“I can’t really talk to you about this, Tom,” he said. “You just don’t understand the typewriter business.”
That was as good as telling me that we had to accept losing money all the time. So I went to Dad and said, “You can’t go on with that guy. The only way he can run the business is in the red. Let’s get somebody else in.” I had in mind H. Wisner Miller, a man I had known since before the war. Wiz was a few years older than I, and I admired him because he’d had to fight real adversity. He came from a prominent family and he’d been a Princeton freshman in 1929 when the stock market crashed. His father lost everything and Wiz had to quit school. The only job he could find at first was selling vacuum cleaners door-to-door in the Bronx. One of IBM’s directors knew him and introduced him to Dad, who hired him to sell typewriters because he liked Miller’s spirit.
Choosing Wiz to run the typewriter division had been a big gamble. Dad went along with me, even though it meant jumping Miller from a fairly low position over men Dad knew a lot better. But Wiz had exactly the sales approach necessary to sell those machines. The IBM method for selling punch-card systems was much too analytical for electric typewriters. You couldn’t overcomplicate it. What Wiz brought was zip and enthusiasm and leadership. I loved to watch him inspire his men at sales conventions. He’d have a typewriter onstage under a spotlight, all by itself. Wiz would walk up in his blue serge suit, look at it, and extend a finger to flick away an imaginary speck of dust. Then he’d step back and say, “A magnificent machine. I hate to see even a speck of dust on that machine. It’s so beautiful.” He taught his salesmen to use this blarney on the secretaries, and started making the typewriters in different colors like red and tan. He even made a white typewriter that my father presented to Pope Pius XII. A lot of systems men called Miller corny and simplistic, but he was one of IBM’s great sales leaders. In 1949 the machines caught on and for years after that the division grew at a 30 percent annual rate. My first important personnel move had been a great success.
The newspapers in the late 1940s were full of talk about laboratory computers with funny names like BINAC, SEAC, MANIAC, and JOHNNIAC. Scientific conferences on computing and electronics were jammed. IBM had no plans to build such machines, but we kept hearing of projects at American and English universities, big-name companies like Raytheon and RCA, and some tiny startups whose names nobody recognized. All of the new machines were cumbersome and enormously expensive, none was intended to be sold commercially, and for quite a while the ENIAC, the celebrated University of Pennsylvania machine that Charley Kirk and I went to see, was the only computer that actually worked. But this didn’t stop people from speculating what the “giant electronic brain” was going to mean for mankind.
Of course, calculating devices had been around since even before the Chinese abacus was invented. And there were a few giant calculators, such as the MARK I machine IBM built for Harvard during the war, that could do a wide variety of mathematical jobs. But they did it, in essence, by counting on their fingers. Their inner workings were electromechanical, like those of an ordinary tabulating machine. When ENIAC was unveiled it created a huge stir because it was fundamentally different. It had no moving parts, except for the electrons flying at close to the speed of light inside its vacuum tubes. All these circuits really did was add one and one, but that’s all they needed to do. The most complicated problems of science and business often break down into simple steps of arithmetic and logic such as adding, subtracting, comparing, and making lists. But to amount to anything, these steps have to be repeated millions of times, and until the computer, no machine was fast enough. The quickest relay mechanism in our punch-card machines could only do four additions per second. Even the primitive electronic circuits of the ENIAC could do five thousand.
This boost in speed promised to change the lives of everyone who worked with numbers—I heard one engineer compare it to the difference between having one dollar and one million dollars. A Time magazine writer who was at ENIAC’S debut wrote that its “nimble electrons” opened a whole new frontier. Until then there were known principles in science and engineering that no one used because they required too much figuring. Aircraft designers, for example, knew perfectly well how to predict wind resistance theoretically. But it was so much trouble to do the calculations that they used the hit-or-miss method of making scale models and testing them in expensive wind tunnels instead. So when ENIAC appeared, people had visions of computers helping to break the sound barrier, predict the weather, unlock the secrets of genetics, and design weapons even more terrifying than the A-bomb.
My father initially thought the electronic computer would have no impact on the way IBM did business, because to him punch-card machines and giant computers belonged in totally separate realms. A computer revolution might sweep across the scientific world, but in the accounting room the punch card was going to stay on top. Dad was like the king who sees a revolution going on in the country next door to his own, yet is astounded when his own subjects get restless. He didn’t realize that an old era had ended and a new era had begun. IBM was in the classic position of the company that gets tunnel vision because of its success. In that same period the movie industry was about to miss out on television because it thought it was the movie industry instead of the entertainment industry. The railroad industry was about to miss out on trucking and air freight because it thought it was in trains instead of transportation. Our business was data processing and not just punch cards—but nobody at IBM was smart enough to figure that out yet.
I don’t mean to say that Dad totally ignored the challenge that computers posed. He believed that no one could beat IBM when it came to building giant calculators for science, which is all he thought the new computers were, and he set out to prove it. In the spring of 1947, when I was still only a vice president and spending most of my time overseeing the sales force, he called in the engineers who had worked on the MARK I machine for Harvard. Dad told them he wanted a new “super calculator” that would be the “best, fastest, biggest—better than the Harvard machine, certainly better than ENIAC.” They could use vacuum tubes if that would make the machine work better, but he wanted it finished in eight months.
Dad had his engineers pretty intimidated, and they didn’t dare ask for more time. Instead they went flat out. They put all their other projects aside, worked virtually around the clock for the rest of 1947, spent almost a million dollars—and built a machine that worked. It was called the Selective Sequence Electronic Calculator, or SSEC, and it was a weird gigantic hybrid of electronic and mechanical parts, half modern computer and half punch-card machine. In was 120 feet long, with 12,500 tubes and 21,400 mechanical relays. It could do the equivalent of ten years of paper-and-pencil work in an hour. In some ways the machine was extremely innovative: it earned a spot in computer industry history as the first big calculator ever to run on software. This made it much more practical than the ENIAC: you could switch the SSEC to a new problem just by feeding instructions into its memory, whereas programming the ENIAC involved manually resetting hundreds of switches on its consoles. Yet because of its mechanical innards, the SSEC was a technological dinosaur—it was inherently slower than the all-electronic ENIAC, and speed was the thing users craved.
Trying to make sure the SSEC would get as much public attention as the ENIAC, Dad had it installed in our showroom on the ground floor of IBM headquarters in Manhattan, in full view of the sidewalk. Passersby on Fifty-seventh Street could look in the window and watch the SSEC work. It was an amazing sight to come upon in the middle of the city—three long walls filled with electrical consoles and panels, all studded with dials, switches, meters, and little neon indicator lights that flashed whenever calculations were going on. Hundreds of people stopped to watch it every day, and for years it was the image that popped into people’s minds when they heard the word “computer.” When Hollywood started to put computers in science fiction movies, they looked just like the SSEC—even though it didn’t exactly qualify as
a computer. Dad dedicated the machine “to the use of science throughout the world” and ran it on a nonprofit basis. Anybody with a problem of “pure science” could use it for free; for anybody else—such as an oil company needing to do a statistical analysis of a drilling field—there was an hourly charge of three hundred dollars that covered the machine’s operating costs.
The great missionary for the SSEC was an influential Columbia University astronomer named Wallace Eckert. In the late 1920s Eckert (who was not related to Presper Eckert, the ENIAC inventor) pioneered the use of punch cards in solving scientific problems. He was a small, retiring man, easy to underestimate. But he played a major behind-the-scenes role in the fight against German submarines during World War II, by calculating naval almanacs of unprecedented detail. These navigation tables enabled convoys under attack in the North Atlantic to determine their positions quickly and precisely and radio for help. Eckert became the first scientist with a Ph.D. on the IBM payroll. After the war my father hired him as director of pure science and set him up with a research lab next to the Columbia campus. His work gave many scientists their first glimpse of the possibilities of machine computation and brought large numbers of people to the SSEC.
Dad really thought that the SSEC was the calculator to end all calculators. And in a way, that was true. It was like a vintage car I once owned, the Stanley Steamer—quite remarkably advanced for what it was, but not the technology to carry the day. The SSEC marked the end of an era at IBM. It was the last great achievement of a talented group of inventors who had spent their lives working for Dad. They’d designed the punch-card machines on which IBM built its success, and now they’d produced one of the most advanced machines ever. But even though they reached the threshold of the computer age, few of them stepped across. The SSEC was built in splendid isolation. Its design was kept secret, so that in spite of its success, it did little to change IBM’s image in the technical community. The new generation of electronics engineers continued to think of us as a stodgy company that was wedded to punch cards and the past.
My father was highly skeptical when ENIAC’s inventors, Eckert and Mauchly, quit the University of Pennsylvania and went into competition with IBM, setting up their own company in a Philadelphia storefront. But before long it was clear that they were good salesmen as well as brilliant engineers. They named their new machine the Universal Automatic Computer, or UNIVAC, and claimed it was going to be useful in both the laboratory and the accounting office. The first UNIVAC wasn’t due to be ready for years, but with nothing more than a paper description Eckert and Mauchly won financial backing from two of our ten biggest customers—the Census Bureau and Prudential Insurance—and at least one other insurer besides. When Dad found out about that, his skepticism turned into fury.
On the Wednesday before Labor Day of 1947, I walked into his office to find him ripping up Frank Hamilton, one of our senior engineers. There was a secretary present who took down the scene verbatim. Father started in by saying, “Now, I understand these fellows who built the ENIAC machine are being backed by insurance companies to build something for them. Why don’t we build a machine to meet their specifications?”
“I think we intend to do something on that,” Hamilton said. He was acting a little sullen, because my father was forgetting that he and the other engineers had been working day and night trying to get the supercalculator built. Meanwhile Dad was getting madder and madder.
“We can’t think and intend when insurance companies are backing this outfit to build machines! We can’t afford just to think about and intend to! This business wasn’t built that way! What is the quickest way to go ahead and build a machine to meet their specifications in the very shortest possible time?”
“The best plan is to survey the specs and see what they want.”
“We know their specifications and we have already lost three months on it. If we can’t build it, let’s drop out. If we can do it, let’s do it at a price those other fellows can’t meet. If we can’t build a machine and give it to them on a better basis than anybody else, then we are not entitled to the business. It’s an indictment against IBM to have these two fellows backed by those insurance companies!”
Hamilton finally realized how angry Dad was. He started yessing to save his life. “There is no question that we can do it,” he said. “Not a bit of doubt about it.”
I knew what was upsetting Dad the most about the UNIVAC design. He felt it was an insult to our main selling proposition: the IBM card itself. Eckert and Mauchly were saying that punch cards were not appropriate for use with modern electronic equipment. Instead, the UNIVAC was going to store data on the new medium of magnetic tape—the same stuff being used in the early tape recorders of the day. This method was still largely unproven, but it was called for in almost all the new computer designs. Eckert and Mauchly explained to customers that magnetic tape had big advantages over punch cards. First, it was fast—it could pump data into and out of a large-scale computer at a rate more closely matched to the speed of the electronic circuits. Second, it was compact. A single reel the size of a dinner plate could hold the policy records of an entire insurance district, which normally required around ten thousand punch cards stretching several yards.
I doubt that Frank Hamilton had much time off that Labor Day. The following Tuesday he appeared, looking haggard, at a meeting convened by Dad in the big walnut-paneled boardroom next to his office. All of IBM’s officers were there, along with the company expert on insurance industry accounting. Hamilton presented an ambitious plan for a machine to go up against UNIVAC. It would use tape in combination with punch cards, and would cost $750,000 to build. People gasped. From a punch-card standpoint, this was a staggering amount. Our average installation in those days cost about $20,000 to manufacture and rented for about $800 a month; by that yardstick, the rent on a $750,000 computer would have to be $30,000 a month!
Dad complimented Hamilton on his hard work and then picked the plan to pieces. It was obvious that he didn’t like Hamilton’s machine because it resembled the UNIVAC. Having built his career on punch cards, Dad distrusted magnetic tape instinctively. On a punch card, you had a piece of information that was permanent. You could see it and hold it in your hand. Even the enormous files the insurance companies kept could always be sampled and hand-checked by clerks. But with magnetic tape, your data were stored invisibly on a medium that was designed to be erased and reused. Imagining himself in the customer’s shoes, Dad said, “Why, you might be going ahead and thinking you are storing information on that magnetic tape and when you try to get it off, you might find you have nothing there!” Frank Hamilton’s design died on the table while Dad told the marketing men to call on Prudential and persuade them that the UNIVAC idea was not sound.
At that point I wasn’t sure that building computers like the UNIVAC—or abandoning punch cards for magnetic tape—would ever make business sense. The new computers were clumsy, extremely expensive, and involved so much exotic and unproven technology that there was a real chance they might never be dependable enough for business use. I shared many of Dad’s misgivings, but I was compelled by the tremendous speed of electronic circuits. Customers were snapping up our little 603 Electronic Multiplier, which Dad had put on the market at my urging. Compared with the UNIVAC, it was a tiny mouse of a machine, designed to fit in with ordinary punch-card equipment and renting for only $350 a month. But it was a success—my first at IBM—and I thought it might be a sign of things to come. We had a small team of electronics engineers in Poughkeepsie working on an improved version called the 604, but I started getting concerned that we might not be doing enough. I suppose on some level I was thinking that I was only thirty-three years old, and that a decade or two down the road IBM might run out of string.
In 1948 I got even more nervous. My friend Red LaMotte sent up a letter from the Washington office saying he’d assigned a man to attend engineering conferences around the country. By that man’s count, there were no
fewer than nineteen significant computer projects underway—most of them involving magnetic tape. Red said this made him wonder, “Inasmuch as IBM is the leader in the field of calculating equipment, does it not seem reasonable that it too should be kept abreast of all developments by active participation in this field?”
I started getting warnings from customers as well, that the punch card was headed out the window. Jim Madden, a vice president of Metropolitan Life and a friend of Dad’s, invited me downtown to his office. “Tom,” he said, “you’re going to lose your business with us because we already have three floors of this building filled with punch cards and it’s getting worse. We just can’t afford to pay for that kind of storage space. And I’m told we can put our records on magnetic tape.” Roy Larsen, the president of Time Inc., said much the same thing. I was working for him in the New York Hospital Fund drive, and his company was one of our big accounts. He explained that the success of Time and Life magazines depended on their getting to millions of readers each week while the news was still hot. Time Inc. was using IBM equipment to handle the mailing lists and address labels, but each subscription required three punch cards, and with the lists growing by thousands of new subscribers each month, the machines were barely keeping up. “We have a whole building full of your gear,” Larsen told me. “We’re swamped. If you can’t promise us something new, we’re going to have to start moving some other way.”
I didn’t think it would be prudent to run to Dad with the idea that punch cards were dying. He’d have thrown me out of his office. Instead I used a systematic approach that I knew would make sense to the old man. In 1949 I organized a task force of eighteen of our best systems experts to study whether we should add magnetic tape to our product line. With Dad it was almost a religion that ideas for improving the product line should come from customers. Of course, customers weren’t always asking for the same thing—some wanted the machines to be faster, some wanted them to print more neatly and handle more carbon copies, some wanted them to be less noisy—and if you panicked and did everything they suggested, you’d go broke. You needed to sort out what improvements made sense and would really pay. The task force studied the magnetic tape issue for three months. When they came back, their answer was that punch cards were the best thing in the world for accounting jobs, and that magnetic tape had no place in IBM. I tried again, bringing in top salesmen and describing what magnetic tapes could do, but they all ended up saying no, it’s better to use punch cards. They gave me nothing I could take to Dad.
Father, Son & Co. Page 22
