' and were trying it out with us. We learned about things called linear programming, simulation (which I knew about already), inventory modelling, queues and assignment problems, sequencing and replacement models, zero sum games and other things, some of which will appear in more detail in later sections.
The company was building a 'flying saucer' which really looked like a saucer, and like the things we now call UFO's, but I did not get chance to work on it. Instead I spent some time designing a ballistic anti-aircraft missile with no guidance (in some ways very similar to what was developed in the late 1900s by the US anti ballistic missile program). I also became familiar with the North American defence system that used a 'Sage' computer. Again, to hear the propaganda, you would think it would keep the people of North America safe from attack, but in reality, (even if it happened to be working, by no means a certainty) it could not handle more than 100 objects in the air at any one time and the early warning systems with which it was associated were so unreliable that the probability of any hostile aircraft getting through was quite high, more so if it came in at low altitude.
One of the proposed defence missiles was called the Bomarc, another unreliable piece of hardware that 'perhaps' could get off the ground, if all other systems were working. I worked out that if all the Bomarcs purposed to be stationed in Canada were fired at an enemy the probability was that we would shoot down one and no more than two invading planes.
Our recommendation, approved by the Royal Canadian Air Force, was that the Douglas Sparrow missile should be installed on the Arrow, and design work on that configuration had taken place. The Arrow was, and would still be today, one of the most beautiful and versatile aircraft ever built. One weekend, however, a group of managers went down to California and, when they came back, had decided to change to the Falcon. This was an inferior missile and there was no technical reason for moving to the Falcon. It had the effect of delaying the first flight of the Arrow by at least a year (as much of the electronic systems had to be reconfigured) and of course escalated the costs, leading later to the program cancellation.
Apart from the tragedy of losing the Arrow (and replacing it with totally useless Bomarc missiles) Canada at that time had designed the most powerful engine in the world, the Iriquois, which was to have powered the Arrow. That engine, at the time, was achieving 18,000 lb thrust, while the British were only getting 12,000 lb and the Russians 10,000 lb., the USA at that time weren't even in the game. Aside from the fact that Canada had designed the best aircraft in the world Canada could also have had the world lead in engine design.
There was a great deal of bamboozling the public at that time. One was in the myth that we were well defended from attack. I had a long dialogue with the late Basil Dean, editor of the Calgary Herald. He had been on a boondoggle to review our 'glorious defences' in the North and then wrote a newspaper article saying how wonderfully we were defended. I pointed out to him that his article contradicted that assumption when it referred to the fact that planes had to be at a certain height to be seen. His first reaction was to ask how an ignoramus such as me could question our glorious defenders. I told him I had worked on North American defence systems and explained to him the problem of radar 'ground clutter', following which he shut up.
Another public obfuscation was the idea of public evacuation in the event of an atomic attack by the current potential enemy, at that time, the Soviet Union. It was suggested that major cities should be evacuated quickly in the event of a nuclear threat. To hear them speak you would think that a missile with an atomic warhead aimed for Winnipeg would in fact hit Winnipeg. American missiles at that time were quite inaccurate (inertial guidance systems were just beginning to be introduced) and the Soviet missiles were considered to be even more inaccurate. There was an equal, if not greater, probability that the missile would hit the evacuated area rather than the city to which it was supposedly aimed. The best approach would have been to have a fallout shelter with an adequate food supply, but that would have cost money and evacuation was cheaper, even if those evacuated might be subject to a direct hit. But even without a direct hit their exposure to radioactive fallout would have been much higher than being in a shelter, let alone an evacuation during a Prairie Winter at -40C.
While at Avro I took two graduate courses, one in Advanced Aerodynamics, the other in Advanced Airplane Design, both offered by the University of Toronto Institute of Aerophysics.
Computer security was also something that needed close attention. Computer programs, when running, emit electronic signals which can be picked up by a receiver and analysed. In consequence any sensitive computer installation has to be protected by shielding to prevent unwarranted acquisition of those signals. The room in which the computer is housed must be 'signal proofed' and in many cases what were known as Tempest terminals had to be installed, which were shielded devices for data input and output.
The USA and UK have sophisticated listening devices, and even Canada does its share. Listening devices were located close to the Soviet embassy in Ottawa and they, in their turn, had known 'listening trucks' at strategic points.
Even electrical outlets and cables had to be protected from signal transmission. At one time, a few years later than this section, I was recommending a secure computer system that would be attached to another secure computer system several miles away. I suggested the use of a 32 channel optical cable to transmit data randomly on any one channel, making it virtually impossible to tap the data, instead of using the very slow encrypting systems then available. While I am sure my suggestion was viable I was shown how optical cables can be tapped without it being known. My suggestion was turned down.