10 - UNIVERSITY CALLS
by BERNARD A HODSON
After almost six years with Imperial Oil I received in the mail a letter asking me if I would consider an academic position at the University of Manitoba. They wanted to establish a teaching program in computer science and also upgrade their computer operations. Apparently I had been recommended by Dr. Scott at the University of Alberta in Edmonton, who knew of the teaching I had been doing at the Calgary branch of the University. Not having the usual paperwork associated with University professors, and being very happy and making progress with Imperial, I discussed the situation with my employer. Their reaction was very interesting.
They indicated to me that I was on their management development program (which of course I knew, having been on their Executive Course) and that I had an assured future with them However, they said, if you go to the University for a period, and then come back to us, your career with us will likely accelerate. While it may flatten out slightly (remuneration wise) at the University, when you return to us it will grow much faster. They suggested that at least I go to Winnipeg for the interview. Given a situation that was a win-win situation no matter what I did, I went for the interview.
I met with the University selection committee and discussed the issues quite frankly with them. I told them that I could offer an in depth knowledge of computers and management capability, but that I was lacking the usual academic paperwork, that I had a fair degree of research experience and was adept at compiler writing and generalised computer programming (the origins of GENETIX). This did not seem to phase them and I was offered the position of Associate Professor with a mandate to establish a teaching program and upgrade their computer facilities, which at that time consisted of an old Bendix G20 computer and an IBM 1620, used mainly by electrical engineering. There was no staff but I would be allowed to submit a business case for expanding the department budget.
On my return to Calgary I discussed it further with Imperial and we concluded that I should "give it a shot". I received very favourable letters from the senior staff at Imperial, congratulating me and wishing me well. Whereas normally, when anyone left, they left immediately because of "oil secrets" I continued with my position in Calgary until my departure.
Consequently we moved to Winnipeg. We now had four children in the family, three of whom had been born in Calgary. We purchased a house and began to settle in. The first advice we received was to "ignore the winter". The city gets quite cold in Winter, which lasts from October to May, at times reaching minus 49F, the coldest we ever experienced while there. Even at that temperature I would walk the mile to work, arriving with two inch long icicles hanging from my eyebrows, caused by vapour from my mouth becoming ice on my eyebrows. In such temperatures the water vapour crystallises in the atmosphere and it looks like fairyland when the sun shines on and is reflected by each crystal. Calgary used to get cold at times but there the Winter was ameliorated by the "Chinook" winds which could send the temperature soaring by 50 degrees within a few hours, so that minus -10F could become 40F between going to work and coming home. The reverse situation could also occur on occasion.
Being the first employee of this brand new venture by the University I set about giving a series of computer courses to the academic staff. At the same time I drew up proposals for undergraduate courses (initially in Fortran for engineers and mathematicians), and developed a five year plan for both the academic and computer facilities (I was responsible for both). The proposal indicated we should install a very large computer within two years, and explained briefly how it would be financed. The University accepted my proposals and so I set to work to implement them.
The Bendix G20 was of little academic use but had been used by a local engineering firm to design transformers. I decided that I would swap the G20 for a 750,000 volt transformer which the Electrical Engineering Department had a need for, and this transaction was duly completed. I then received approval to hire an operator for the IBM 1620, and advertised the position. One of the applicants was an interesting fellow by the name of Karl Schmidt. Karl had been studying mathematics in Germany when he had decided to help out in the Spanish Civil War. Being associated with the "wrong" side his studies had been forcefully terminated by Hitler, and he eventually immigrated to Canada. In Winnipeg, at the time, he was employed by Manitoba Hydro, in a position well beneath his apparent talents. I said to Karl that he was way over qualified to be an operator but that when we had an opening, expected soon, for a programmer, he would be considered. He asked for a copy of the programming manual for the IBM 1620 and, in typical German thoroughness, came back about a month later and pointed out scores of errors within the manual. I eventually hired him and let him have his head with the Mathematics Department. He eventually finished up giving joint academic papers with several academics.
One of my early projects, which received approval, was to establish a course of training in computer programming for blind persons, described in another section. I also established a number of undergraduate credit courses, initially in science and engineering but later extended to other disciplines. I then outlined a Master's degree in computer science, which was approved. This meant that I needed academic staff and I advertised in Canada and the United Kingdom, eventually hiring five academics.
The next requirement was to choose a computer and I decided that the only sensible thing to do was to install the largest computer in Canada, which would leapfrog us over the more established Universities such as Toronto, Queen's and McGill, and the newer University of Waterloo. It was ambitious but I believed it to be financially achievable. I proposed to pay for it in part by letting local industry have access to the system, as there were no large scale computers in the area, the closest being an IBM 7044 in Calgary, and that was fully used.
The main contenders were the UNIVAC 1108, the newly developed IBM 360/65, and the Control Data 6400, with Honeywell also proposing one of their machines. I went on a round of presentations and discussions in New York, Minnesota and Massachusetts, each vendor doing their best to show off the capabilities of their systems, and royally entertain me. Each machine could have served quite well so it eventually came down to who would make concessions to the University, who would best support its research effort. In the meantime, through organising a conference, discussed later, I had become friendly with Bob Fano, who was at the Massachusetts Institute of Technology (MIT), and heading a time sharing project based on Honeywell computers. He wasn't overly impressed with the IBM 360/65, saying it would likely be a maintenance nightmare. Despite this advice I finally chose that system, mainly because of what I personally considered its good capability, but also because of the attractive concessions I negotiated with IBM (In a sense both I and MIT were proved right, as events transpired).
Carl Corcoran was the Regional Manager of IBM at the time and later was Vice President of IBM Japan and President of IBM Canada. He told me several years later that they would have given even more concessions had I requested them, as they were anxious to install a major system in a Canadian University, indicating that my order had triggered at least three additional systems within the following twelve months. The salesman involved, Mike Burns, made so much commission on the sale that he took a year off to wander the world, later finishing up in Vancouver. He was a very tall and very affable individual, but he, Carl, and IBM senior management knew that if I wanted service it had to be good and delivered quickly (more of this later).
It was necessary that the system be in place before commencement of the 1966 academic year. This was so important to us that I negotiated penalty clauses for non delivery on time. It was also important that the computer "software" worked satisfactorily and I negotiated what was unique for the time, penalty clauses if the software did not work, scaled according to its importance, but with the total penalty value not exceeding the cost of the computer. Shell Oil Company, in particular, wanted to know how I had achieved such a clause, phoning me from Toronto to discuss it.
In June of that year IBM said that it would be late, but surely I would not hold them to the penalty. I assured them that "Oh yes, I would hold them so" at which point they asked if they could deliver and install an IBM 360/50 at no cost to the University. I agreed to this proposal and the lesser system was delivered in August, to be replaced two months later by the larger system. A similar large system was delivered to Canadian National Railways two days before ours was installed, otherwise we would have had, for a short period, the most powerful system in Canada. As it was we had to contend ourselves with having the most powerful University computer in Canada. In order to staff the system operation I had hired a group of operators and programmers, several of them from England.
In hiring academic staff I gave preference to those who had industrial experience, because I thought it would benefit the students to know that there were practical sides to what they were learning. At that time there was only the beginning of computer science programs and very few people with doctoral qualifications in the subject areas. Also, with non academic staff, I deliberately hired people with divers experience and qualifications. These included a Classics graduate, one with a degree in German, people from different industries, such as engineers, scientists and mathematicians. We also finished with a staff that spoke most of the main languages, including French, Spanish, German, Ukrainian, Russian, Arabic, Hebrew, Italian, plus one or two languages from Africa. This later proved to be very advantageous as visitors came from around the world to see what we were doing in research and in computation.
Once we became known, which did not take long, we started to receive master's degree applications from around the world, and I was able to fund their studies from the cash flow that I had generated through a variety of research grants plus rental of some of the computer processing time to industry. I also received many applications from academics wishing to do research at the University, and several joined us. One was Dr. Brodin, from the University in Paris (later in Toulouse) a specialist in aeronautical engineering. After he had returned to France and I happened to be on a visit, we walked over much of Paris, returning exhausted to the French Officer's Club, of which he was a member, for dinner. Dr. Carol Abrahams joined the University from CSIRO in Australia and did pioneer work on time sharing. He developed what was then known as "rollin/rollout algorithms" that enabled the computer to do several applications at once, rather than one at a time. Also joining us for a period was Dr. Jan Blatny, Head of Radio Engineering at the Technical University of Brno, in what was then Czechoslovakia. He returned to his country to establish a computer science program there, being appointed Head. In 1968 he invited me for a visit and I was due to land in Prague the day the Soviets invaded that country. A trip later that year to participate in the 500th Anniversary of Charles University, also had to be cancelled, because of the Soviet presence.
I encouraged all staff, whether or not academic, to do research related to computers, and non academic staff was given half a day per week to do whatever they liked. This resulted in a very strong and well motivated technical group, similar in many respects to what we had built up with Imperial in Calgary, with similar world recognition to many of their efforts. Our philosophy was that if we thought it could be achieved we would attempt it.
We did hit one small snag in the purchase of the computer when IBM said it had to be paid for in total over five years, rather than the seven I had requested. This meant either reducing the computer size or doing some creative accounting. We solved this by paying for it over five years and two days, the first payment being on the last day of the first budget year, the second payment during the next year, the sixth payment being during the fifth full year, with the final payment being the first day of the seventh budget year.
By today's standards, even though it was the most powerful system in Canada, it had very little memory (128K words initially), several magnetic tape drives, a card reader, high speed printer, and several disc storage units with limited capacity. However, despite its hardware limitations, we used our intellect to make this computer do whatever we wanted to be done. Apart from graphics, which was too expensive at the time, we were told that the things we were doing with the resource we had, was only matched by NASA. We also had lots and lots of time sharing between applications, as well as doing day to day processing for scores of students. A chapter will be devoted to describing most of the research attempted but a few early examples will be illustrated here.
The Physics Department had installed an atom smasher called a Cyclotron which, when it was operating, generated reams of data which had to be quickly reviewed so that the Cyclotron could be controlled. They had a display computer for the system from Digital Equipment Corporation, called a PDP9, which presented some graphics on a screen related to the Cyclotron operation, but which was not powerful enough to do the heavy calculations needed. I was approached to see if there was some way we could do the heavy processing. I agreed provided that it did not interfere with student processing.
We came up with an innovative solution in which the PDP9 would, while the Cyclotron was running, strip extraneous data from its output and transfer a data stream to the IBM 360/65 for heavy processing. The more powerful system would do the heavy computation in parallel with its other work, and transfer the results back to the PDP for display, enabling the scientist to modify their experiment in "real time" (i.e. while the experiment was running). IBM were reluctant to have "foreign" computers attached to their "computer child", citing all sorts of problems we might experience with warranty violations, so I asked them who owned it, and was assured the University did, so I politely told them to "get lost".
Later I added two Control Data computers at the Medical School, attached over phone lines, and later on still some IBM remote entry equipment located at Pinawa, about seventy miles away, which was home to the Atomic Energy of Canada Limited (AECL) experimental nuclear reactor. I was invited to tour the reactor site before it went critical, one of the last human beings to have actually been in that location. I was also asked to teach their scientists the Fortran language, which I did once a week This meant driving out in the depths of the harsh Manitoba Winter, where temperatures fell to below -40F, making sure I carried survival equipment in my car, in case of a breakdown at minus 40F in a howling wind miles from any help (there were no cell phones in those days). I almost had to use it on one occasion when a vital part of the car engine broke, but I was able to slowly clank my way back to Winnipeg, where I sold the car for scrap.