My leave of absence from Boeing was up by April, 1965, by which time Regnesentret was a going concern of about 25 people, with a giant Univac 1107 on its way (so we thought) from St. Paul, Minnesota, and computing was spread all over Gløshaugen and down into the town. That was quite an explosive two years, and perhaps before all the details are finally confined to the oblivion of my rapidly fading memory, it might be of interest to one or two of the old timers to record some of the events of those years.
I don't think, however, that the youngsters of today are a bit interested. Computing today is as available as the air they breathe. It comes as some God-given right. It's literally at their finger tips. And this is as it should be. It's what we were striving for right from the start. But I think a great deal of the fun was in getting it that way. And I don't think that many people today know what a computer is. Indeed, the irony is that the larger and more complicated the computer the simpler the things we do with it. A secretary couldn't have used the EDSAC to type a letter, but it helped discover the shape of DNA. Language facilities and software generally are so powerful today that hardly anyone needs to know what a bit is, a word, an instruction, an exponent, a floating ponit number, an index register or an overflow indicator. And rightly s~. But there goes the fun. But why should computers be fun? They're too expensive. Nevertheless, the Golden Age was fun.
So how do you start? You have something that looks like a broom closet, a secretary and a single computer man just about to call in the movers. You don't speak a word of the language. You haven't a clue how NTH works. And anyway, the whole thing is being organised by something called SINTEF for NTH because the latter hasn't the power to do it itself. You have no budget apart from the secretary's salary and half yours. There's no provision for any courses on the timetable. All input is via papertape, i.e. every program change means an entire recopying of the tape instead of a one-card insertion. There are seven Flexowriters at which sit seven users making seven times seven errors per minute. The professor of Numerical Analysis is a German and his English is equivalent to your Norwegian. There is no printer. Output is made via punched tape, printed later character by character on the Flexowriters. But something called a carrousel has been ordered from Denmark consisting of 128 minimagnetic tapes on a rotatable wheel, highly reminiscent of a fairground. And Algol to you is as a red rag to a bull.
We have arrived at the fairground.
However there are two glimmerings of hope. One is that no one actually expects anything, so no matter how much you muck it up, no one will ever know. An secondly, here and there in this and that niche were people who had already acquired consid erable experience on the com puters. If we could bring about a balanced hardware configuration, close the shop and give some courses, we might, in five years, develop ourselves into a computing centre.
A mong the primae val users of GIER (as the machine was called) were the structural engineers in Professor Holand's institute, Professor Borgen's chemists (who used to sleep on the computer room floor) and Professor Balchen's control engineering boys. Indeed the latter had been largely responsible for building the machine at Dansk Regnesentral in Copenhagen. And here I would like to pay an especial tribute to Nils Ivar Bech whose generosity and helpfulness make it possible to acquire the computer. In a little while we shall find out that GIER was a far smarter move than I originally thought, blinded as I was by the vastness of the 32 K of the IBM 7090's that I'd left behind at Boeing.
The first thing to do seems to be to put the actual hardware in the hands of fulltime operators. This is the only way of getting any efficiency out of it. But efficiency doesn't mean much at first on the computer itself, because so little work is reaching it because of the vast incompetence of the users at the keys of the Flexowriters. So step one is to hire some secreatries to man the Flexowriters. But you can't do this. Remember, you have nothing to pay them with.
Another ray of hope, you can have night shift, and if you can earn money on it you can hire staff. But how can you earn money if you don't have any staff? At any rate, please let me keep Knut Skog. So Knut stayed. I remember he was on Kr. 32000 per annum, but where the money came from I simply don't recall. Probably Karl Stenstadvold sold a violin.
Step zero, then, is to bring in some work from the big wide commercial world. But how do you talk to people about their computer problems when they don't know what a computer is? How do you talk to the blind about colour? But somehow we managed it, and the roll of honour of those two first years reads somewhat as follows: - Pølsemaker Helgesen, Norges Fiskarlag, Norges Geologiske Undersøkelser, The Institute of Statistics at the University of Uppsala (we went international right from the start!), Norsk Brændselolje, Kristiania Portlandsement fabrikk, Elkem, Fjellanger, Sydvaranger, Felleskjøpet and several others who have now escaped my memory. Think of it! They actually trusted us and our computer. They must have been crazy. But we delivered the goods, on time and within budget. We were probably the only university computing centre in the world in the service bureau business. And we were at least as good as any commercial outfit. But without those wonderful customers there would never have been a Regnesenter as we knew it. The goverment didn't put up the money; the money was obtained from individual private companies and institutions as payment for damned hard work well into the small hours. Fortunately we didn't have the Arbeidsmiljølov in those days destroying our enthusiasm and preventing us getting the job done.
All this didn't happen over night, and hiring people went hand in hand with getting jobs. At first three keypunch girls. Immediately a wave of work starts hitting the computer and the bottleneck moves to the priner. It is now Easter, so we fly down to see Bech in Copenhagen. Help! Let us have a highspeed priner with the carrousel money. You can't do that, it's an official Norwegian government order. O.K. send us up a printer, but bill us for a carrousel. Done!
When we get back there is an outcry. Poor old Professor Romberg. He's set his heart on the carrousel. His pockets were to be bulging with mini tapes. But this can never be. With all that data in the computer and no way of getting it out it would have been a case of data constipation without equal. When the printer arrives and 20 characters per second becomes a thousand lines a minute all doubts are dispelled, the ranks are closed and the government will never know.
But now we must start to squeeze some efficiency out of the main frame. We train the girls to operate the machine and lock the door for hours per day. No one seems to complain though. Probably the novelty of actually pressing the wrong buttons yourself has worn off in favour of someone else pressing the right ones. Turn around time is pretty short. The error rate is down. Why, the keypunch girls, with no Algol training, even spot errors in the original coding sheets and correct them while punching. You don't get that service today with your on-line terminals.
And so we start to hire the boys. In addition to Knut Skog, today professor at Tromsø, Karl Schjetne from Kjeller, later to be director of RUNIT, Harald Michalsen later to get his Dr. Techn. on computer school timetabling; Nils Høeg, now Managing Dirctor of Hydro Data, and Øivind Solvang, today Managing Dirctor of Guru Papp, both from Boeing; Pluto Østlie, NTH's first candidate to write software for his diploma (and what a fight it was to prise him loose from his NTH institute!); Ralph Høibakk, founder of NorData and later Managing Director of Tandberg Data; Tore Bough-Jensen, today a senior consultant with Habberstad; Arne Sølvberg, today Professor of Databehandling at NTH; Olav B. Brusdal who kept the machine going, today Assistant Professor at NTH, and part-time third shift student operator Kristen Rekdal, later Associate Professor with SINTEF. That's a pretty star-studded team to have stacked arund the computer itself, in addition to which we had a rapidly growing user group typified by people like Johannes Moe, then Professor at the Division of Marine Structures, later Rektor of NTH and now Director of SINTEF, and Øivind Bjørke who combined mechanical engineering with control engineering and who is now Professor of Industrial Automation.
But there were many others, and it becomes impossible to name even a worthwhile fraction. Once the stopper had been removed from the computer djinne bottle star performers started pouring out.
A nother component of Regnesentret's early success was courses, both to NTH and the outside world. At NTH some of the older members were a little reluctant to dive in, but the students weren't. But we needed a teacher, and I'll never forget my first meeting with Jim Tocher who came limping out of the lift on crutches, his left leg encased in plaster, his face covered in a red beard - years before beards ever hit Norway. Three weeks later both the cast and the beard disappear. Conclusion - Jim Tocher shaves with his left foot. He's just finished a post-doctoral year with Professor Holand and is about to return to Berkley. I promise him a good job at Boeing if he will stay on six months and do the teaching. He's still at Boeing twentyfive years later, and the teaching that first year was a revelation. Actually, the first programming course was given already in the autumn of 1962 by Aimar Sørenssen who later left for C.E.R.N.
There was no place in the time table for teaching programming, so we announced a spare-time course, three times a week for eight weeks. We expected twenty people to show up; 250 came. When that was over Knut Skog said that that was it. There will be no more interest this year. Well let's find out Knut. Waste of time. Well let's at least offer it, and if twenty show up we'll run a course. 150 turn up. That's absolutely it, says Knut eight weeks later. Come on, let's try once more. Look, I know NTH. But there's no harm in asking; if twenty show up we'll run the course. As we approach the lecture room where the initial meeting is to be held there is a large crowd outside the door. Damn it, they've forgotten to open the door. But no. This is the overflow; 450 people have turned up. Getting on for one thousand people learn to program the computer during its first year at NTH; all in their spare time. One of the professors complains that the students have deserted his lectures for the computing centre. They don't even drink any more, he says. GIER is better than beer observes Karl Stenstadvold.
Two other courses were held form the start, Brusdal's in hardware, which became the first ever credit course in computing at NTH, and a general awareness course called something like, "How to use the computer".
In addition to the internal courses, we started giving courses to practicing engineers, and that became our most successful way of getting new customers.
No one anticipatec the explosive growth of computing at NTH SINTEF, and the GIER computer was probably intended to see the campus through to the 1960's. But already by the end of 1963 we could see GIER getting jugged by 1965, and the long, frustrating process of getting a replacement started already in January, 1964. Step one is to get the money from the Department of Education. At first there is absolutely no question of a new computer for five years. So we go down to Oslo loaded with paper - the list of project numbers and titles. The first thing we ever did in January, 1963 was to set up a project number and accounting system. We know very accurately who is doing what on the GIER, and when we open our bag and dump our statistics on the official desk the day is won. The only question is how much?
One of the problems that beset us was that there were already two enormous computers in Norway, the Univac 1107 at Norsk Regnesentral in Oslo, and the CDC 3600 at Kjeller. We must be compatible with one or the other. To get an IBM computer and present Norway have been an act of criminal folly. So which should we get, a 3600 or 1107? There were two factors that made the decision, one was that we managed to talk Univac into an 1107 for half price, and the other was that the Exec II operating system was admirably suited to large numbers of small jobs, and it was that that characterised NTH computing at that time. So the 1107 selected itself, but not after a great deal of toing and froing involving half a dozen computing companies, all of whom had seen the desirability of getting their machine in at NTH where tomorrow's decision makers were cutting their computing teeth. And that is why, a generation later, it is still Univac at NTH, and partly why Univac grew so strong in Norway.
As a post script, and to demonstrate to today's users that computers didn't come quite automatically, when I returned to Boeing 1st May, 1965, a telegram was awaiting me from Nils Høeg saying that Univac had cancelled the 1107 because it was half price. A new man had taken over the Univac company, General Baker. A tough guy. A new broom. And his first action was to sweep away all unprofitable actions that the company had undertaken. And ours was the most unprofitable. So I had to go halfway back to Trondheim to rescue our 1107. As I said at the start, there's no fun any more. The most fun I've ever had was rescuing that 1107. It's all very well having learned theories about acquiring computers, but it's what goes on behind the scenes that maters, and I'd like to pay another tribute, this time to General Baker whose fairness and generosity made it possible for NTH to aquire a large, modern computer.
There must be thousands of anecdotes of those times, and I'm looking forward to Olav B. getting out his old log book and refreshing our memory. However, one that is ever vivid in my memory is the way we put one of the Trondheim banks through the wringer. They hired me as a consultant, (but not for long). I didn't know anything about banks so I started to find out. To help the process along a bit I started signing my cheques Charlie Brown, and so did the rest of Regnesentret. I don't know how many Charlie Brown cheques were cashed, but not a single one was either refused or caught by the auditor's department. Not all problems are computer problems!
We had our frustrations of course. One thing that people perhaps aren't aware of is that the 1107 at Norsk Regenesentral was a world pioneer in remote computing. At one time there were four terminals simultanously coupled to it over telephone lines, one in Sweden, our 1104 in Øivind Solvang's office, and two others in other parts of Norway. The technical problems were solved, but the problems Øivind had with the telephone operators! They'd listen in suspiciously after two hours of silent data transmission, and he'd have to start all over again. And that brings me to my third debt of gratitude, to Leif Olausen who ran Norsk Regnesentral in those days. His energy, drive, enthusiasm, and the tireless way he supported us in Trondheim until we could stand on our own feet deserve our deepest appreciation. He knew how to inspire a bunch of people and get the last ounce of effort out of them. No arbeidsmiljølov down there either.
The seed that initiated the Regnesentret crystal was undoubtedly GIER itself. It had a core memory of only 1K. It was a microcomputer in a broom closet. But it had two vital properties, it was a highly reliable piece of equipment and it had an excellent compiler-cum-operating system. I.e., it was both hardware and software. And that was a lesson that the Norwegian computer industry still didn't understand a decade later. A computer consists of its software. Had programming been tedious or had productivity been low because of downtime, computing would have got a bad name right from the start. But Algol is a good language, despite an article appearing in Datamation, January 1963, and, as a result, computing took on like wildfire from the first day. GIER's task was to convince NTH - SINTEF that they needed a computer. No computer that I know of could have achieved that goal more admirably.
Those were the good old days, of course. In particular, they were noisy. Mostly laughter. Today as you tiptoe through RUNIT all is solemn silence. I suppose they're far more productive than we were, but I doubt whether they work with more enthusiasm, and I hope I don't hurt anyone's feeling if I opine that we worked far longer hours. At any rate in those days people put in enormous amounts of time, and the reason they did so was that it was tremendous fun. We had an esprit de corps. We had a feeling of being wanted. If any politician had come in to try to shove any arbeidsmiljølov down our throats she'd have been kicked smartly down the stairs. Fortunately, in these days people were still being treated as adults. Today Aunty Norway wags her schoolmarmish finger at her children and denies them their most fundamental sacred right, the feeling of being wanted.
I'd like to end by paying my fourth, and last debt of gratitude. Without the unceasing, total, committed support of Karl Stenstadvold, then Director of SINTEF, Regnesentret, NTH, would never have come about. In theory Regnesentret was an impossibility, but Karl Stenstadvold is a man who makes the impossible possible. One of the very few. You can write your academic books on management, then chuck'em on the fire. They aren't worth a fig. They've nothing to do with managers, at least managers of the calibre of Karl. I've known hundreds in my time, but none like him. He's the best of all my memories of Regnesentret, NTH.
That's all memory. Now a greeting to the jubilanter. The whole point of pioneering anything is to provide the following generations with something of value. We've gone from GIER to CRAY in one generation, a factor of 100 million speed-size units. An unimaginable growth. You've inherited something close to a miracle. I hope you use it well. I hope it's all been worth the effort.