Broad Band Read online

  Pres thought John Mauchly’s vacuum tube idea was interesting. Everybody knew vacuum tubes were too delicate for use in a computer—like light bulbs, they blow out—but Pres figured that if the tubes weren’t pushed to their limit, they’d hold steady. The pair started designing circuits. John took a teaching position at Penn, happy to be closer to Pres, and, once settled, discovered the human computers at the Moore School. In their backbreaking calculations, he found the perfect application for his vacuum-circuit computer. He dictated a proposal to his secretary, Dorothy, who shunted off a memo to the university’s civilian liaison with the military.

  The memo was lost in the shuffle, as memos sometimes are. It wasn’t until an informal conversation between the differential analyzer’s maintenance man, a friend of John’s from Ursinus College named Joe Chapline, and Colonel Herman Goldstine, a military liaison with the Ballistics Research Laboratory in nearby Aberdeen, that the idea came up again. When Chapline mentioned his friend John’s electronic computer, Goldstine saw the potential immediately. He tried to hunt down the lost memo, to no avail. Fortunately, the secretary, Dorothy, managed to re-create it from her shorthand notes. Like most secretaries in her day, she was trained in shorthand, a type of rapid writing, which is scribbles to the uninitiated. If it weren’t for Dorothy’s ability to code and decode what was at that time a largely female language, the original proposal for the electronic computer might have been lost.

  The reconstituted memo was brought to army brass, who didn’t need much convincing. John and Pres secured their funding in 1943 and started building the ENIAC right away. They hired engineers and former telephone company workers, who were good with relays, but most of the people who actually wired the ENIAC were women, part-time housewives with soldering irons on an assembly line. The most important hires were the human calculators, chosen from the best of the Moore School group, who would translate the ballistics computations they knew so well for the new machine. Nobody thought much of assigning women to this job. It seemed only natural that the human computers should train their own replacements. Further, the ENIAC looked like a telephone switchboard, reinforcing the assumption that its “operators” should be women, their task “more handicraft than science, more feminine than masculine, more mechanical than intellectual.”

  By 1944, construction on the ENIAC, then known as “Project X,” took up most of the first floor of the Moore School building. One night, Pres and John conducted an after-hours demo for one of their new hires, Kay McNulty. They brought Kay and a colleague into a room where—behind a sign warning HIGH VOLTAGE, KEEP OUT—two of the ENIAC’s accumulators were wired together by a long cable with a button on the end. One accumulator displayed a five. They pushed the button. The five jumped to the other accumulator, moved three places over, and transformed to a five thousand. John and Pres looked excited. Kay couldn’t see why. “We were perplexed and asked, ‘What’s so great about that?’ You used all this equipment to multiply five by one thousand,” she said. “They explained that the five had been transferred from the one accumulator to the other a thousand times in an instant. We had no appreciation of what that really meant.”

  It meant that the ENIAC could calculate at speeds previously unimaginable, by human or machine. And although it was funded by the military to churn out firing tables as fast as the army could manufacture guns, the ENIAC was much more than a ballistics calculator. Pres and Mauchly had designed a general-purpose computer—think of the difference between Charles Babbage’s one-note Difference Engine and the speculative Analytical Engine, which so entranced Ada Lovelace. It could perform an essentially limitless number of computational functions, as long as new programs for it were written. In its time at the Moore School, it would calculate the zero-pressure properties of diatomic gases, model airflow around supersonic projectiles, and discover numerical solutions for the refraction of shock waves. Hardware may be static, but software makes all the difference. And although it took some time to settle in, that truth came with a corollary: those who write the software make all the difference, too.

  The ENIAC Six were an odd mix, thrown together by the circumstances of war. Betty Jean Jennings grew up barefoot on a teetotaling farm in Missouri, the sixth of seven children, and had never so much as visited a city before pulling into the North Philadelphia train station. Kay McNulty was Irish, her father a stonemason and ex-IRA; Ruth Lichterman, a native New Yorker from a prominent family of Jewish scholars; Betty Snyder, from Philadelphia, her father and grandfather both astronomers. Marlyn Wescoff, also a Philly native, had been hand calculating since before the war, and she was so adept that John Mauchly said she was “like an automaton.” They all met for the first time on a railroad platform in Philadelphia, on their way to the Aberdeen Proving Ground, a marshy plot in Maryland the army had converted into a weapons testing facility. Bunked together, they became fast friends. Even after long days training on the IBM equipment they would be using to tabulate and sort ENIAC data, they stayed up late talking about religion, their vastly different family backgrounds, and news of the secret computer. “It was just a great romance, I think,” John Mauchly hazarded when asked why these women volunteered for a job they knew so little about. “There’s a chance to do something new and novel—why not?”

  The reality might have been more pragmatic: in the 1940s, a woman with mathematical inclinations didn’t have many options in the job market. When Kay McNulty approached her college graduation, she had a hard time finding any employment that might make use of her math major. “I don’t want teaching,” she explained. “Insurance companies’ actuarial positions required a master’s degree (and they seldom hired women, I later found out).” If the only other options are teaching math at a secondary-school level or executing tedious calculations for an insurance company, the opportunity to work in a brand-new, relatively well-paying field represented a hugely exciting change of scenery for all the women who signed up.

  Computing was so new a field, in fact, that none of its qualifying attributes were yet clear. During her job interview with Herman Goldstine, Betty Jean Jennings recalled being asked what she thought of electricity. She replied that she had taken a college physics course back in Missouri and knew Ohm’s law. No, no, said Goldstine, from behind the desk: Was she afraid of it? The job would require her to set switches and plug in cables, he explained, and he wanted to make sure she wouldn’t get spooked by all the wiring. Betty Jean said she could handle it.

  The ENIAC Six trained on paper, writing programs for a machine they hadn’t met. When they were finally shown the finished ENIAC in December 1945, what they encountered was a massive, U-shaped assemblage of black steel housed in a room big enough to hold it along with some miscellaneous furniture. It had forty panels, grouped together to create thirty different units, each addressing some basic arithmetic function: accumulators for addition and subtraction, a multiplier, and a combination divider and square rooter. The sprawling visual effect of the machine was overwhelming. Programming, the six learned, would not be a desk job. The women would stand inside of the ENIAC to “plug in” each problem, stringing the units together in sequences using hundreds of cables and some three thousand switches.

  There were no instructions to read, no courses to take. The only manual for the ENIAC would be written years later, long after the women had reverse engineered it from the machine itself. Built by electrical engineers, the ENIAC came with nothing but block diagrams of circuits. Just as Grace Hopper had before them, they taught themselves what to do, becoming hardware adepts in the process.

  They started with the vacuum tubes and worked their way to the front panels. Betty Snyder borrowed maintenance books for the machine’s punch card tabulator from a “little IBM maintenance man by the name of Smitty,” who told her he wasn’t allowed to lend them out but did anyway, just for a weekend, so she could figure out how the ENIAC’s input and output worked. They found a sympathetic man to let them take a plugboard apart a
nd make their own diagram for reference, even though his supervisor wasn’t sure they’d be able to put it back together again (they were). It was hot and there was construction everywhere, including in the room above the one in which they worked. One day John Mauchly popped in and said, “I was just checking to see if the ceiling’s falling in.” They started going to him with questions, and eventually made headway.

  Knowing how a machine works and knowing how to program it are not the same thing. It’s something like the difference between an intellectual understanding of internal combustion and being a fighter pilot. John Mauchly and J. Presper Eckert essentially built a jet, gave the keys to six women without pilot’s licenses, and asked them to win a war. It was daunting, but it presented an opportunity for the women to claim space for themselves in a field so young it didn’t have a name. “At that time it was new and no one knew what to do,” explained Betty Jean Jennings. Not even the men who designed the ENIAC had given much thought to how it would run. They’d ignored the actual workflow of setting up problems. In 1973, Mauchly himself admitted that he and Pres had been “a little cavalier” about programming, saying that they “felt that if we had the machine capable . . . there would be time enough to worry about those things later.”

  As it turned out, Mauchly found other people to worry about those things—six people, in fact, in wool skirts and thrilled by the challenge. “How do you write down a program? How do you program? How do you visualize it? How do you get it on the machine? How do you do all these things?” wondered Betty Jean. It would be up to the ENIAC Six to figure it out.

  Betty Jean Jennings (left) and Frances Bilas (right) operate the ENIAC’s main control panels.

  Today, programming can be tricky, but it’s accessible. To write code, you don’t need to study circuit diagrams, take apart components, and invent strategies from scratch. Instead, you simply need to learn a programming language, which acts as an intermediary between coder and machine, just as a shared spoken language can bridge a gulf of understanding between people. You tell the machine what to do in a language you both understand; the machine then translates and executes your commands on its own. The ENIAC had no such language. The computer accepted input in only the most elemental of ways, and so the ENIAC Six rolled up their sleeves and met the machine on its level. As Betty Jean Jennings recounted:

  Occasionally, the six of us programmers got together to discuss how we thought the machine worked. If this sounds haphazard, it was. The biggest advantage of learning the ENIAC from the diagrams was that we began to understand what it could do and what it could not do. As a result we could diagnose troubles down to the individual vacuum tube. Since we knew both the application and the machine, we learned to diagnose troubles as well as, if not better than, the engineer.

  Unlike Grace Hopper, who managed a team of operators punching her handwritten code into the Mark I’s tape loops, the ENIAC Six moved around inside the great machine itself. They replaced individual burned-out vacuum tubes from among thousands—several burned out every hour, despite Pres’s design—fixed shorted connections, and wired control boards. They wrote programs, feeding them gently into the machine with much trial and error. The job required a combination of mechanical dexterity and mathematical know-how, to say nothing of organizational skills: punched cards containing the ENIAC’s programs needed to be sorted, collated, tabulated, and printed. The word “programmer” didn’t exist yet, but Betty Snyder thought of herself as a “cross between an architect and a construction engineer.” Betty Jean Jennings was more blunt. “It was a son of a bitch to program,” she wrote.

  Unfortunately, none of this effort did the U.S. Army any good. Although it ran a number of one-off calculations, the war ended before the ENIAC became fully operational as a ballistics calculator. In peacetime, however, the ENIAC was no longer secret, and the computer was unveiled to the public in 1946, with much fanfare and two different demonstrations. The first, for the press, was by all accounts a bit lackluster. The second, for the scientific and military community, was a hit, thanks largely to a demonstration of a trajectory calculation programmed by Betty Jean Jennings and Betty Snyder.

  The two Bettys, as they were sometimes known, were the aces of the ENIAC programming team; after the war, they both went on to long and pioneering careers in the commercial computer industry. As was common in the history of human computing, the pedagogy of the Moore School emphasized working partnerships, with teams of two people seeking out errors in each other’s work. Betty Jean and Betty were ideal partners, because they delighted in finding each other’s mistakes. They both wanted perfect code and never let their egos get in the way of achieving it. “Betty and I had a grand time,” Betty Jean wrote in a memoir. “We were not only partners, but we were friends and spent as much of our free time together as possible.”

  A few days after the first ENIAC demonstration, Herman Goldstine, their military liason, and his wife, Adele, invited Betty and Betty Jean over to their apartment in West Philadelphia. Adele trained the human calculators at Penn and had always struck Betty Jean as an impressive, big-city woman; at the Moore School, Adele lectured sitting on her desk, with a cigarette dangling from the corner of her mouth. Betty Jean was surprised to find the Goldstine apartment rather ordinary, with few personal touches and a set of twin beds. As Adele served the Bettys tea and the Goldstine cat leaped uninvited onto their laps, Herman asked them if they could set up a ballistics calculation on the ENIAC in time for its unveiling to the scientific community twelve days later. It was a big ask, and Betty Jean sensed that Herman Goldstine was nervous about the demonstration. Well-known scientists, dignitaries, and military brass would be there, and everyone was keen to see that the ENIAC worked as advertised. Not much has changed, it seems, about the way tech keynotes are anticipated and prepared.

  The Bettys asserted vigorously that yes, absolutely, they could make it happen. They were bluffing. Although they’d spent the last four months working out a ballistics trajectory program on paper, they hadn’t actually plugged it into the ENIAC yet, and they had no idea how much time the transfer would take. They started the next day.

  Betty Snyder was twenty-eight; Betty Jean Jennings had only just turned twenty-one. They knew they’d been asked to do something important and that everyone they worked with was counting on them. The pair worked around the clock for two weeks, living and breathing the trajectory program. Their colleagues Ruth Lichterman and Marlyn Wescoff supported them by hand calculating an identical trajectory problem on paper, mirroring step-by-step how the ENIAC would process the calculation. This would help the Bettys debug the ENIAC if it made any errors. Men popped by with offerings: the dean of the Moore School left them some scotch, and John Mauchly came in on a Sunday with a bottle of apricot brandy. They didn’t really drink—maybe a Tom Collins on special occasions—but Betty Jean kept a taste for apricot brandy for the rest of her life.

  The night before the big demonstration was Valentine’s Day, but the Bettys didn’t go on any dates. Their ENIAC program had a massive bug: although they’d managed to model the trajectory of the artillery shell perfectly, they couldn’t figure out how to make it stop. When their imaginary shell hit the ground, the mathematical model kept going, driving it through the earth with the same velocity and speed as it had while shooting through the air. This made the calculation worse than useless. If they didn’t find some way to stop the bullet, they’d embarrass themselves in front of eminent mathematicians, the army, and their employers. In desperation, they checked and rechecked settings, comparing their program with Ruth and Marlyn’s test program, but they were stuck. A little before midnight, they left the lab. Betty took the train home from the university campus to her house in suburban Narberth. Betty Jean walked home in the dark. Their spirits were low.

  But Betty Snyder had one trick left: when stuck on a logical problem, she always slept on it. Wearily, she spent her hour-long train ride home that night considering the
artillery problem and its various potential solutions. When she fell into bed, her subconscious began to untangle the knot. The next morning—February 15, 1946—Betty arrived at the lab early and made a beeline to the ENIAC. She’d dreamed the answer, and knew precisely which switch out of three thousand to reset, and which of the ten possible positions it should take. She flipped the switch over one position, solving the problem instantly. Betty could “do more logical reasoning while she was asleep than most people can do awake,” marveled Betty Jean.

  The ballistics trajectory demonstration was a huge success, thanks to the Bettys’ clever ballistics program and a little old-fashioned razzle-dazzle from John and Pres, who placed halved Ping-Pong balls over the ENIAC’s neon indicators. During the demonstration, staff dimmed the lights in the room, showcasing the ENIAC’s thinking face in feverishly blinking orbs of light. The program was faster than a speeding bullet, literally: the ENIAC calculated the trajectory in twenty seconds, faster than it would have taken a real shell to trace it. The Bettys and Kay McNulty hustled over to the tabulator, made printouts, and handed them out to the audience as souvenirs.

  The event made headlines. The women were photographed alongside their male colleagues—they remember flashbulbs—but the photos published in newspapers showed only men in suits and military decorations posing with the famous machine. The press had a field day with the ENIAC, presenting it as a fruit of the war effort unveiled for the better living of the American people. Because of their unfamiliarity with computing, journalists called the ENIAC a “giant brain” and a “thinking machine,” a mischaracterization that has persisted in the popular consciousness, enthusiastically supported by science fiction writers, ever since. The ENIAC couldn’t think. It could multiply, add, divide, and subtract thousands of times per second, but it couldn’t reason. It was not a giant brain. If there were giant brains in the room, they belonged to the people who built—and ran—the machine.