Oral-History:John McPherson
About John McPherson
John McPherson was born in Short Hills, New Jersey, on October 16, 1908 as the first one of four brothers and a sister. His father had a career as an electrical engineer and his mother was the daughter of a graduate of the Harvard Lawrence Scientific School. His father wanted him to go to M.I.T., but McPherson chose Princeton and started his college education in 1925. During the college years, he worked as a railroad apprentice in Altoona. After graduation, McPherson was invited by Mr. Watson to join IBM. He was assigned to the Railroad Department as a trainee and got his own sales assignment in less than a year. McPherson started out in New York and was transferred to Philadelphia. When he came back to New York, he worked with the Railroad Group, whose goal was to provide statistics of the movement of freight and passengers. While working with the Group, he prepared a set of charts called Machine Methods for Railroad Accounting, which was used quite widely. He worked in Philadelphia for about three years and then in the Railroad Group for about seven years. In 1940, McPherson was promoted to Manager of future Demands and in 1945 appointed Director of Engineering. He also took part in nuclear activities, although not active.
Before taking the position of Director, he worked on the Harvard Mark I, which IBM was encouraged in building by an instructor at Harvard, Professor Howard Aiken. In about 1940, McPherson worked with Aberdeen Proving Ground and introduced it IBM punched-card machines. After the war, IBM moved toward producing electronics, and McPherson participated in developing the 603, 604, 607 and SSEC. In 1948, McPherson was made a vice president, and from around 1960, he handled Systems Research Institute for five years as titular head. From 1966, he took interest in the APL business and focused on the job.
This oral history consists of two interviews. In the first interview conducted on April 29, 1992, John McPherson introduces early tasks of IBM, the use of its equipment, and competing companies. He also discusses the Harvard Mark I, a machine dedicated by IBM to Harvard University and explains the roles of various people involved in building the machine. The second interview of May 12, 1992, starts with John McPherson talking about those whom he encountered within IBM up until the end of World War II. In the interview, McPherson focuses on IBM's early move into electronics computing devices and development of various machines. He also recounts his career and important events at IBM up until his retirement in 1970 (he was not sure whether it was 1970 or 71).
About the Interview
John McPherson: An interview conducted by William Aspray, IEEE History Center, April 29, 1992 and May 12, 1992
Interview # 139 for the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc.
Copyright Statement
This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.
Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.
It is recommended that this oral history be cited as follows:
John McPherson, an oral history conducted in 1992 by William Aspray, IEEE History Center, Piscataway, NJ, USA.
Interview
Interview: John McPherson
Interviewer: William Aspray
Place: Short Hills, NJ
Date: April 29, 1992 and May 12, 1992
Family, Childhood and Education
Aspray:
This is the 29th of April 1992. The interviewer is William Aspray. It's an interview with John McPherson in his home in Short Hills, New Jersey. Why don't you tell me when you were born?
McPherson:
I was born here in Short Hills.
Aspray:
Not in this house?
McPherson:
Not in this house. In a small house about a half-mile walk to the station for my father.
Aspray:
What year was this?
McPherson:
This was 1908.
Aspray:
What day of the year?
McPherson:
I was born October 16, 1908. My father had been married before, and his wife was the daughter of a Harvard graduate of the Lawrence Scientific School.
Aspray:
Were there other brothers and sisters?
McPherson:
I was the first one of four brothers and a sister who came along much after the group of us had gotten a good start.
Aspray:
I see. What did your father do for a living?
McPherson:
He was one of the early electrical engineers, trained at Johns Hopkins in the years just before the electrical school finally got accredited. His degree was in practical electricity.
Aspray:
What about your mother? Did she work outside the home?
McPherson:
Oh, no. She was the daughter of a successful graduate of the Harvard Lawrence Scientific School. He worked in Altoona for the Pennsylvania Railroad in what's called their Engineering Test Plant. They had one of those that far back — to measure the power of the engines and their efficiency. He'd gone on to be the superintendent of the Erie Railroad out in Buffalo, over their Switching & Automotive Power Department. He took charge of their rolling stock. Then he went out to Chicago to work for the Westinghouse Air Brake Company, where he was hired, for reasons I don't understand. But he'd become a member of the Westinghouse Air Brake Corporation. Shortly before dad married my mother, he'd been made, head of the Westinghouse Air Brake in Europe with the home headquarters in London. So I had an engineering father. His early work was for a Westinghouse-connected engineering firm in New York — Westinghouse Church Kerr, I think, was the name of the operation. That brought him to Short Hills from New York. He'd previously been in Pittsburgh doing electrical consulting engineering. He continued to do engineering work as long as his eyesight held out. He ran into trouble with a detached retina. Before we were very old, he had to give it up. But he had an interesting time until then on several engineering projects. One was a super-power survey of the Eastern Seaboard under Mr. William S. Murray. Later, with the engineering concern that planned electrification of the western end of the Canadian Pacific Railroad. Later yet, the Illinois Central. I don't know whether he actually got in on the Lackawanna or not. But those were the things that he did while I was growing up.
Aspray:
Tell me about your education.
McPherson:
I was jump-started by my grandmother, who lived with us when I was in the early grades. She had taught before she was married and lived with us for a number of years. She had me up to the fourth grade level almost before I went to school. So I got through grade school and high school locally here without any trouble, but I didn't pass all the college boards. I had one extra year at Hotchkiss and repeated the senior year, in effect, before my college boards were all covered for admission to Princeton. That was my choice. Dad thought I ought to go to M.I.T. because one of his brothers had a home up there near Cambridge, and I could live with him. But I didn't. [Chuckling] I didn't think much of M.I.T. I wasn't anxious for that. I was very anxious to go to Princeton and was lucky enough to get in and have some scholarship help during my years there.
Aspray:
So you started in 1926 at Princeton?
McPherson:
No, it was the class of 1929 — so it must have been in 1925. I finished a B.S. in electrical engineering degree, but worked during the summers up in Altoona. The first summer I worked with the electrical engineer who was designing some of the motive power for electrification use on the Pennsylvania Railroad. From then on, I worked as a special apprentice every summer and Christmas during college.
Aspray:
Did you know that you wanted to go into engineering when you went off to college?
McPherson:
Yes, I wanted to be an engineer. Dad had let me see, understand, a little of what you did in the way of designing electrification projects. I'd been playing with electrical work all this time.
Aspray:
I see. So it was not only being an engineer. It was actually being an electrical engineer you knew.
McPherson:
Yes.
Aspray:
Had you been taking lots of science courses before you went to college, for your college preparatory?
McPherson:
I had no problem with learning my mathematical courses, and Hotchkiss had an extremely good instructor in physics. I did that and had no problem. I really enjoyed it.
Aspray:
When you were a child, did you have hobbies? For example, did you have an interest in ham radio or in tools?
McPherson:
I had a very early contact with radio, when it was dots and dashes. My uncle got me a little crystal radio set back in 1918 or so. We went over to Paterson to a jeweler who was making these things as a hobby, and we came home with one, which I still have upstairs and which I can still get to work.
Aspray:
Oh!
McPherson:
I did enjoy making things and particularly various toys that I constructed for just the fun of making them. I wasn't too concerned about hobbies. I played a little bit of football but not to get anywhere. I read a lot. I was very quiet, and I worked hard on my books. That was fortunate as I got very substantial Princeton support got a couple of the prize scholarships for the last two years, which helped dad. Because with four boys on an engineer's salary, he was always up against it. [Chuckling]
Princeton
Aspray:
Yes, I can imagine. Can you tell me what it was that attracted you to Princeton?
McPherson:
Their football team. [Laughter] Well, that wasn't the thing that triggered me. The thing that triggered me was listening there to Princeton, Chicago and Yale, which was famous for years afterwards, on this crystal set. That was what triggered me. But after looking at the catalog, I realized that Princeton engineering was a pretty liberal course, and that struck me as a distinct advantage. It was an engineering school that called for a good deal of liberal arts.
Aspray:
I see. Liberal in the sense of adding liberal arts to the technical courses?
McPherson:
Right. History and economics and things of that sort, as well as the required engineering course.
Aspray:
So, very different from the course of studies you'd get at M.I.T., for example?
McPherson:
Actually, it was only different to the extent that these extra liberal courses were required in the first two years and were taken with the regular liberal arts college people. It was, I found it to be very interesting.
Aspray:
What can you tell me about the course of study in your own engineering field at Princeton? What kinds of courses did you take? Do you remember any of the professors? Do you remember the textbooks?
McPherson:
I remember at least a few of the professors. You took the courses that were required. I thought that some of the people there were exceptional, although they were all good. But it was a power electrical engineering era.
Aspray:
Were there any courses on the control side? Were there telephone courses, for example?
McPherson:
No, but there was an excellent course by Walter A. Roberts in the design of radio sets. He was an RCA man, and that course, coming in my senior year, gave a real understanding of what the latest in electrical developments consisted of and how you went about doing electrical engineering, engineering design. But in spite of that, since I was a railroad apprentice as well, I wasn't thinking too much about just what I was going to do after college. I continued that line of work. I did go back up to Altoona after college. The summer right after college I was invited to go visit my grandfather in England and take my next younger brother with me. He offered us a small gift. And Dad said, "Well, that's fine, son. But you can probably manage to take yourself and your brother Bruce as well." So the two of us went. When we got back the following September, I went back up to Altoona to continue the various stages of a three-year apprenticeship. The following Christmas was invited to join IBM if I wanted to.
Aspray:
Before we move on to that, let me ask you a few more questions to fill in. At Princeton, did you have to take a course of study in mathematics?
McPherson:
Yes. I took differential equations.
Aspray:
Did you take a numerical analysis course of any sort?
McPherson:
No.
Aspray:
What about physics courses?
McPherson:
You had to take at least a year of physics.
Aspray:
So you'd get basic principles of electricity and magnetism, for example?
McPherson:
Yes.
Aspray:
Were there any people you met in college — fellow students — who were closely associated with you during the rest of your career?
McPherson:
No. But after graduation most of us joined the Princeton Engineering Association. To that extent I kept in touch with a number of members of my class and ones that I had been following for a number of years.
Aspray:
But you didn't go to work with any of them?
McPherson:
No, I didn't. I had a course planned for at least a couple of years, but I didn't have a business association with any of the rest of my fellow graduates.
Aspray:
Are there any particular faculty members who stand out in your mind as being particularly important to your education?
McPherson:
Well, I mentioned this Mr. Roberts. But at the time the man who impressed us most was in history. Professor Hall was one of the outstanding teachers at Princeton at the time, for many years before and a number of years after. The faculties were all competent, but none of them made a lasting impression on me. I did well in all my courses, but I guess I just went right ahead.
Aspray:
For a typical graduate out of the engineering program at Princeton, what kind of career would they have? Where would they have likely gone?
McPherson:
All tended to do essentially some kind of engineering. But I suspect that most of them would go into business, using engineering as a background rather than as an activity.
Aspray:
Yes.
McPherson:
The most successful member of our class was a man named Naughtley in the Aeronautical Department. He eventually became head or chief engineer of Beech Aircraft out in the Midwest. Al Skerlus became an engineer at Grumman and did quite well there. And Blair Birdsell, a civil engineer, became a bridge-builder and had a part in a great many of the big bridges that were built via the firms he was associated with. There were no great engineers in the group as far as the way it's turned out. But they all found good use for their engineering education.
Aspray:
Was your career plan out of the ordinary? Did your fellow students think that this was sort of a standard thing to do or an unusual thing to do — going on to the apprenticeship with the railroad?
McPherson:
[Chuckling] I never gave it any thought at the time.
Aspray:
I'm just trying to get some sense of to what degree your career was typical of an engineer at the time.
McPherson:
It would be interesting, but I know of nobody who thought of it that way. We all just essentially said that if you're an engineer, you'll do well anywhere. If you can think things through clearly.
Aspray:
Before we start to trace the rest of your career, why don't you tell me about the careers of your three brothers and your sister? To the degree that you think it's relevant.
McPherson:
Well, unfortunately, the story is that it started out with the death, the year after I graduated Princeton, of my brother who was graduating from Hotchkiss, where he'd been for four years, in an automobile accident just before graduation. The car was run off the road by somebody into a bridge abutment, and he was thrown out of the car. My other brothers had a good deal of trouble getting their education. They had trouble scholastically. My second brother after me the third member of our family — worked for a while at Westinghouse. He worked for a firm on Wall Street for a while at small brokerage firms. He then worked over in the Westinghouse plant in Newark and went in the service, as did his younger brother. This kind of upset their careers. All three boys went into service during World War II. Edward, my youngest brother, worked for a small bank over in Brooklyn before the service and came back to the bank afterwards. He spent his lifetime as a junior officer of a savings bank. Robbie came back, worked for Fairchild Camera for a number of years, and retired to Florida quite early. My sister was in the service as a WAVE and came back and finished her college work. And then worked for Slumberger for a number of years because she specialized in French in college. That got boring because her boss was in Paris [Chuckling] so much so she decided she wanted to teach and did so for 20 years.
IBM Job Offer
Aspray:
All right. Let's return, then, to your getting the offer from IBM. How did this come about?
McPherson:
It actually came about because I was invited by Mr. Watson the first Christmas I was home to join IBM. Fortunately, I decided I would like to work for IBM. [Laughter]
Aspray:
Had you applied to them?
McPherson:
No, I had not applied to them. I'd thought of it. When it came hiring time at college, I'd wondered about working for AT&T in their research. But I got this chance to go with IBM the winter after I graduated, and there was never a dull moment from then on. [Chuckling]
Aspray:
I don't understand how Mr. Watson knew about you.
McPherson:
Well, first of all, he was interested in people. He must have had a slight knowledge of what I'd been up to. He came down to Princeton at one point to give a lecture for the Engineering School. At that time he'd checked up on how I was doing, I think. At that time our minister's son, Ed Douglas, was working for him, and he came down to demonstrate a machine at this meeting, as well as what Mr. Watson had to say about the big world of business and the use of machines in the office as well as in the factory. I never had a clear picture of why he might know of me. But if he came to Princeton, he probably checked into any Short Hills boys that might be there. He was that sort of person. Interested in people. And my record was good.
Aspray:
At the time that you joined IBM, it was a fairly small company.
McPherson:
It surely was.
Aspray:
What attracted you to it?
McPherson:
The main point about possible work for IBM was that I talked at some length with Ed Douglas, who was already there. You were essentially like a salesman in that you worked for yourself. Your territory was something that you developed. If it grew, you made out all right; and if it went sour, you got hurt. You were on commission. Commissions on rental business disappeared if your business disappeared, but you were in business for yourself.
Aspray:
Did you like that feeling of self-reliance?
McPherson:
It seemed to me to be a big plus, in addition to the fact that the business was something that looked new and promising. The punched cards were beginning to expand into the business world, as well as the statistical work of the census.
Aspray:
Was it customary to hire people with engineering training to work as salesmen?
McPherson:
No I just happened to be somebody that was a college graduate with a good head. Actually, at the time, I think most of the people coming into the company were college people. Certainly the sales force was. It was small enough so you knew everybody. In the course of time, that was the biggest holding attraction of the company, I think. That the people were a team working together. Unfortunately, people were being moved around all the time, but it was necessary for an expanding business to build up in that way.
Sales Work in Railroad Department
Aspray:
What was your first job for IBM?
McPherson:
When I was hired, I came in as a trainee. Within about six months to a year you were assigned to an office and worked with some of the older salesmen as an additional training. I was almost immediately assigned to the Railroad Department, which was the sales group that handled the railroads. This was a major industry area where machines were beginning to be widely used.
Aspray:
Was that merely an accident, or was it because you had that background?
McPherson:
It was deliberate on the part of whoever decided what I should do. They felt they might just as well take advantage of the fact that I knew something about railroads. My father and grandfather had been interested in it. Although Grandfather had been in England the whole time. I never saw him until that one summer I went with Bruce to England. He was supposed to come on the "Titanic" and had to cancel at the last minute. I think he only made one further venture across the Atlantic by boat. For good reasons, I would say. [Laughter]
Aspray:
As a trainee, what happened to you? What kind of training did they give you?
McPherson:
Well, you just went out with the salesman and sat in on his calls on customers. You had to learn how to use the machines: the punch, the sort, the tabulator. To understand, you went to sales school, which essentially went over in great detail both how to use the machine and how the machines could be applied in various areas of accounting work. You attended a sales school for maybe two months at Endicott or one of the plants. It was small, but they also had some real good help and case studying. Part of the sales training was studying cases that the Harvard Business School had put together. Dr. Ben Wood was one of the Harvard people who helped organize material for the sales school and teaching. It was very well organized to prepare you to be able to talk on equal terms with the people that should be buying the machines, even though you were an outsider and a salesman. You were trying to teach them how to use a tool that was different from the desk-type machines that were usually the only thing you saw, or just hand calculators. When you look back, I mean, we've come a long way in a hundred years.
Aspray:
We certainly have.
McPherson:
I still remember one of our neighbors in Short Hills, who was a banker, saying that when he started his training in the 1890s, only a few people used the adding machines. They were just coming into use.
Aspray:
Right. Comptometers.
McPherson:
Comptometers were widely used in the railroads because they had some fairly heavy need for multiplication. It's amazing to me how short a time since the office was aided only by paper and ink, and mental arithmetic.
Aspray:
So it was about a year before you got your own sales assignment?
McPherson:
Less than a year.
Aspray:
What was your sales territory?
McPherson:
I started out in New York and in less than a year I was transferred to Philadelphia and given a railroad territory over there, the several railroads that had headquarters either in Philadelphia or Reading. There were two of them. I had two accounts up in Bethlehem. So I had to travel a bit from where I was over in Philadelphia. My principal account was working with the Reading Railroad and helping them really expand their use of IBM equipment substantially to all phases of their business I incidentally found a little Philadelphia Norfolk Steam Ship Company that hadn't heard about computers. They were my one essentially new customer in the year I was over there. This was just the beginning of the Depression, and if you could do anything at that time, you were awful pleased and you were very lucky. I made out all right those first three years. Then I came back to New York to work with the group there which we called the Railroad Department. It was half a dozen people who traveled the country to talk to railroads from New York to San Francisco. During World War I the federal government had operated the railroads, so their ways of doing things were pretty much put on a common basis. What applied to one railroad, applied pretty generally almost directly to any other railroad. They were called upon to provide statistics of the movement of freight and passengers, which was a fairly big and continuing additional burden of paperwork to the regular task of keeping track of how to distribute the money between the railroads that shared in an interline business. There was also the problem of collecting the business — either by cars or destination — for the railroad transportation provided.
Aspray:
Can you give me some sense of what your everyday work was like in Philadelphia? Would you spend most of your time at your customer's site or in your office?
McPherson:
Most of your time was spent either out with customers to show them how to get more use out of the machines they had because their business was short of it. If they didn't get more new applications, why, you tended to get the machines back. Sometimes we had to hold the hands of the railroad people and take machines away because they tended to hold them when they shouldn't. They needed the economies to live through the Depression. We occasionally took some pretty bad setbacks deliberately because it was the right thing to do from the standpoint of the customer. When you're in a rental business, you've got to always put his interests first.
As I say, you spent your time with customers trying to study and examine. Most of our work was done by going and asking questions. Every time we saw a chance to do something, we had to try to help them justify it. We had to know how their job was being done, and how we could put it on a machine. We would design a card with the fields having the necessary information to do the job. Sometimes we would go out and find new customers in our field or in our territory that might be persuaded to try an initial group of machines.
Aspray:
Were there other IBM salesmen in the Philadelphia area covering other kinds of businesses and products?
McPherson:
There were probably ten or fifteen salesmen covering a hundred to two hundred existing customers and looking to double that number as fast as possible.
Aspray:
It sounds like you had to really learn quite a bit about the actual everyday workings of the companies in order to help them.
McPherson:
Absolutely. This was one of the most attractive things about it. You ended up knowing almost as much about how a business ran as the people running the business. One of the points of the sales school was to essentially impart some of that background to the salesmen so that they were competent advisors, able to talk to the top management people of a business — at least to the top accounting people. In the case of a big railroad, those guys were pretty powerful. They had two and three hundred people under them, and a lot of work to be done. There was a lot of opportunity.
Uses of Punched-Card Equipment
Aspray:
What different kinds of tasks were done using punched-card equipment by the railroads?
McPherson:
There was a list of forty or fifty different operations. But essentially the railroad work was in three major areas: handling a freight account, which was a major job; handling payroll; and, as time went on, keeping car records. Initially, that was a completely manual operation, where you had a great big ledger-like desk with a line for each car and the days of the month, and you said where it was every day as you got wheel reports from the train crews indicating what had happened. The office could then trace and collect money for the rental of their cars and, to some degree, answer questions when something was lost. If the car was a carload of sausage that didn't get where it was supposed to in short enough time, there were obvious problems you’d have three or four different kinds of payrolls. Some of them were salary payrolls. Some of them were payrolls for the train crews that would have special rules governing their pay on a piecework basis: how many miles you worked that day and hours. Then labor in the repair shops, which was a straight maintenance operation. Later on you had to worry a little bit about keeping the books for a company, and using the information in the freight accounting to do some elementary sales analysis, seeing how much business you were getting and trying to get some of the business routed favorably to your own railroad. I've still got upstairs a manual of the Uses of IBM Equipment. It included a whole spectrum of industries, by customer, by application. For a long time when we were working with the railroads, we'd keep charts up to date and use the example of Railroad A in convincing Railroad Z that they could use this application — or all these applications. So essentially the task was one of introducing, in the railroad case at least, proposing a method of doing the job that was perhaps more efficient, less expensive than the existing hand method.
Aspray:
A customer like Reading Railroad, how much IBM equipment would they have?
McPherson:
In those days they'd have two or three groups of equipment involving anywhere from $2,000 a month rental to $50,000 a month rental. Actually, a big installation would probably have four or five accounting machines of one variety or of one size or another. And, say, groups of fifteen keypunches for making the initial entries into the system. The Reading Railroad instance was a medium-sized railroad. It was bigger than the Lehigh Valley or Lehigh New England. Lehigh New England just had one machine. That's where I learned freight accounting. We put in 80-column freight accounting there at almost the beginning of my period up there. They were very nice. They taught me the railroad end of the business while I was helping set the machines up for them. They used these machines for a long, long while until they finally got absorbed by someone.
Competition and Sales Methods
Aspray:
Was there any competition? For example, was Remington a competitor in this business?
McPherson:
Remington-Rand was our principal competitor. They had Powers accounting machines. At the time we started, they had pretty good representation. They were probably somewhere around 20 to 30 percent of the railroad business. But I think in the years that I was with the Railroad Department, as a member of it and then as an assistant to the head of it, the railroad that wasn't using IBM was the exception. Only a handful of big railroads left.
Aspray:
- Audio File
- MP3 Audio
(139 - mcpherson - clip 1.mp3)
What was the advantage of the IBM equipment?
McPherson:
Electrical. Flexibility of electrically setting up to do any job you wanted to in your own shop. All you had to do to start up a new job was design a new card form which would tell you how you punched some figures, what figures you punched, and how they'd be arranged. You could then take your own selection of what you needed of that information on a card to any particular report you wanted to run. You'd never use all of it at one time, but if your card was properly written — prepared — you could count on it. The principal benefit was if you used the same numbers in several different ways and in different combinations, it was entirely within the user's choice. They didn't have to go have something built in the way of a mechanical connection, which was necessary in the Remington-Rand or Powers situation. In varying situations, Powers was probably just as good as we were. They had some advantages. They had flexible ways of loading the data.
Aspray:
Did you get to know your competing salesmen from Remington-Rand? Did you know them personally?
McPherson:
No. That was a rule.
Aspray:
It was a rule that you didn't know them?
McPherson:
You didn't know them. You were really careful, as you are now, of not disparaging them. But there was really no interaction. Remington Rand was, unfortunately, a diverse company. The punched-card business was only a small section of it. Most of their people were selling bookkeeping machines. You didn't run into them at all. On a freight account you don't want them to displace 15 or 20 comptometer operators with an IBM machine that might do the work or it became a different problem. But we just barely got to that point where we could use our multipliers — card-operated multipliers — to do that work, because the comptometers were pretty efficient, and the jobs were a lot of little jobs. We stuck pretty much to just the adding-machine type of operation in the railroads, except for the payroll which you could put through. That was the first place we used the multipliers.
One of the things was that as IBM grew as each new machine came along you'd suddenly have to review how you could improve the IBM picture — do a better job — by announcing it. I remember one time we had a group of railroad people at Endicott. The collator had just come out as an addition to the sorter for resequencing cards or merging groups of cards. I think we ended up with 40 different specific places where we could use the collator, which would put us ahead of the job in getting these cards prepared for end-of-the-month work. Much of the card work had that nature; that you'd punch cards all month, and then have to put your figures together for the month by railroad, by station, by boarding station or receiving station. That's where the punched card had a really clear economic advantage. It still does. I don't think we pay as much attention to it now as we did then when you had to essentially displace manual keyboard machines with an initial recording and use it as many ways as it could be used to make it productive. One of the last things I did while I was with the railroad group was prepare a set of charts which we called Machine Methods for Railroad Accounting — 70 or 80 aid charts showing the card forms you used, the steps through which you got the data from, it and the steps through which it earned its keep by making the reports or summaries. The job was all right, and it was used quite widely. I applied it line for line at one point on the Lackawanna Railroad. And I helped the local Newark salesman sell the account. We installed part of it at the Hoboken Station and part of it up in Scranton, where they'd had an earlier initial IBM installation of the less powerful machines.
There was a big difference between 1930 and 1933. Up to the early 'thirties, you never had more than 45 columns of information on a card. The expansion of the cards by making the usual round holes into rectangular ones, giving you 80 columns, was just enough to essentially do a good job on the railroad accounting and in many other instances. The information you needed in one spot was really not adequate on the 45-column card. The 80-column freight accounting was the thing that really gave us the superiority factor in that particular job. At the same time we began to get alphabetic printing, which made a big difference. That started just after I joined. The first alphabetic printers were 1931 or 1932. Actually the railroads did without the alphabetic printers except for payroll — almost the whole time. They never really cared about the names of the stations. They went by numbers. Unfortunately, they had to have people on the table because of the difficulty of sorting alphabetically.
Aspray:
What was the relationship between the salesman in the field and people working in home office, or in Endicott? What kind of contact did you have with those people?
McPherson:
Well, the Railroad Group in headquarters essentially were in the position of advisors to the local sales people who had the accounts and who got the financial benefits of pluses or minuses. Our headquarters railroad groups were looked on as consultants, assistants in that sense. They'd go to the annual meetings of the Railway Accounting officers where at least all of our railroad reps and so on from our group and headquarters would show up. We had a very good rapport with the senior people in large accounting organizations of the big railroads. That meant that when we came to town, we helped the salesmen who might be dealing with, say, two levels further down, to essentially feel they had a real support, even if they were spending most of their time down in the keyboard factory. It was a very friendly and cooperative situation.
Aspray:
So that if an individual salesman came across a problem with a customer, an application that they couldn't handle, they could call on somebody from the Railroad Group in — ?
McPherson:
As a matter of fact, it was usually the other way around. If there was something new coming up such as that, it would probably come up in a lot of railroads. If it comes up in Railroad One and we've been called in on that, then the Railroad Group immediately had to suggest it to all the other areas — St. Louis, Cincinnati, Cleveland or Rochester, R&L — and the various railroad shops. Of course, railroad shops, which maintained and repaired and built engines and cars, were essentially in a common business, but run independently, helped by each railroad. I haven't mentioned passenger accounting, but there was a similar group of activities there. One of the things we came up with in the Railroad Group was an idea that you could work on ticket sales of $300 apiece — in those days that was a lot for a single unit — if you kept track of the ticket numbers and could keep track of the stations selling them. They'd essentially be able to compare consecutive numbers. As now on many bank accounts you see you have a little list of your checks with a marker saying "missing check." That was one of the things that started far back in Railroad Passenger Accounting. I guess I was the first one to suggest we could do this. I borrowed it from somewhere; I didn't invent it. I saw how it could essentially do this: having this ability to check equals and check consecutives. It worked out fine.
Aspray:
Were there suggestions that either salesmen or people in the Railroad Group provided to the engineers who were designing equipment for improvements? Was there that kind of interchange?
McPherson:
Yes, very definitely. Some of them started when some of our customers in the early days came up with quite critical improvements in the machines that we built. In at least one case, we found a railroad auditor with patents, which we needed to acquire to make the machine the way it should be made. Then it could really be done once you'd seen the possibility. We did have an organization back then which did accept requests for features that would help, and by a slight modification were possible. One of the most prominent of those things was the business of checking meter readings for the electric utilities. In other words, they took a meter reading each time, and we devised a scheme, for instance, for determining from last month's. Normally, if you just record the two readings and they're in separate counters — separate parts of the machine — you had to be able to also take the two and subtract one from the other. And that feature was one which was a request through the Utility Special Rep Group, which was applied very widely. Knowing what Roberts was doing there suggested that we could make sure the punching of our freight cards by using that same idea just one degree further. We could add in advance in two different fields and compare them with debit and prepaid, if those two pairs of numbers agreed, the card was correct and all four fields were right. We just added a few wires to the machine, but it gave us a way of verifying the amount fields of all the keypunches of the freight card punched. It was trivial, but it was a tremendous boost. Instead of having a second person re-key the whole card, you just took the stack of cards. The way it finally worked out was that as they were stacked, if a card didn't check out, it was offset in the pack so it would stand out for correction.
That was the sort of thing that made life a little bit interesting, and where I made some slight use of my engineering insight. I always was interested in how the machines were built. I talked to servicemen and looked at the diagrams when they were lying around the machine. In those days we didn't let customers see the wiring diagrams. That caused some real problems.
Aspray:
So how long were you in the Railroad Group?
Manager of Future Demands
McPherson:
I was in Philadelphia for about three years and then in the Railroad Group for about seven, which makes it about 1940. I think 1940 was the time that Mr. Shackleford was my boss. When I first joined the Railroad Department, there was an old-timer handling it, Mr. Quackenbush, Senior. A year after I joined it he retired, and Mr. Shackleford, an auditor of one of the railroads, became the head of our Railroad Group. He'd been a customer for a number of years, and we hired him. He was a great big man. He did a great job. He was quite methodical, but also very progressive. I enjoyed working for him and made many trips with him. But it was about 1940 that I was made Manager of Future Demands. I don't know why it came about, but this was apparently just an assignment to try and see what new things we ought to try and build.
Aspray:
Why do you think it was that you were given this assignment?
McPherson:
At the time I had no idea why other than it was just something that Mr. Watson wanted done. I think he wanted me to understand a vision of what and where the punched-card business might go, and how effective it could be. He was constantly trying to find out what else we could do. He tried to find out what we could do in education, in medicine, for the government. We had a group at the time called Commercial Research, which was a group, which would gather special requests from the field. This was essentially set up as a one-man operation. I had a part time secretary too. I was working in close collaboration with them but without the sort of day-to-day chores of listening to somebody with a specific thing and doing something. On that particular aspect, I would keep in touch with what they were doing to see if there were patterns in what they had to do that should be done better or done in a new way. I don't know whether I found very much in the period. I kept quite close in touch with them, and I think I got acquainted at least with the development groups in Endicott.
Aspray:
When you speak of Future Demands, were these demands and applications that could be met with existing equipment, or with new equipment that was to be designed?
McPherson:
Future Demands was intended to be for things that needed to be designed. For instance, the collator had already come about by then, but if there had then been someone in Future Demands five years before that, they would have essentially tried to formulate that machine in such a form that it would be of maximum use as a tool for expanding the business. As the card-to-tape and tape-to-card machines that came along later, they represented things that were Teletype in nature. Or emerging punched cards in communications activities. It was hard to find completely new things to do. But I'm sure that as time went on, it was worthwhile. I had a challenging opportunity.
Aspray:
And how far into the future were you supposed to look for uses? Was this supposed to be something that should pay off within the next two or three years? Or were you looking farther into the future?
McPherson:
No, I think the Future Demands see were essentially something on which development should be started for a hopefully useful product in one, two, three or four years.
Aspray:
Suppose you came up with an idea for some new equipment? What would happen then? What would the process be of its being considered within the company?
McPherson:
There was no formal process. What I would probably do would be to make them a good description of it and just say, "Mr. Watson, here's a good idea." He was a perpetual source of things; "we should do this," and "we should do that." Of course, the more important ones represented the way the business moved ahead. He originally backed the introduction of electricity for the keypunch. Before that you had to bang those holes out with your fingertips instead of just touching something that cut it for you. A whole series of products in the past had been that sort of meeting a future demand, where we didn't have the product to sell yet.
Aspray:
And those were ideas of Mr. Watson's?
McPherson:
Well, in the time I was in Future Demands, I must have had to field about an average of ten a month from him. Most of them you couldn't do anything about. He was just a real source of "let's do something different and do it better or do a better job for the customer with it." It was very much user-oriented in the sense that the customer should be king, but we had to have a solid product before that. We could not have complete diversity, or we would be spread all over the map.
Aspray:
Who did you report to at this time? I know reporting was not so formal in earlier years.
McPherson:
I never knew who I reported to, but I always reported to the people directly under Mr. Watson. Mr. Nichols for a long time. Mr. LaMotte. Mr. Stevens, who was General Sales Manager. In those days I didn't have any idea whether anybody held my card or not. Today I think everybody worries about that. That came in later in the history of IBM.
Aspray:
Did you see Mr. Watson very often?
McPherson:
No. First ten years, probably once. The next five years, probably once a year. He would speak to me if passed in the hall. But if he wanted to talk to me, he would call me in to talk about something or ask me to call. I just hated to undertake the job of carrying a story to him because it was an interminable wait before his secretary could get him to call me in. I found that other people that he wanted to talk to would displace me on the schedule, which was annoying. The business was essentially run from his office. He kept the things going because he picked the things that needed most doing and talked to people about those things. He trusted that the projects that were in good hands.
Director of Engineering
Aspray:
How long were you in Future Demands?
McPherson:
Three or four years, then World War II came into view. Mr. Watson wanted the engineers to devote their time entirely to aiding the war effort. There was a Number 2 said to me that Watson thought that Future Demands could be put on hold for that period. About 1943 he decided that I should be Director of Engineering. I think it was Larry Harrison who was just ahead of me on that, and really had no qualifications for it. Mr. Watson had been running the show — directing these projects to be carried out by the engineers at Endicott — almost entirely. He also had a group of people in the typewriter business who were trying to carry on continual improvements of the electric typewriter. Those things he either ran directly or through the man from whom he had bought that Electromatic Typewriter end of the business. But he felt he could run it. It was his baby. So for one reason or another, I was appointed Director of Engineering there in about 1945.
Aspray:
Do you think that the reason for your appointment was that you had both some general engineering background from your education, and a long line of involvement with applications and customers?
McPherson:
I had consistently made suggestions to basic Commercial Research people of things we ought to publish — better ways of using the machines. I was essentially the object of and not the source of what happened. I really don't know how it worked. But I never gave it much thought. When I started out at IBM, I vowed I would never ask for a raise. I wondered from time to time if that was foolish, but that was one of my principles. I never felt that I was in danger of being fired; I never felt that I would not give the company anything that I came up with.
Aspray:
Can you describe in a bit more detail what your duties were as Director of Engineering?
McPherson:
What I considered my job was essentially being in touch with the laboratories and the individual engineers there more frequently than Mr. Watson had time to do himself. Part of that was filing the monthly reports from Engineering of the progress of each of the projects, which were written, on a diary basis. And coming up with suggestions to Mr. Watson, or the other senior people in headquarters, of things that we ought to be doing that we weren't doing, or things that we needed to push on with. At that time there was no question of so much budget for this or budget for that. You had the staff, and you made the progress you could. I was trying to represent Mr. Watson to the engineers, and simply give them the assurance that somebody down in New York thoroughly understood and appreciated what they were accomplishing. I would be up there two or three times a month. When I went up there, I talked to the local Manager of Engineering who was Wally McDowell most of the time. Then talked to and visited the individual engineers to see what was making progress. I took a direct interest in seeing just what the status was and what their problems were. And they might just talk, too. I was a sounding box, as it were. Somebody who understood and was encouraging, and who knew what he was evaluating. Interested in and valued the things they were doing. And appreciating it. It was no more than that, really.
But it was also, at the same time, building up a knowledge of what else was going on outside. And in the course of that I was thrown into the nuclear program. I was expected to find out and take part in these nuclear activities; we were then looking into the nuclear power plant. There was always a future in that area.
Aspray:
You were doing this during the war even?
McPherson:
No, this was after the war. This was peaceful uses of the atom. After two or three years I went to a number of meetings, trying to assess the question: Should IBM become a part of the nuclear industry's activities? Mr. Hayward and I wondered what we should do with Carlson, the man who invented Xerox — I mean, the copying process. Mr. Hayward probably decided that for me. But anyway, we didn't take an active part in that, and my basic reason for it was that we had such a lovely and profitable business with what we had that we'd never be able to give this new thing the backing that it needed if it wasn't going to be just, "let's build this, too and give it to the gentlemen in engineering." It looked like a lot bigger problem than we could handle. One that would require a different attack.
I was in quite close contact with the Watson Lab, which we started up post-war on this question of computers. I followed that from the very beginning and made several trips to Europe even after I was no longer Director of Engineering, and I was more or less on my own.
Harvard Mark I
Aspray:
Let me go back and take more slowly some of the things that happened in this period. In the period from 1943, when you took over this position, until the end of the war, what kind of work were the engineers doing in Endicott?
McPherson:
They were doing a whole series of jobs, some of them suggested by the Office of Special Devices in the Navy. I don't have a clear memory of what they all were because we didn't usually write them up. If they were secret, we just let them sit where they were. But one of the things we worked on, I remember, was an odometer device so you could keep track of where a vehicle was by essentially drawing a map of it as it went along. I know we did a number of special jobs in the area of typewriters, for connection to pantographic equipment.
Aspray:
By the time that you took that position, the work for the Harvard Mark I was pretty much finished?
McPherson:
The period should be absolutely pinned down in my mind, but it isn't. I didn't keep a diary then, but I wish I had. It would be helpful now. Right now my diary is essentially a bunch of newspaper clippings and things like that, which are very disorganized and disorderly.
Aspray:
- Audio File
- MP3 Audio
(139 - mcpherson - clip 2.mp3)
Let's go off target for a minute. Last time I was here to visit you, you told me a little bit of the story about the Harvard Mark I and the dedication ceremony. Do you want to tell me that story from the beginning on tape now, as you remember it?
McPherson:
I guess my principal point was to emphasize the fact that IBM was encouraged in building it by Professor Aiken — Mr. Aiken — an instructor at Harvard. Mr. Bryce had been one of the people who'd gone up to Harvard in the very early days when it was a "Should we do it or not?" sort of situation. It must have started actually about 1938, and it was one of the things that was continued without being interrupted when the war started and all of the rest of the engineers were told to do war work. But we got started fairly slowly. The machine was built in Mr. Lake's department. The only problems we got into were essentially those at the bitter end. The construction went ahead fairly steadily during the early period of the war, but when Aiken went and put on his Navy uniform, he essentially left this thing to take care of itself and let some of the younger people try to be the contacts at Harvard until it was finished. We had real difficulty in getting any further without this direct guidance from him. There's a dispute now about why the Harvard machine did 24 places in arithmetic, which as far as I can recall was determined entirely by the fact that he wanted a strictly digital machine with no floating point capability. He wanted the decimal point to stay put.
We later varied it in various ways and with greater knowledge. Aiken was the sort of person who was literally minded. But it's obvious to almost anybody that you don't need a 24-place position for describing almost any number. But if you don't move the decimal point, the number may be a hundred billion or it may be a billionth of an inch. I'm sure that was the thinking that guided him in specifying that it should be that big. I guess he didn't realize the amount of complication that it induced. Because it really meant you'd have to have a lot more equipment in the machine by a factor of at least two. And only have time by comparison with what you might have done. But this was essentially the first. This was to do precise work. And you could argue with it and think of other reasons for it. The whole Harvard problem didn't burst until Aiken decided to pre-announce the event of the dedication at Harvard. It had gone ahead very smoothly about a year before when the machine was running at Endicott and the senior people at Harvard came up to take a look at it. We had a very interesting one-day meeting. I was fortunate enough to be included with them and showed them what they would be having. At that time the machine was open, in a test assembly area and running, but I think it was fully equipped with the stuff that was going to go into it. But then after that meeting, Mr. Watson decided it had to be covered. IBM would not ship a machine that looked like it was incomplete — without the cover. I had to get hold of an industrial designer to do a cover job for us. It was interesting.
Aspray:
It's actually a very beautiful machine.
McPherson:
Well, it was a big machine to cover up. Physically its geometry was bad. A great big box for all the central hard gear and then nothing. A tremendous series with a long, long drive shaft. But we were good at doing mechanical construction, as well. The Lake relay and the new counter for the machine were really big advances in capability, which, in turn, made the postwar punched-card equipment a lot better.
Aspray:
So these were things that were designed for the Harvard machine but then were used in future products?
McPherson:
No, they were things that were already in design. As a replacement for the existing accounting machines, which essentially were first incorporated into a single machine, I think, in the Harvard machine. In other words, they could use that as a place to build and as an excuse to get the stuff tuned and built because the unit — the counter — was something about 2 inches square. And the relay, of course, was very small but dual contact, dual-wire contact, which turned out to be quite a good unit. Quite an effective relay for reasonably high-speed operation. In fact, the wire-contact relays were the thing that gave us the tremendous advantage over the Telephone Company and the relay multipliers that we built for them and that the Telephone Company built for Aberdeen. I figured we had something like a 5 to 1 advantage over theirs with speed of calculation, which was quite satisfying.
Aspray:
IBM paid all the costs for the machine's construction.
McPherson:
Yes. Well, they didn't pay the cost, they gave it. I was a little upset with Mr. Watson — that after we had given them the machine, asked for Mr. Watson to raise money for the building. But after the gift he felt that was somewhat of an imposition. Asking too much.
Aspray:
What role did Howard Aiken play himself? Did he have any role in the actual construction of the components — in the design of the components — or the system? Or was that really Lake and his team of people?
McPherson:
I don't know whether or not I'm trying to answer the question because the full answer was arrived at by Mr. Watson over a six-month period of cross-examining everybody in IBM on this basic question. And there was a little pamphlet, which essentially was the result of possibly 40 revisions, trying to get the correct understanding of what part Aiken played in the machine. [I can get a hold of a copy of it and give it to you. It's a little black folder.] And it essentially just said that Aiken merely helped. No question about it. The conception of what was needed was there. He asked for some machine to do large-scale calculating. And he also helped in our understanding of what was needed in the way of mechanisms for computing the trigonometric functions, for instance, by tables or by calculations. I remember at that time we were just getting away from the idea that you do the whole job. All computing was done with tables, either logarithmic tables or trigonometric tables. I had a system that I wrote, a paper on IBM use of all of their computers for the mechanical tabulation of trinomials. And the idea of creating a table-making machine might be managed with original Babbage engines by pushing some cards repeatedly through a big tabulator. It came just after the need for tables no longer existed. It makes it kind of interesting. But I finally had the punched-card method working, where you set these up, and ran the cards through the tabulator. Somebody punched cards with the proper results, then reran those if necessary for the next stage. You could build up tables of tables by differencing. Since I've retired I have written a simple APL program to generate the thing on a personal computer. [Chuckling] It's a nice easy little half a dozen lines of code and the machine goes in and comes up with the answer. Either evaluating a single polynomial that was given for it, or a whole table that could keep it running long enough. Aiken's support was one of the things I wanted, to have an idea of what he had to do with it, having an idea of what was needed to do the jobs that he had in mind. He did spend a few weeks up in Endicott one summer with the engineers as they were describing to him what they thought — what their interpretation of his requirements were. He had a clear understanding of what we were doing. His original idea was that we would take a bunch of Monroe typewriters and Monroe multiplying machines and put them all together. He was delighted with the shift of it to an essentially integrated machine, which in a sense was the same thing but electrically controlled. In other words, a much more elaborate control — where the control was an integral part of the arithmetic function and number storage situation. So I think it was an electrical version of the Babbage machine. Although when you get down to fine points, there are some divergences.
Aspray:
Were the people who were building this machine, were they familiar with Babbage?
McPherson:
Bryce was. As a matter of fact. I finally found an old broken-down copy of his invention and it's quite clear from looking at that that Bryce might very well have gotten his idea for the original design of the carry mechanism in the early punched-card machines from the design in Babbage's work. The mechanisms appear to be identical. I would hate to raise the question of whether a lot of patents were issued that perhaps shouldn't have been. I was really fascinated with the fact that the carry mechanism, the reading mechanism, was something that was the essential way of doing things.
Aspray:
You don't know for a fact that this was the origin of Bryce's idea?
McPherson:
No, no. I just said that my conclusion was that the mechanisms were the same idea. And the fact that Bryce had the book and was the sort of man who I would have expected to invent it, made it possible. On the other hand, he may have felt — may have been surprised — the same way that I was if he ever ran into the same point. But the big point was really that Aiken essentially implied the position that he was the inventor of the IBM embodiment of his request so that receiving a gift was essentially receiving his due. [Chuckling]
Aspray:
I see.
McPherson:
And I read the remark in the latest British Computer Bulletin about Grace Hopper. They gave a nice two-page kind of tribute to Grace written from a distance. They made the remark there that Grace had written most of the manual that Harvard put together on the machine, which I hadn't realized before. I'd raised the question just a few weeks before the dedication, "Shouldn't we have something like this describing our machine?" Art Davis, our publicity man, said, "No, we shouldn't." For some reason he didn't want to mess in with it. But it just seemed to me that if we were going to have a big machine like that we ought to have some sort of a technical description of it to hand out. But, as I say, when I suggested it to him, he said, "No, that's in Ed Douglas's hands." This again was the same man who got me to IBM, who had been a vice president of IBM, and was a Harvard graduate. I guess what Art Davis was saying was that Ed Douglas was then the contact man with Harvard for Mr. Watson. Maybe he didn't want to get in any difficult spots. I think that must have been it. But I was quite surprised that what I thought was a very logical suggestion wasn't acted on, and it left the door wide open for the treatment we got, without it being intended, I'm sure.
Aspray:
You partly answered the following question, but maybe you can pull things together: What did IBM learn from building that machine?
McPherson:
I can only say that it learned that a big quantity of these Lake relays and Lake counters could run successfully on a major project, which shouldn't have been built if you were sensible. [Laughter] The things should have been tested out rather more thoroughly on a much smaller scale. But it was a gamble that paid off in the sense that it worked. And I think we learned something clearly about the possibilities of a stored program, whether it was stored on tape, or whether you were going to use tape storage for the necessities of the controller. And where you used electrical circuits freely to exchange information from place to place. It was a major test case of a mechanical relay and storage device operation. I certainly think it gave Mr. Watson the courage to essentially say we could do anything. I suspect that after the machine went to Harvard, that Harvard learned more than we did. I know that Aiken did, because he was able to follow the ideas here of using relays in the machine, and was able to essentially modify it with some additional relays — whose details I don't know — but I think they probably covered it in one of the later volumes of his series. After the machine went to Harvard, we took hands off, but we set up Donald Piatt, who was Lake's assistant engineer at the time as a person to essentially supply any needs that Aiken had for IBM assistance or equipment. And I think through Don we probably supplied substantial quantities of additional relays to speed up the multiplier.
Aspray:
This was even after the dedication ceremony?
McPherson:
Oh, yes! Two or three years after it. Aiken made modifications, which I would assume would be to speed up operations at some point or another. Because after all you don't want to go through 24 steps of multiplication when you want to multiply two 2-digit numbers. That was a point that should have been obvious. But it certainly wasn't part of the overall spec. All you had to do was just cut the thing in half. I think it was a 6-second cycle for a multiplication, roughly. Maybe it was 12, and it was cut to 4. It was an order of many seconds rather than thousandths of a second. You have to reorient to back there. Our multipliers were doing 900 multiplications an hour, and we were doing fine. [Laughter] When they were reading, computing and re-punching and then sorting another, we did better than that for the Aberdeen relay calculator because we started one calculation and the succeeding calculations before the first one was completed. Pass it down through a zigzag pass hole or line-up of the relay answers. Of course relay answers were faster than turning a counter around 360 degrees. I feel Aiken failed to get full value out of the Harvard machine. And although I shouldn't make the comment because I never really had a chance to analyze what were the totality of the things that machine was used for, to us it seemed as though it was used for almost 90 percent of its time on doing Bessel tables — Bessel functions. And, there again, the questions is, "Should they have ever been computed because just around the corner you didn't need to compute any more?" To compute the table, you just put in individual values.
Aspray:
Yes, you just did it when you wanted to.
McPherson:
And that's why it's weird. Now the question is, "When are we going to start to see multipliers cease to be multipliers?" And are you essentially going to take these big memories and do your arithmetic directly, giving you a factor of what is the answer? It's possible. It's been possible right along. We tried a little bit of that. In one of our intermediate computers we built we used a table look-up for arithmetic and a table look-up for multiplication. The sixteen hundred, I think, Wayne Winger's machine. We only built a few. It was the one decimal machine we had for a while. It was very cheap, but it wasn't very useful. It wasn't one of the top-of-the-line machines. But that's getting off the subject.
Aberdeen Proving Ground and W. Eckert
Aspray:
Yes. I understand that you had close association with what IBM was doing at Aberdeen. Could you tell me that story?
McPherson:
Well, in the interest of using IBM machines for calculations, I came across the Meeting of the American Mathematical Association; I think it was, up in New Hampshire, when some of the people from Aberdeen talked about the Bell Lab Stibitz calculators. The names of the mathematicians were such that you knew that there was some fire down there. [Chuckling] I took it on myself to get permission to go down and see if we could introduce some digital calculation in the work of the Proving Ground.
Aspray:
This would have been what year?
McPherson:
I think it was in 1940 probably. It may have been after the beginning of the war. I really don't remember. I'd have to dig out that Aberdeen file, which I didn't keep. All my basic records are in the archives up at IBM. They're available, but I just haven't looked at them. I don't recall whether I was in the Railroad Department still or whether I was in Future Demands. At any rate, I asked the General Sales Manager's permission to go down and see what I could do. I went down to the headquarters of the Proving Ground and said, "Here, I think you might be able to use some IBM punched-card machines. You have some over in the record-keeping office, main office. But I think they're useful in computation." In the meantime I'd learned about the modifications of the multiplier, which a Census Bureau employee had made, which gave us the ability to do multiplication and to round products — which seemed to me to be necessary — without the ends to the plus, and handle the final plus or minus calculations instead of just one sign. Those two changes made the multiplier useful, therefore, as a calculating instrument. On the basis of that being available, I made my pitch. They were a little reluctant to enter into it because they said, "After World War I, we'd all thought we'd lost all of our funds for ballistic calculations and if we go with the use of IBM machinery this way, we're apt to be deprived of it at the end."
Aspray:
Because they were renting it rather than buying it?
McPherson:
Well, yes, that would have been it. Anyway, the decision, which was finally worked out, was that the company would give them the machine for the duration of the war, as part of our war contribution, but would like rental afterwards. On that basis they accepted it and set up some additional equipment over in the Ballistics Research Lab, which had similar equipment with its own building somewhere along there. It was a General Simon? Well, he was a major or something about that grade when he was in charge of the Ballistics Research Lab. Leland Cunningham worked for him, and so did the other names you're familiar with, who never worked for the military exactly. But General Simon went on up pretty far as a result of the successful work they did in various ways. But anyway I opened the door there and got IBM in. In addition to the headquarters operation there over on Long Island, IBM added to the Lab calculating facilities. I went down there from time to time. One time was when Eckert and I went down together. When we were through that day, we ran into Johnny von Neumann on the train platform, both waiting for the same train back to New York. I argued about a decimal machine against his binary machine. Eckert was apparently familiar with von Neumann's work. He was new to me entirely. It was interesting. It was the beginning of our acquaintance. It was certainly of value, and I'd see him from time to time.
Aspray:
When did Wallace Eckert join the company?
McPherson:
He certainly became an employee of the company in 1945, but what I'm uncertain about is whether he was working for IBM when he went to the Naval Ordnance Lab, Almanac Office — the group in Washington that was doing the table work. My first thought is that he should have been working for IBM, but probably wasn't. We provided him with a set of machinery there for the Watson Astronomical Computing Lab. He may not have been an employee with IBM until after the war, but he certainly worked so closely with us that whether we paid his salary as Director of the Watson Astronomical Computing Bureau or not, I don't know. I think we probably did.
Aspray:
Is there anything more that you can tell me about Aberdeen? Were there additional changes made to the equipment for them during the course of the war?
McPherson:
The only changes we made for them that I know of were, essentially, providing this rounding and sign control on the multiplier — which was trivial. I don't think we put division on the machine. The only important thing we did during the war, after we got the punched-card installation going smoothly, was to give Halsey Dickinson the chance to build relay calculators for the Aberdeen Proving Grounds, which got there in the last year. So they were working for approximately the last year. Our point to doing it was that it would be helpful a) in the war effort, but b) it would also give us a more powerful relay computing operation which was quite diverse. Well, Halsey Dickinson had some ideas on multiplying and dividing which he was patenting which needed implementation in a hardware version so on the basis of that, there was the advantage to IBM of putting that project through as a special machine construction. It involved a slight change in the reproducer so that you could read and punch in the same card path. I guess the actual Halsey machine was actually built in Lake's department with Lake's electrical designer putting the stuff together.
I got my nickel in on that machine after it was all finished by pointing out that they didn't need carry relays because I think you could use an electrical version of the Babbage machine the entire way. The machine was probably built with the relays in, but they did get an efficient answer, which was that they could have left them out. As a matter of fact, if you went far enough, you would have a pretty heavy load of accumulated ones when you go from 99999 to 0 so you still might have wanted a repeater in there, but you didn't require a static relay. This was just a note between me and Ben Durfee. It never needed a patent. I always recall one of the chief things I had to do as Director of Engineering was never invent anything. This was a private one between me and Ben, but it didn't hurt my reputation with him. It just proved that I understood what they were doing.
And about the multiplier — those relay multipliers were really quite a step forward. They had dividing circuits as well as multiplying and worked fairly well. There's a story in the NTAC there of IBM contributions. You know, it was put in a publication, which only a few people interested in computing or computation for itself ever looked at or knew about. I've often thought that that's one of the things that the Annals of the History of Computing should probably revive. That is, go back and do nothing more than just essentially run an issue or a couple of papers describing — not just a bibliography — but one that essentially kind of reviews that history as it was put on a record. One of the sillier parts, of it was Eckert's scheme for verifying the air almanac. He printed some parts of the almanac and then you re-keyed the product and used the reproducer to verify the cards that went in and the cards the keypunch operator generated afterwards. It was about the most positive section you could get on anything, and it was there. It's in the NTAC. I think this whole story of computers and computing and its transmission from tables to mechanical machines to electronics is a fabulous story of what's happened.
Engineers at Endicott (NY)
Aspray:
This is an interview on the 12th of May, 1992 with John McPherson. It's a continuation of an interview that was started several weeks ago. Let's begin today's session by talking about some of those people that you encountered within IBM up until the end of World War II. Maybe you can begin by telling me something about Frank Hamilton.
McPherson:
Well, Frank Hamilton was one of the group of engineers at Endicott who, I believe, at one time was Mr. Lake's chief engineer's assistant. But I remember him best as an independent engineer with his own little area of expertise. He was one of the more accomplished people of the old school. He was not a scientist but called one, I'm sure, by Mr. Watson from time to time. That was one of the ways he used to emphasize his interest in lifting the level of technology in the company.
Aspray:
Was he mainly mechanically skilled?
McPherson:
Frank was the big mechanical, but he got pushed into a major part in the Harvard machine, which was finally called the Automatic Sequence-Controlled Calculator. At that time he must have been Lake's assistant. But he was a principal on it, and after that was done he got into the problem of further designs, particularly the initial efforts on the magnetic drum technology. He was in charge of putting together the SSEC, Selective Sequence Electronic Calculator.
Aspray:
What can you tell me about him as a person? Do you remember any stories or any personality traits or hobbies?
McPherson:
No, I don't really. He was just a very effective one of the senior engineers who I saw regularly.
Aspray:
What about Clare Lake?
McPherson:
Clare Lake was essentially the head the chief engineer at Endicott and had his own department for 20 years or more. He was the man who put together one of the early punched-card machines, and he was strongest in the early days when he was designing the straight punched-card system machines. I think he and Mr. Mills were the two key men. Mr. Carroll stuck in there with Lake. He had the best people — had the choice of the best assistants and designers, draftsman, on the product. He was a man who came up with a good many of the ideas, including the scheme for the improvement of the original electric counters. He reduced them to something about 2" by 2" and gave you a static readout of the device and the wire-contact relay which turned out to be a real success. I think it competed very favorably with the best of the Bell Labs designs, although it was much simpler but quite effective. It was the basis of all of our last group of punched-card machines, particularly the 407. He was very reserved, a very senior person, and he ran a kingdom, as it were.
Aspray:
What about Ben Durfee?
McPherson:
Ben Durfee was one of the good electrical people in Mr. Lake's department. He was the man who did the electrical circuit design or wiring of the various features that were included in the machines, both for controls and for selection operations. It was all immensely interesting. Lake, I think, started back in those days when you had to mechanically lock in a counter if you wanted to total — or turn out. I mean, you'd run a card, and then when you got to a place where you wanted totals, you'd move a collar around a shaft and engage the turnout — the reset mechanism to bring the counter back to zero. Instead of manual clearing with a hand crank, you just engaged it and called for a reset. But he went right ahead from those very simple concepts to a great deal of electrical circuit work. Ben Durfee was the man who essentially implemented most of it. He was the assistant engineer on the electrical side.
Aspray:
The other person I think of from this period is James Bryce.
McPherson:
Jim Bryce, of course, was part of the business, almost from the beginning of Mr. Watson's career with it. He was a brilliant inventor. His whole interest was in designing new features for the punched-card business, and in the course of doing that he had a very wide range of interests. I still remember first hearing about Babbage's computing engines, something Jim Bryce talked about. I think his invention record was over 400 patents. He was prolific and his job essentially was there at headquarters simply sitting down and trying to find better ways of making the machines and patenting anything that came along in ways of multiplying, of adding, of carrying information from one column to another in various forms. He dabbled in electronics, I believe, from 1920 on. Not deeply, but he kept looking for where it might be used in IBM. It came in very slowly, though. He was aided there by Halsey Dickinson, who did a great deal of the early electronic attempts to find ways of using it effectively in the punched-card business and who was a good electrical circuit designer of electronic devices.
He worked on transistors for a number of years before we finally got onto design. He used multiple tubes of essentially the scheme as Paul Palmer and Byron Phelps who built the first electronic counters, and electronic keypunch. Halsey had circuits for dividing that he wanted to patent. He'd been working on various electronic devices, which early in the war were considered significant enough to be given Patent Office secrecy restrictions so none of us could find out what he was doing. But when they wanted him to collaborate on the 603, we had to bring Phelps and a couple of other people in from Endicott and add them to Halsey's group so they worked on the job of putting together that first electronic multiplier. In fact it turned out to be rather key because it really teed off the interest in electronics that we'd been trying to find a use for. Electronics seemed to be too fast to match into the relay operation or even electro-mechanicals — even further off. It was a matter of two or three orders of magnitude difference in speed and that was a real handicap. The 603 was a kind of compromise. It was given a keypunch that worked a hundred cards a minute and the thing moved ten characters a second and enabled us to get some reasonable use out of the electronic circuitry.
Early Electronic Computing Devices
Aspray:
Do you think that part of the reason that IBM's early move into electronic computing devices was on the scientific side had to do with the fact that the electronics were used not in the input-output, but rather in the internal processing, in doing the arithmetic units? So that meant that you could take advantage of it more in scientific computations than you could in business computations?
McPherson:
Well, at that point in time, I think the initial interest in trying to use vacuum tubes or electronics was to try to get a better reading of the holes in the cards. We tried to run it on the collator at more or less the same speeds as the sorters, which is an area of 200 to 400 cards a minute. We made several efforts to try to avoid erroneous operation by using a group of tubes in the reading circuits to read quickly and to read without error, or without arcing, which is one of the problems of reading with brushes. A card going past a roller and a brush was a pretty elementary way of doing it, but it served up to that time.
Aspray:
When were these trials with electronics for reading done?
McPherson:
They were done in the early 'forties. Or maybe 1938. About then. I think you've got probably more information than I do about how Phelps got started. It was at Wally's suggestion, I think. Just building the first electronic counter unit, electronic adder. But that was very early, and at that time we were thinking of it essentially as an attempt to understand — to see if it could be done — without a clear picture of how we could use it. A couple of years later we got around to the idea that Halsey could make a multiplier and divider which worked electronically, and we went into the problem, with Bryce's department, of building a unit that was essentially a prototype, or a plate model we used to call them. Electronic doesn't sound quite right. There was a period of two or three years when we couldn't see any use in the product line on something we knew how to do — at least something that looked practical in the way of a group of tubes — doing the addition or multiplication, putting in the proper circuitry.
Aspray:
What was Mr. Watson Senior's attitude toward electronics at this time?
McPherson:
From the very beginning it was, "Let's try and use it." Let's try and use it? How can we use it? That's why he used to keep needling Bryce on this score. He said that back as early as the 1920s, which was just after the trigger, circuit was first commented on.
Aspray:
And Bryce's attitude was that, yes, it was a good idea, but it was hard to implement?
McPherson:
No, Bryce's idea was essentially, Is this something patentable? Let's get a patent. That was his mission in the company, to protect the company with patents on such future avenues of development. There's another person in this group that you ought to be interested in. There isn't much written about him, but John Hayward was a brilliant man. He was a very close friend of Mr. Watson's for all the years. He was a lawyer on the Board of Managers and a friend and advisor on essentially technical matters. He could read Chinese. I found out late in the program he was a Phi Beta. But his advice was good, and he helped run the development of our patents. He helped very often with the patent negotiations with the various outside inventors that came to IBM, or we found and looked into. John had time and again had the problem of helping Mr. Watson decide which were important and which patents we should acquire. When I was put in touch with him he used to give me the job of telling him some things, looking at some of the electrical or technical aspects of the thing. He was always interesting. Mr. Bryce was a practical inventor. Mr. Hayward was a good negotiator and gave good sound advice to Mr. Watson. And Mrs. Hayward and Mrs. Watson, of course, were friends. I think John Hayward and Mr. Watson may have actually roomed together before Mr. Watson was married. Their friendship went back a long way.
Other Key IBM Figures
Aspray:
Are there any other people who deserve special remembrance from the 'twenties and 'thirties? Well, 'thirties in your case since you weren't there in the 'twenties. But if one is keeping in mind the key figures of the 'thirties.
McPherson:
The key figures of IBM at that time would certainly include Mr. Carroll, who was the man who invented both the printing press that produced highly-accurately sized punched cards at fairly high rates, and of the sorter. The other major accounting machine developer was Al Mills, who was the man who built the first alphabetic accounting machine about 1930. He and J. Roy Pierce, from outside the company were the two people who built alphabetic machines. Pierce for Remington-Rand and Mills for IBM. One of the other men up there was. Samuel Brandt, who was building the 602 multiplier, which was an improvement over the original one, put together in 1930 by Lake's group. Sam Brandt followed up on some of the improvements of that machine and then was going to build a much better one, which was interrupted by the war. He had the design partly under way when Mr. Watson said, "I don't want any more development of IBM products during the war. I want you to use your services to help the government in any way you can." Sam brought the 602 out within six months after the end of the war. He had it that far along. Ray Johnson, in the meantime, had come along with his test-scoring device, and he went on, of course, to build and set up the West Coast laboratory and build the RAMAC, RAndoM-ACcess memory, and the first disk machines.
IBM Postwar Strategy
Aspray:
Let me turn to a different topic. Mr. Watson Senior wanted to do more work in electronics and he came to you to tell you that in 1944. I wondered if you could simply tell me more generally what kinds of actions was the company starting to take? People were realizing we were getting towards the end of the war by that time, and presumably some people were starting to think about peacetime activities again for the company. Can you tell me about that period and what the thinking of the company was at the time?
McPherson:
The place that I pick up my thread is just towards the end of the war. Mr. Watson wanted to essentially get a new accounting machine added to the line, and he decided that Mr. Fuller was the man to do it. He was an old NCR man, an inventor whom Mr. Watson had kept on the IBM payroll for a good many years. He had his own laboratory in East Orange and I think, lived somewhere in Llewellyn Park. He built the proof machine, which is a set of 24 adding machines with slots alongside of it to catch the checks as you added the values on the tape or on the adding unit. Mr. Fuller had been dabbling with the idea of building an accounting machine with pretty much that sort of a mechanical background.
Aspray:
Why did Mr. Watson think there was a need for a new accounting machine?
McPherson:
I can't answer that. [Chuckling] Maybe we wanted something new. We needed another way because during the 'thirties we'd been making rapid advances in essentially rounding out the punched, sort and tabulated system into the improved sorters and improved accounting machines. When Fuller came in IBM had just numerical machines — 5- and 7-bank tabulators and printers. He got an alphabetic tabulator going in the beginning of the 'thirties and then went on to the collator and the summary punch, which was a very useful addition because it enabled you to break down the end-of-the-month load that the punched-card system was kind of bad for. What was found was a difficulty. You had to have a battery of machines to take care of a two-day or four-day rush at the end of the month. And that wasn't very effective cost-wise. It was the fact that you could make the two-or three-way distribution of information offset it. It kept things in order. Mr. Watson was saying, "What's the next step we can do? Can we improve this?" We wanted a better accounting machine.
I'm a little puzzled now; I said that the first punched card and first alphabetic machine came up in the 'thirties. I think it did — I'm sure it did. It was pretty cumbersome, and Mr. Fuller, unfortunately, wanted to make it even more cumbersome by using a 5/32-inch spacing of characters, which wasn't either fish nor fowl. That was where I took the liberty of suggesting, as we were going back from the first visit out there in East Orange that we needed to do it at least on typewriter spacing, and we had to do it electrically. And Mr. Watson didn't say, "no," so we did it. He was quite satisfied with the result. Mr. Fuller was the head man, but Bud Levy and Page were the two engineers there at the time who could shift the gears and who came up with this wheel-printer concept, one on which you could print at much higher speeds than the older machines and which were more useful in the sense that you could use Lake's new electrical counters and relays extensively in building a machine which could be set up by the customer to do practically any sort of job. It was that which was the principal move we had in the works when the war came to an end. At that time, of course, we were a little puzzled as to what was going to happen with all the machines we had in the government — were they going to come back? In everything else the business had been lacking. The new business demands came in very heavily on the new printer for a while. There were not a lot of features in the 407 that were quite exceptional, but at that time there was no serious concern except for this area of competition with Harvard: how vacuum tubes or electronics would fit into the business. Mr. Watson wanted it. It didn't look like a commercial product or project was there, so the Commercial Research people weren't particularly interested. They thought the punched-card business was not going to be affected or that the electronics could really pay off.
SSEC, IBM-603 and IBM-604
Aspray:
I suppose the first step towards an electronic product was the SSEC then, is that right?
McPherson:
No, there were two things running in parallel. The SSEC was essentially a one of a kind and looked upon as something that wasn't owned by the government that business could use. But they could use it by coming to one machine and running their problems there. And that they'd probably researched. They'd be just a sideline of the business.
Aspray:
Was there any thought that it would pay for itself?
McPherson:
Actually the use of the SSEC was essentially a gift of IBM's, I think, in almost all instances. I mean, I think there was no fixed fee, but there was a problem of getting approval to use it simply because the amount of that sort of work was fairly limited, and we felt it should be used to try and help the people who didn't have government access to anything government might have had in the same area, which we assumed of course was probably greater than we knew.
Aspray:
Who was behind the idea for the SSEC? Whose idea was it to build?
McPherson:
Well, Frank Hamilton was called upon to come up with it. Fortunately, he had Rex Seibert who had been at Harvard, but who wanted to essentially make the Harvard machine — the original Automatic Sequence Control Calculator — essentially a sequence-controlled machine, and had been turned down by at that point by Aiken. At least he was unhappy with the reception of his doing it there. He had been hired by IBM; I don't know whether he quit first up there or not. But he was at the Watson Lab when Mr. Watson said, I want an answer to the fact that Aiken is going ahead and building more computing machines, his Mark II and Mark III. So we did. That whole area was one of essentially knowing what we could do with tubes.
But at the same time — and in fact earlier that we were going to work on the SSEC, we had the 603, which was a box of 300 vacuum tubes, which was originally built, as the initial plate model, of a multiplying and dividing machine. The first machine we had built in New York had dividing in it as well, but because of the problems of the amount of tubes they'd had in the machine, the answers when you did division would only give you three-place accuracy because of the comma-shift problems, so we didn't announce the division feature as part of the product, but we knew it could divide, and we knew if we built a little bit bigger one, we would have that. I've always felt a thrill that we showed Mr. Watson and Tom the machine in March. Mr. Watson said, "We'll show them at the business show in September." We had three machines made by September 1st. We announced the machine then and took orders. We shipped the first machines to customers within nine months before the end of that year, which was an unheard fast response. And while all it did was take care of all five payrolls and things like that, it had our feet in the door. We were exposed to the use of tubes in quantity in a business product. And Red Dunwalter got in the picture of the first research people.
Wesley and Warebridge and Dunhill got in the business of designing a 604, to specify the 604 which was a 500- or a 1000-tube machine which really put their stored program on 20 or 30 or 50 steps, which you could wire. So you could do a considerably elaborate computation. Essentially our way into electronics was via the 604. And then almost immediately after it got out, the card program calculator which would essentially reach out and let you put a much larger program on punched cards which, again, controlled the 604 and gave you some additional mechanical counter-like storage of additional data for the 604. That was big. Then after the 604 we built a 607 which, I think, was the first transistor version of an electronic machine, but it stood on its own feet and was used just as the 603 had been in the sense of a multiplying and dividing machine. It put us up against the problems of what kind of transistor would last and how do you get them from being $15 apiece to 5 cents apiece?
Aspray:
One last point before we go on. You mentioned being able to build and start shipping within a nine-month period of time. What made that possible? What was different about the way that you designed and built this project?
McPherson:
In the first place, all we had to do was buy some tubes and set them in a suitcase on some sort of a rack. But we also had — and this is where Mr. Hayward comes in — he had a nephew down in Washington who was in the Navy and, knew some of the problems that they had in the Intelligence area. Mr. Hayward had shipped the problem to the East Orange lab without my even knowing about it. It was probably secret by necessity. But what they did was took a reproducing punch which had two feeds, one for the cards to be copied and a blank one to punch the holes and they modified that machine so that you could read and then punch in the same path, the card path. I happened to know that was in existence, and that was the perfect tool for the 603's input-output mechanism, because you took a card, took data from the card, and before it moved 3 inches, you had the answer, ready to slap into it. That was the key to the 603's usefulness. Instead of having a column-by-column keypunch, which had a maximum speed of 900 and an average speed of 500 operations an hour, you jumped up to 6,000. And this was enough to essentially justify the 603 as just a pure multiplying machine.
I called attention to the fact that even on this 8-by-8 multiplier, we got two small multiplications at the same time if you spread the numbers far enough and ignored some of the garbage in between where they didn't overlap or where you didn't need to get the overlap. But that was another thing I've always thought we'd missed ever since; that all through the war and the years before, we had a permanent activity in Commercial Research called Pointers, which every week or two came out with a four-page suggestion on how somebody had done something tricky with the existing equipment both showing him how to do it, but also explaining enough about it so they could look at other places where the same idea could be generated just by the way the machines were wired on the plug board. We kept all our customers in close contact and the rental business — with the product they were paying rent for. We've lost it since then.
Aspray:
What was your role in the 603, 604 and in the SSEC developments?
McPherson:
My role was rather minor in the sense that Dave Rubich was the man who came up with the suggestion that we all make Halsey stop playing with individual tubes and try to make a useful multiplier. And my part was only in getting Halsey. Now Halsey, I think, says that Mr. Watson told Mr. Bryce to do something about the electronic multiplier. But what we did was make sure that Halsey did something that was useful in that connection, and he did. Later on he gave us the designs for the relay calculators, of which there were several — five of them, I think, built. That was a slightly different project but again, that was one Ben Durfee put together after Halsey provided the scheme for operating it. But again, it was a very interesting and ingenious adaptation of the relays. They built a sort of a whiffletree so that you could solve a problem going down the whiffletree as you did these digit-by-digit multiplications. But before you got the first answer down and it would be finished, you could start a second one on the next . So this was pipeline, if you want to call it that, in the relay fashion. How Halsey came up with it, I don't know. But he came up with it, and Ben Durfee wired the relays through. I had the job of putting them together, and Endicott put them out. My part in all those things was really to just know what needed to be done and let it go ahead.
Aspray:
That was also true of the SSEC?
McPherson:
No for the SSEC we really had a get-together of engineers. And after they'd discussed it, Mr. Watson essentially put the job to Hamilton to specify it. But with the advantage that now Hamilton has the advantage of essentially a consensus of Eckert and the Watson Lab people as to what was needed to make a better machine than the ASCC. I was just in the background, just keeping in touch with the project and encouraging its progress. Perhaps people thought I was running front for them, but whether I was or not, the results — that's all I had. In other words, very much of this was — the old-style where Mr. Watson said: "I want it."
Watson's Inventions
McPherson:
I was also fielding every day or two more of Mr. Watson's inventions, most of which we couldn't do anything about but which we had basically had to have somebody look at. Try and see what they could come up with as a practical solution to the things he wanted us to build.
Aspray:
These were all computing devices?
McPherson:
Well, they were anything.
Aspray:
They were in all sorts of fields?
McPherson:
One that I particularly remember was a railroad ticket machine. We should design a small punched-card ticket. Use that for all railroad passenger traffic. But they were just endless.
Aspray:
These were his own ideas?
McPherson:
Where he could identify a need, it became an invention. He could essentially twist it around just to the point where it would be an invention. One of the wilder ones was the instruction I had to electrify the abacus for Madame Chiang-Kai-Shek whom Mr. Watson had seen the mere night before. Yes, we had a couple of engineers work on that for quite a while. They didn't do it, but out of it we did get two or three balls from a Chinese typewriter. It was just a drum that had 6,000 characters on it, on which you had to be able to know a four-digit code to get the word printed. But it was there. It was an attempt to try and get something done regardless of whether it was practical or not.
Aspray:
Do you think of any other examples of some of these ideas you investigated?
McPherson:
No. I'm afraid not because most of my problems were trying to forget them. [Laughter] And not get beaten over the head for it. I can assure you that anything we could do and was useful, was acted on. But there were an awful lot of them that had to be ignored.
Role at IBM to 1950
Aspray:
What was your attitude about the SSEC? You said that Mr. Watson wanted it and therefore went ahead, essentially, and there was consensus of the engineers. But what was your own personal view about it?
McPherson:
My view then was definitely very much in favor of doing it. I felt, as Mr. Watson did, that too much of the advances were in the hands of the government. There were a lot of things we didn't know at the time, but we assumed that the government, that they may have had other efforts. We had the Bureau of Standards' efforts.
Aspray:
Are you thinking of SEAC?
McPherson:
Yes, SEAC. And SWAK.
Aspray:
Who kept track of what was going on at all these other research centers in the 'forties in electronic computing? Did somebody in particular in IBM have responsibility for that?
McPherson:
I don't know whether I had the responsibility, but I took it.
Aspray:
You took it?
McPherson:
No, I made a serious effort to keep in touch with what did appear. I'd been in the habit for years before that of reading the Bell Labs Records and two or three other Telephone things like that, The Bell Labs — Bell System — Technical Journal, and there's an English one. And I'd already bumped into this NTAC activity and followed what they were doing. I'd call on the Works Progress Administration Table-making Project that used punched cards. Did you ever hear of the story about the fact they threw me out?
Aspray:
Is that right?
McPherson:
It was talked to me. I remember that fellow was Lang _____?
Aspray:
Who eventually became part of the Bureau of Standards activity?
McPherson:
Yes. But when it was in New York they didn't want to do anything other than have a lot of people do it step by step. I've made a number of trips to Europe over the years because at the time of the Airlift I was attempting to attend the first computer sessions at NPL, National Physics Lab in Teddington. Lawrence Hanos joined me on two occasions. We got interested particularly in the Williams storage tube. It was our job to really. We really went over and tried to talk to as many people in England and on the Continent as would talk to us.
Aspray:
Do you remember the people you spoke with about that?
McPherson:
I still have some notebooks on the visits. Actually, at one of the later meetings, Arthur Samuels was asked by Wally to send him notes of what we did, and he did the job for both of us. I didn't keep as good notes as he did, and I didn't write them down. I just kept a little diary that I had over there for all of those trips.
Aspray:
So did you meet people like Maurice Wilkes, at the time, and Tom Kilburn?
McPherson:
Yes. Well, I met the Manchester people. We worked through the NDRC over there. Lord Halsbury and his patent man took us around to see the TRE people up in Great Malvern. We talked to Lee Owen. The whole group. We didn't go to Bull because they were definitely a competitor. We were a little cautious about bothering them or getting into an awkward situation with them. Now, I see we have an agreement with them, which I would have never dreamt could happen.
Aspray:
Did you happen, by any chance, to meet Alan Turing?
McPherson:
No. I read a lot about him, but we didn't run into him.
Aspray:
What were your regular duties during this period from the war until 1950?
McPherson:
I made it my business to keep in touch with developments in the IBM field of interest — computing primarily — and working with Watson and Wallace Eckert. I think he [Eckert] looked on me as his right-hander. I wouldn't say he reported to me because he was reporting to Mr. Watson. But he was keeping me fully informed and looking at me for help in any way that I could. We made a couple of the recruiting trips up to M.I.T. at the end of the war together to get some of the new, really good people to join us at the Findlay Radiation Lab in Boston.
Aspray:
Who were some of those people?
McPherson:
John Lentz, Bye Havens and there was a third person. Bye Havens was by all means the most important, the most capable. Have you seen our little history of the Watson Lab?
Aspray:
Yes.
McPherson:
That's a pretty accurate description of all the people who joined us.
Aspray:
How close contact did you have with the designers of computing equipment at the other American laboratories? Did you go to see the labs? Did you contact the people?
McPherson:
With Art Samuels's help again, I got the first computing meeting going. I was chair of it. In 1951? The first Joint Computer Conference.
Aspray:
Oh, yes! And you were the chairman? I didn't know that.
McPherson:
Yes. The next year Norm Taylor came down to talk about the first meeting and what they could do to help with the second. He also at the same time wrote me the question: Would IBM help on the Sage project? So I introduced the Sage project part. I should get the finder's fee on that one. [Chuckling] I never did. Once I was a vice president, I didn't really have any clearly defined job but to do the best I could and try to move things ahead. After Wally succeeded me as Director of Engineering, he'd occasionally give some of the chores that he thought I could handle and handle well. I conducted a couple of study groups that were useful and I ran programming systems for a couple of years. This was when they were partly in Poughkeepsie and partly in New York and before Coster said, "Oh, we're going to have to have a much bigger activity than just the Fortran Group and an operating system." I kept worrying about whether we were doing the right thing in programming.
IBM-6407 and IBM-407
Aspray:
This was much later, though? This was towards the end of the 'fifties, right?
McPherson:
Yes, this was in the 'fifties now. During the late 'forties I was primarily concerned with keeping the 6407 from being scrapped which is what actually almost happened. The Sales Department didn't think highly of it because it looked like an expensive machine, and it looked to be about double the cost of the older machines — 800 against 400. Charlie Kirk and the sales managers were at one point in agreement — they didn't want it. Mr. Watson finally took the matter in hand and told me to come in just before a Board Room meeting of everybody on the subject. The day before, he called me in to say, "How good is the 407? It has more type wheels?" "Yes." "It's faster?" "Yes." "Twice as fast." It adds what it prints, which was a feature, which was lacking in the past. In the past we'd made a minor and a major total. You'd use separate counters to add the two numbers and reset the counters as proper — one counter more frequently than the other. So it looked as though all the information was from the same source, but occasionally the master counter was wrong or failed. You just figured that had to be rerun. So anyway, Mr. Watson said, "That's 8 times as much — 2 times 2 times 2. You tell them at the meeting tomorrow that's how good it is." I did. I believed it. I still believe it. And I'm still very proud of it because I thought then I'd lost it. If Tom had ever heard this story at that time, I would have figured, "I'm done for life. I've ruined my future." I would have gotten out — as far as he was concerned.
Aspray:
Why do you say that?
McPherson:
Because I think Tom [Junior] was siding with the boys who wanted to stop it.
Aspray:
I see.
McPherson:
I can't accuse them of that, and this you can't write up please, because to this day we've never talked about this. I don't think he [Tom] ever realized how valuable it was to the company. After all, it was his father's last big achievement. I was interested it, and I sold it. I followed it as the years went by. I got enough figures from the accounting people to determine that the revenue from that, of the 407, and its accompanying machine that was numeric only, made it the first product that brought in a billion dollars revenue for IBM. This was by 1954-55. Those were the years we grew the fastest. If you look in Tom's book, it says that. There was an occasional "why we did." Except a lot of people wanted our machine. But this really gave me a lot of personal satisfaction. I think that today we get a billion dollars first year of a product, but in those days we were wondering about how soon — after the war we'd make a billion dollar revenue. Once it came and the electronics came, it just seemed to be unstoppable.
Career until Retirement (1970-1)
Aspray:
Why don't you trace your career for me, from that point on?
McPherson:
I was made a vice president in 1948 and went abroad with Mr. Watson and Lansing Dick and a group of IBMers that was head of the foreign part of the business, and another associate of mine, Doc Granback. We were there all summer. When we came back, I just continued with my activities with the Watson Lab, contacted with the engineering people but not on an intimate basis or a scheduled basis as I had been before. Shortly after that I was asked to care of fine programming for a period of two or three years. I remember going to Detroit before I was made an officer. And I was following up some of our special developments. One of our special developments was a stabilized gun mount for tanks under Ward Leather's direction. I was out at the Detroit Automotive Arsenal, when I was called back to New York for a meeting the following day by Mr. Phelps — he was a vice president. George Phelps had been Secretary of the company at that time. I woke up in Albany station and said, "I know why I've been called back" and it turned out to be true. Tom had had a promotion that day, but I also wasn't told I was going to be made a vice president, and it wasn't a surprise. This was one of these weird second guesses that you can't forget. Well, as I say, I wasn't very constructive. I was usually involved in running some of these things for Wally.
Aspray:
It sounds like you're being very modest to me. You know, the company had many talented people, and they recognized production. So you must have been doing a very effective job for the company.
McPherson:
Well, I had the confidence of everybody I worked with. And I very carefully made a point of never inventing anything. Yet I made the effort to understand — I wanted to understand — everything was really being accomplished in each of the projects. This support was the sort of thing that Mr. Watson had done when it was very small, but he later found he had less and less time to do it, and had to devote quite a bit of his time to the sales end chiefly people because they were numerically fairly much larger. But when it came around 1960, Bill Lawless suggested we have this Systems Service, Systems Research Institute, for training our systems service people in what systems were. I handled that for five years as titular head. I think that Frank Beckman was better — was a real teacher — and he came out and showed them some programming systems and actually picked out most of the people. But I took my offices over there for five years. I tried to make sure that we had a good clean record and ran it efficiently.
And after five years of running SRI and getting the Systems Journal started, I succeeded in getting assigned back to Armonk and I tried there to set up some competition for the Telephone Company as a supplier of special electrical relay gear and that sort of thing. They didn't like it. They didn't accept it. We sold one or two information systems to them where it was essentially in handling directory work. "Information, please." I think Southwestern Bell were the only people who put it in, even a small chunk of it. It seemed to me that we should have become involved. The Telephone Company did so poorly with their ideas and desires to automate the billing of toll charges — it was a logical place for us to try to worm in there. But their management — Bell Labs people — just didn't want it. I can't blame them for freezing us out. Then I got very interested in the APL business, and from 1966 on I was captured. It started in 1960 when I first came down and taught APL in one of our early classes — first or second. Abe Holcroft took over and gave it another time. I started out from not believing it, or as accepting the general view of it as being too complicated, to realizing that it was a very important and continuing possibility. I still think that we need it and will need it in the future.
Aspray:
When did you retire from the company?
McPherson:
I think it was January 1st 1970 — or was it 1971?
