Century of Science; The paperwork revolution

This is a giant electronic computer. One of the sounds of science in the 20th century. WGBH FM in Boston presents a century of science produced under a grant from the Educational Television and Radio Center in cooperation with the National Association of educational broadcasters. This is an exploration of developments in 20th century science and of the implications they present for contemporary American society. Your host voted Tory former editor of Popular Science and now director of radio television programming for the Massachusetts Institute of Technology. Such events are sending satellites around the earth and predicting their orbits would still be dreams if ways had not been found to make mathematical computations extremely fast. Big electronic digital computers are now being used in nearly every kind of scientific research. Business concerns are using them too for a great variety of purposes. Professor J.W. Forrester helped to build one

of the first of these high speed digital computers. It was called whirlwind as head of the digital computer division of the Lincoln Laboratory of MIT. He also helped to develop the sage defense system. He introduced a magnetic core memory system which has since been used in nearly every big high speed electronic digital computer. Professor Forrester was educated as an electrical engineer but is now in the MIT School of Industrial management. The newspaper's Persis Professor forester and calling these machines that you've worked with so much giant brains can they really relieve us of any thinking not of greedy of thinking but they relieve us a lot of a lot of the tedious computation that would otherwise need to be done by hand. More and more we're beginning to realize that. A lot of the work which we

ordinarily call thinking is carried out according to very simple or a straightforward rules. And if one happens to know the rules then he can instruct one of the present day machines to carry out the work instead. The trick then is to instruct the machines perfectly. Yes the machines have to be told in detail every thing that they are to do. They have to be instructed as to how they are going to perform a series of computations one must be able to sit down and do the computation by hand before he knows how he would instruct the machine to do it. So in order to use one of these machines you actually have to know the shorthand of higher mathematics and reduce it to simpler mathematics is that true with that. Yes in general these machines I add subtract multiply divide to the ordinary. Steps of arithmetic and anything that they accomplish is made up of a series of these earth

magical steps and choices that they make on the basis of rules that you have set down for them. Well when and how did you first become interested in this sort of thing in these big computers. It was perhaps by chance it was a normal evolution out of some of the previous work that I had been doing during World War Two on the control of military gun mounts and radar and we came to a problem that was too difficult to handle by the methods we used before and were led into the digital computer field as a way in which we could solve some of the complicated problems in military research. And this led gradually to the engineering civilian and commercial applications of the machines. What can you tell us about the sort of evolution of these machines. Well I think we can divide the evolution

of machines of this type into five year periods fairly nicely. And if we say that the period from one thousand forty five to one thousand fifty was the period in which the experimental work on the machines themselves was done we found out how to make electronic circuits that would operate a million times a second and hire with found out how to make high speed storage devices for remembering information inside the machines. We found out that machines with 10000 vacuum tubes in them could be made to operate at all these tubes would indeed operate at one time and we could get reliable equipment. So the period from 145 to 1950 was the period of electronic development. In fact the technological progress. Is a great deal more rapid here than in most other scientific fields. The machines of the type that we have today have improved in their performance by

almost a factor of 10 every year. This means that they have gone up in speed by a factor of 10 or in reliability or in storage capacity by nearly a factor of 10 per year for the last 10 years this means that the total rate of progress in the performance of these machines the total progress is greater than the step one sees in going from chemical to atomic explosives it's truly a tremendous step forward in machines of this kind. From 1950 to 1955 was the period when the pioneering was in how to use machines for engineering and scientific applications this was the period in which. Problems and aerodynamics problems and heat flow problems in gas turbines in atomic energy were being put on machines for the first time. We found out how the machines could be applied in scientific problems now that's going on that's

expanding all the time but it's no longer the frontier the frontier now from one thousand fifty five to 1960 is in the application of these machines in commercial problems to bookkeeping filing payrolls accounting the kinds of clerical activities that go on in our businesses. From 1960 to 1965 the five years coming up very soon I think is the time when we will see these machines applied to what we call real time control systems. By this we mean a system in which the machine is participating in the control of what goes on we have an example of that out in the sage air defense system where the machine calculates the instructions for fighter airplanes as to where they should fly. We will have experimental work between one thousand sixty one thousand sixty five in applying these machines to

civilian air traffic control. We will see application of devices of this sort in chemical and petroleum process in the strays. Will see it in machine tool industry so that this is a field in which the machine controls activities that are going on at the moment. The troll is executed then from Nineteen sixty five to nine hundred seventy eight. Is the period when all of these will be brought together in their application to the management process itself and that sort of thing. I myself am interested in so the research we're starting now is research aimed at a period of application beginning in 1065 and running on out 10 15 20 years beyond that. These machines is binary numbers where you come back to what people refer to as boolean algebra. Where are these used for other

purposes before Professor first or is this all knowledge that suddenly applied or was this fairly new knowledge. The binary number system and Boolean algebra is very old and has no great significance really relative to today's machines the machines operate on the binary number system simply because it is an engineering convenience. It doesn't introduce any new philosophical. Ideas relative to the machine and a binary number system binary counting system is very common in one's everyday life. For example a light switch is a binary device that is either on or off it's also a memory device it remembers whether you have left it on or left it off until you come and change it in a way this is entirely comparable to the basic circuits that go into today's digital computer. Is this going to throw a lot of people out of work or create more jobs or what. If this is going to take over a lot of I take it you're referring to white collar work. Millions of people do.

That's right but I think we have to bear in mind that the machines for. Mass production whether it be the production of goods or of numbers come as the result of a demand they come because we wish to do more of this work than we have people to accomplish. And in general machines of this sort have made more work for people rather than less just as the telephone company has a tremendously greater employer of people now with all of their automatic telephone exchanges than they were in the times when telephone exchanges were strictly manual. As we develop new machines we develop a greater demand for the outputs so that in fact as many people are used running the machines as were previously used to doing the work now done by the machines. Also we open up all the jobs connected with the building in them and tenants of

such machines so that in general. The demand for human activity is greater. What is different and usually of a more interesting type well-liked are likely to be some rather serious for individuals temporary dislocations though for instance if a big insurance company or any other kind of business suddenly puts in one of these machines and succeeds in using it to do their work it will say a hundred or a thousand men and women did formerly with hand machines and the other way are a good many of those people going to lose their jobs. This has not been true so far. So far the organizations that have put in such machines have undertaken to do more than they did previously and in general their total employment has gone up rather than down. Although there can easily be exceptions to this any new major technological change can create local cases of hardship in the short

run. So this this is possible here but in general the reverse has been true even in the short run even in the local situations. The shifting patterns of employment have not been much more drastic than occurred normally by people leaving seeking new jobs or retiring and so forth. Do you expect these computers to continue to become bigger and faster. Or have we sort of reached a leveling off point in their size and capacities nowadays are they big enough for most jobs now. Well in fact there are many applications that require machines considerably greater performance than any that we have today. So the development of machines the technological progress is certainly going to continue very rapidly. But at the same time the machines have moved ahead of our ability to apply them in many of the commercial and industrial applications so that the present.

Technological performance far exceeds all really knowledge of how to use the machines in many of the applications so that at the same time that higher speed machines are being developed we're going to find a great deal of activity especially in business devoted to the question of how we should use those that we already have. Is that what led you from electrical engineering into industrial management or how did this come about. I went into management myself because the background which I've had in engineering seems very well suited to solving some of the major problems that exist in the understanding of industrial activity and the understanding of our economic system and how it is. There are three things from engineering that I have. Participated in over the years that I think will have a major significance in the management field. One of course

is the field of computers itself that we are discussing. Another is our increasing knowledge about decision making and judgment and thought and our knowledge of the parts of this process of thinking which indeed are repetitive and the parts which must continue to be intuitive creative judgment. Knowing more about where the boundary dividing line exists between the is we realize that a great deal of the present routine work. That is cold decision making can indeed be established on the basis of sound rules. So we have a knowledge of decision making that we haven't had before. And third and most important of the three is the growing realisation that everything that people do can be characterized by what we call

feedback information systems feedback control systems and by this we mean a situation in which you make a decision to change something which in turn affects the circumstances that led to the decision. Now we have a feedback control system when you try to stand vertical as you sense yourself beginning to fall over. Want to just as balance so that he can continue to stand vertical as one drives an automobile. One observes the street corrects the steering wheel corrects the position of the automobile. These are all information feedback systems. So in our social economic and industrial activities we have information feedback systems in which one place is in order to build up an inventory so that he can deliver goods. One invests in a new plant so that he will have. Production capacity to meet the demand led to the

decision for the new plant. These are all information feedback systems and they have very special characteristics. They have the characteristics of potentially being unstable or fluctuating because they have time and they have distortion or amplification in the information channels. And we've come to understand this behavior a great deal better over the last 20 years and the application of this to the field of economics and industrial management is an entirely new field it's going to have a tremendous impact on our whole economic system over the next 20 or 25 years. There must be some limitations and there's a little tricky there when you talk about decision making and feedback way. It occurred to me when the decision to get married is certainly an important decision and there's certainly a lot of feedback from that. And yet that isn't the kind of decision that I would want to trust a machine to make for me. I'm an artist in this sense how do you decide what kind of decisions you can

in effect. Tom the ties are turned over to a machine and what kind remains. So qualitative as distinct from qualitative that they have to be made on other bases. Well I think we should be very careful here about any implication that these decisions are all to be made automatically on the other hand. We know approximately the basis on the average for a lot of these decisions we know in the form of insurance tables and insurance rates how many people will die during the next year. We know approximately how many will be married and it doesn't matter to the economic system precisely which person it is during the next year. But it is important to know approximately the number and perhaps the influence that various outside factors might have on this especially when we refer to the production and sale of goods. The more we know about the effects of advertising the effects of consumer credit and the effects of interest rates the better it will be our

basis for judging what is about to happen. Just as the setting of interest rates and the Federal Reserve System the setting of tax rates the setting of government expenditures all aimed at. Just instability in our economic system just as we attempt to do these now we can do these things better if we understand better what the system is that we're working with. OK then you still have really good leadership in the business or in the governmental organization. If you lean so heavily on this kind of a system for instance it would seem to me that if you were let's say running a newspaper or a magazine where you could find out a lot I suppose by modern techniques about what people wanted but you might not have a great newspaper or a great magazine unless you gave them something they didn't even know they wanted because they had never heard

about it before. Doesn't this feedback tend to limit leadership a little bit. No not at all quite the reverse. The point here is that if people's attention can be freed from the routine repetitive crises so that they have time enough to think about the creative aspects of our life. The things that you refer to I think we will find a great deal more reward in what we do a great deal. More satisfying progress in what happens at the present time people are so busy doing the same things over and over every day and getting themselves out of the difficulties that could easily have been foreseen if they'd been able to look a little further ahead. But there is usually very little time for the kind of new creative activity that you refer to. If one goes back to an agrarian society a several hundred years ago when everyone raised his own food made his own shoes it was perfectly

evident to each person how his work was related to his needs. But with the coming of an industrial society people build goods products in factories for other people to use it becomes less evident what the relationship is between their work and ultimate consumption of the consumer product. As we go to more and more automatic machinery for the construction. Of household appliances for the packaging of food we find that the activity that people engage in becomes further and further removed from the actual consumption of these products and this means that people must look further and further ahead that the decision made today may affect consumer products for five 10 or even 15 years from now. Rather than effect the production of consumer goods for next summer's use doesn't this creative very tough

educational problem I think are great merchants and great industrialists of the past have been very largely people who learned their business learned how to do a certain thing and maybe sometimes in the so-called School of Hard Knocks. Well how can they acquire this close grip on reality and close familiarity with the day to day problems and still be able to in effect utilize the kind of information or feedback from their computers and so on that I take it you're talking about seems to me that is asking an awful lot of men. Well let's not look upon this as anything new. We've had this trend. For several hundred years. After all public education is relatively new it came in since the beginning of this country that public education has been widespread. It's only since about World War 1 that a high school education has been common for most people. It's only now reaching the point where a substantial

fraction of our population goes to college. It's part of the trend that goes with a continuing development of the kind of society that we live in. There's nothing new here. It just intensified it means that some of the new. Aspects new knowledge of our economic system new knowledge of how to use devices like computers must be acquired by the people that are managing the enterprises on which our economy depends. How long does it take to and where do you get a good computer. Operator I take it your term for this is programming but we think of it as the person who actually operates the thing. There's a very wide range of skills that are employed here. And they extend from the best and highest trained people there are down into the

high school level. And I think most people would be surprised how many. Boys and girls out of high school and just above the high school level can find very interesting work in this field that includes the whole field of machine maintenance which is an electronic technicians and machine operators in general are not people who have gone to college they are people who take problems which are already given to them decks of punch cards who actually operate the machines but the problems in get the results out push the buttons in other words the actual running of the of the machine. The people who prepare problems for the machine at the present time are generally college trained but very often not highly trained in mathematics. A very wide variety of kinds of training have gone into the computer

programming field. I know one instance of a young lady from a women's college who had majored in music who became a very good computer programmer. So it is more an aptitude and an interest kind of person who likes to solve puzzles. That succeeds here and the actual amount of mathematics will depend a lot on the particular field to which the computer is being applied. I think the point here is that for any particular individual the better his ability and the higher his training the more challenging and interesting can be the job that he finds himself doing. But for those who are not able or not trained for some of the more advanced tasks there is a complete range of supporting work as goes along with this new field including of course the construction of the machines in the first place. In factories and maintenance after they come out and

the day by day aspects of running them so there is a complete gradations of skills required. So pretty seriously in taking a pretty hard look at it. The probable trend of developments both in the business world and in engineering and science. What studies do you think should be most emphasized by a high school student if he wants to be a businessman there is what should he take for sure. Or if he wants to be a scientist do you draw a distinction there. I don't think there should be such a sharp distinction as there is at the present time the trend in the colleges is away from specialization. And if the trend at the college level should be away from specialization and I feel even more so the trend at high school level should be away from specialization we think that will surprise a good many people because I think a good many people have the idea Professor Forrester that things are becoming work plex and we're all becoming

more and more specialized. I think this is one of the more serious deficiencies in our present society the tendency for people to see so little of the whole picture that you don't understand where they fit into it. And going back for a moment to my work in industrial management it's aimed primarily at this deficiency that we are trying to develop a much better feeling on the part of people as to how advertising and research and production and marketing and the behavior of people at the consumer level how these are related to each other how all these relate to investment banking and money flow. And too often we find people that study just one aspect of this and don't understand how it's related to two other things. I don't think there is any great conflict between what one needs to study for business or for science and engineering. It seems to me that

more and more our whole society. Is moving in the direction of scientific activity even if one is not going to participate at all in actual scientific endeavors he needs to understand some of the aspects of science and mathematics just so that he can understand what goes on around him. So I would say at the at the high school level it seems to me there should be increased emphasis on engineering and science but not necessarily because the individual person expects to become a scientist. Now this means that the way we teach engineering science in our high schools needs to be carefully re-examined it needs to be made interesting it needs to be evident to the student how it does relate to anything that he may do in the future and I'm afraid this isn't true at the present time. In most of our schools the challenge here on the educational side is to develop on the part of the

student an awareness of how interesting how closely related to everyday living some of these subjects are going to be. Well thank you very much Professor for a stripper telling us this about what the computers are doing both to engineering and business. You have been listening to the paperwork revolution with J.W. Forester builder of the world's fastest electronic memory and a professor of industrial management at MIT. This has been a part of century of science a recorded exploration of developments in science and their import for the 20th century American. This series is prepared by WGBH Af-Am in Boston for the Lowell Institute cooperative broadcasting Council. Your host voted Tory a former editor of Popular Science and now director of radio television programming for the Massachusetts Institute of Technology Director for the series Lily and Racine O. Producer Jack the Summerfield Bill Gavin a speaking century of Sciences produced under a grant from the Educational Television and Radio Center and

distributed by the National Association of educational broadcasters. This is the ne B Radio Network.