Dealers of Lightning Read online

  CHAPTER 3

  The House on Porter Drive

  In May 1969, about the time Xerox shareholders voted to approve the purchase of SDS, the provost of Washington University in St. Louis was reaching the end of his rope.

  The academic year just ending had been the most trying of George E. Pake's career. The 1960s were not easy on anyone in a college administration, but Pake felt that unrest on his own normally placid liberal arts campus had reached a high-water mark during the previous semester. A group of protesting students had occupied the chancellor's office. Someone tossed a Molotov cocktail into the ROTC Quonset hut and burned it to the ground. Pake spent the year contending with all sorts of reactionary trustees and alumni who, he recalled, "wondered why we didn't just fire the students and keep the faculty." As spring drew to a close, he said, "I was a case of battle fatigue."

  In a more peaceful era Pake would have seemed the ideal college administrator. Narrow-shouldered and retiring, he possessed a clipped and slightly distracted manner of speaking that reinforced his donnish air. But this diffidence was deceptive. When the faculty got fractious he could dig in his heels and hold his ground, especially when called upon to uphold his standards of academic propriety. Fairness, he insisted, was the key. No administrator of a large academic institution could possibly know enough to mediate every issue purely on academic or scholastic grounds. The trick in refereeing among powerful faculty with their overdeveloped intellects and underdeveloped social graces was to remain unyieldingly impartial. When all else fails, split eveiything down the middle.

  This was a skill he had tried to hone in the years since he had come to St. Louis from Stanford, where he had held a physics professorship. Washington University had installed a dynamic new chancellor deter­mined to enhance its reputation as a first-class academic institution, and Pake had accepted his call to join the crusade as a senior administrator. At first the change fed his idealism. He imagined himself promoting the social benefits of higher education in ways that would be closed to him if he remained merely a teacher and laboratory researcher. But by the spring of 1969, when he was next in line to succeed that chancellor, he had also become profoundly disillusioned.

  "I hadn't visualized myself as running a command post in a military oper­ation," he said. "I knew I did not want to be a candidate for chancellor, not to lead that goldfish-bowl kind of life. My wife would have hated it."

  So that fall he returned to teaching. On Thanksgiving weekend, just as he was finally re-acclimating to the milieu of classroom and chalk­board, he got a phone call from his old friend Jack Goldman.

  "George," Goldman said, "I got a proposition for you."

  Jack Goldman's relationship with George Pake dated back twenty- five years to when they had worked together on wartime projects at Westinghouse Research Laboratory, Goldman as a senior fellow and Pake as a Westinghouse undergraduate scholar.

  After the war Goldman remained in industry while Pake moved on to Harvard for his doctorate. But they kept track of each other's careers within the insular community of working physicists. Just as he was start­ing his search for a director for his new research center, Goldman heard that Ford had offered his former job to Pake, and that Pake had turned it down. Goldman guessed Pake's reasoning had something to do with Ford's erratic commitment to basic research. "I figured I could make a better case for Xerox than Ford could make for Ford," he recalled. After a speaking engagement in Chicago he swung down to St. Louis in Xerox's new Sabreliner corporate jet (his favorite mode of transportation). "I met Pake at the airport, invited him aboard the company plane, gave him a couple of drinks and proposed that he join up."

  In truth, Goldman's pitch was more focused than this breezy descrip­tion suggests. From the airport the two of them repaired to a nearby hotel, where Goldman spent the better part of a day spinning a seductive vision of computer research conducted in a pristine setting with Xerox's copious cash. "We're talking real money, George," he said, showing him the growth plan for a lab that would employ 300 professionals within four years.

  Pake hesitated, wondering about Xerox's resolve over the long haul. "I had a lot of friends at other industrial research establishments and the usual thing they were worried about was the feast-or-famine effect," he said later. "You know, in the good business years the company invests in research, but in the bad years they want to pull out." That was a recipe for wasting millions of dollars. "Research is a steady-state thing. You can't just turn it on and off."

  Goldman tried his best to be reassuring. He reminded Pake that in 1944 the Haloid Company had offered Chester Carlson research support when no other corporation would. Carlson had scarcely anything to demonstrate the potential of his invention other than a tiny scrap of paper on which he had duplicated his own scrawled "10-22-38 Asto­ria"—the date and place of his first successful xerographic copy. The company that was now Xerox had invested in that improbable invention for fifteen years before the first Model 914 came off the production line in 1959 and made its fortune. Long-range research? Was there an enter­prise anywhere in the land that understood it better than Xerox?

  "Yes, but it seems to me the corporation has got it backwards," Pake replied. "If you're going into the computer business, you should have got the researchers first to help you identify the right corporation to buy."

  "Unfortunately," Goldman said, "it's too late for that."

  Pake ended the meeting insisting how deeply he enjoyed the life of a college professor in St. Louis, but he was beginning to crack. Before Jack Goldman reboarded his plane he extracted Pake's agreement to visit Rochester and Stamford to meet Peter McColough and the chair­man, Joe Wilson. If anyone could charm his wavering quarry into join­ing the company, they could.

  'When I went back I asked Peter McColough why he wanted to start a new research center," Pake recalled later. "I said, 'You've got a research center here that has developed xerography. To build a new one you'll have to have a new research library and new research machine shop and all the other things. Lot of fixed costs you have to duplicate. Wouldn't it be easier to expand the laboratory in Rochester?'

  "McColough turned to me—and I remember this conversation very well, it's indelible in my memory. He said, 'George, I think these peo­ple here in Rochester have had a heady success with xerography. But I'm not sure they're adaptable enough to take on new and different technologies. If we're going to bring new technologies into Xerox it would be better to do it in a whole new setting.'"

  McColough's reply might have come directly from the Jack Goldman playbook. Despite himself, Pake was utterly taken with McColough and Wilson and deeply flattered by their apparent willingness to place him in charge of a multimillion-dollar corporate asset after only one interview, especially since he told them he would expect to be held to a liberal standard of success.

  "I said if you hire me you will get nothing of business value in five years," he recalled. "But if you don't have something of value in ten years, then you'll know you've hired the wrong guy."

  Pake understood that managing a research center devoted to finding a common ground between his first love, physics, and the intriguing new field of digital computing was a once-in-a-lifetime opportunity. In comparison, the charm of closing out one's teaching career in the Mid­west seemed meager indeed. Just after New Year's Day 1970, he tele­phoned Jack Goldman to accept the job.

  The first order of business was to find a site.

  Goldman's plan to locate the lab in New Haven had collapsed even before Pake came aboard. Yale, as it turned out, was afflicted by a strain of that old malady known as the town-gown syndrome more virulent than Goldman had suspected. The university, it was true, was famous for the snobbery of its faculty, but he was still shocked at its unfriendliness to enterprises located outside the grimy stone campus walls. Faced with the prospect of being shut off from the very resources for which he sought an academic setting in the first place, Goldman decided to look elsewhere.

  Several other po
ssibilities were culled early. These included Webster, where Goldman feared his new lab would come under the intellectual domination of the copier bureaucracy still entrenched in Rochester. Also rejected were Princeton; Stony Brook on Long Island, where the State University of New York was building a new campus; and several other East Coast sites that were either too far from an established Xerox facil­ity or lacked the cachet Goldman craved for his would-be Bell Labs.

  Pake directed Goldman's attention westward. Teaching at Stanford in the early 1960s had given him a glimpse of the phenomenon that would shortly make the Santa Clara peninsula famous as "Silicon Val­ley." A few weeks after signing on, he proposed that Goldman charter the company plane for a California excursion. Ardent corporate way­farer that he was, Goldman agreed with alacrity. Soon he and Pake were working their way south from Berkeley to San Diego, stopping at every major university campus in search of the ideal spot.

  But at Berkeley there was no available real estate to support a corpo­rate research facility. At Santa Barbara, where a new state university cam­pus was sprouting on the dazzling coastline, there was real estate but no major airport. "Oxnard . .. dismal," Pake recalled. "Pasadena . .. Smog was terrible. Xerox had a division called Electro-Optical Systems there with a fairly big site but it was not something that could interact with Cal-tech—too industrial. So we didn't see anything very encouraging."

  That suited his purposes fine. For the whirlwind tour on which he led Goldman was mostly window dressing. Pake's primary objective was Stanford and its vibrant home town, Palo Alto. Goldman had initially ruled out the site for lack of any nearby Xerox facility, but Pake goaded him to reconsider. He knew from experience that the university was anx­ious to develop strong relationships with the industrial enterprises springing up like anthills all over the valley. Then there was the salubri­ous physical and cultural climate—not a trivial consideration if one hoped to attract gifted researchers to an embryonic lab.

  As for Goldman's objection that Palo Alto was too far from any Xerox property, Pake countered with a neat equivocation: Let proximity mean being close enough to reach a Xerox facility in time for lunch. SDS was in Los Angeles, an hour's flight from the Bay Area. Anyone could leave Palo Alto in the morning, lunch at SDS, and get home in time for dinner. And was not the original rationale for the lab to be SDS's research support?

  "That's a very interesting thought," Goldman said, bowing to the inevitable. In early March, Pake invested his first two staff members, a pair of administrative officials from the Webster research division named Richard E. Jones and M. Frank Squires, with the task of flying to Palo Alto and finding a building suitable to rent.

  "Nobody at Webster wanted the job," Rick Jones chuckled, remem­bering how he became PARC's first official employee. "I was the admin­istrative manager at research and development in Webster. Everyone else had kids in school in Rochester and I only had a nine-month-old son. I had married a Rochester girl in 1966, but when I said, 'How about leav­ing Rochester and moving to California?' she said, 'Sure.' Squires was similarly unencumbered by a growing family, having only recently mus­tered out of the service, so Jones tapped him as personnel manager.

  On their first reconnaissance trip they found that cutbacks in govern­ment and military research spending had left plenty of vacant research facilities to choose from. In a couple of days they visited thirteen empty locations before settling on one in a development known as Stanford Industrial Park. This was a compound of one- and two-story buildings occupying a parcel of land the cash-strapped university had decided to

  "He also acceded to Pake's repudiation of the designation "Advanced Scientific and Systems Lab," the name the lab bore in Goldman's original proposal, in favor of the bucolic-sounding "PARC." As Goldman acknowledged later, "The acronym of the for­mer would have invited ridicule." lease out to small businesses. It was located just beyond the campus boundary, in a dale surrounded by orchards and horse pastures where the grass had turned brown in the dry peninsular spring. Its main street, Porter Drive, meandered in gentle curves among the squat industrial buildings before disappearing over a low hill.

  About halfway down Porter stood a two-building complex that had been vacant since the Encyclopedia Britannica moved out a couple of years before. Facing the street was a cinderblock building windowless on two sides and with a concrete floor sturdy enough to support heavy lab equipment. Behind it was a somewhat larger structure that presented an exterior of floor-to-ceiling plate glass to the bright California sun. Trailing behind the rental agent, they stepped inside, disturbing a layer of dust and filth that seemed to have remained untrammeled since the begin­ning of time. A musty stench pervaded the air. The floor, littered with pieces of crumbled ceiling tile, traced a large square around an interior courtyard adorned with one lonely olive tree. Squires and Jones contem­plated the squalor, which was illuminated by a few dim rays of sunlight straggling in through streaks in the windows. The place needed work. But every other site they inspected would have needed more. And at a total of 25,000 square feet, the two buildings together were the roomiest they had seen. They gave the real estate agent a handshake deal and flew home to Rochester to pack up.

  In mid-May Jones returned with his wife and infant son as Pake's advance guard. He temporarily parked his family a mile or two from the site at Rickey's Hyatt House, a motel on El Camino Real that would serve as a transitional home for scores of PARC recruits over the next dozen years. After picking up the keys to his new workplace from the rental agent, he headed over to the property. As he coasted up the long driveway he could make out a stranger peering through one of the big windows.

  "Can I help you?" Jones asked.

  "I must be lost," the man replied. "Do you know where there's a Xerox research facility around here?"

  "You're at it," Jones said.

  "Really?" An expression of grave doubt passed over the stranger's face. He introduced himself as Frank Galeener, a newly graduated Ph.D. from Purdue who had been hired as a materials scientist in the new physics lab.

  "Oh, right, I recognize your name," Jones said. "But you're not sup­posed to start for a couple of weeks."

  "No ... I was in the area and thought I'd stop by and see what it looked like." He cast another anxious look through the window at the debris- strewn interior.

  "We're not set up just yet," Jones hastily reassured him. "But don't worry. It's going to be great."

  "Thank goodness," Galeener said. "For a minute there I thought I'd made a terrible mistake."

  Jones, Squires, and Gloria Warner, a senior secretary who relocated from Webster to work for Pake, spent the next week working like char­women. With brooms, buckets, and mops purchased from the nearest K Mart, they swept up the accumulated filth themselves and installed a rickety table and chairs in the clearing. A day or two later a man showed up from Pacific Bell to install the first telephone and a van arrived from Webster with a load of surplus oscilloscopes and other castoff equipment that Jones had redeemed from the Webster storage sheds. By the time Pake arrived on June 30 a local contractor had been in to fashion a few office cubicles and a large library space out of the bare interior. The next day Xerox's Palo Alto Research Center officially opened for business at 3180 Porter Drive.

  Pake had also been busy. While closing out his final semester's teaching obligations at Washington, he wrestled with the challenge of getting up to speed on the science of digital computing. He felt like an old dog trying to learn new tricks. "I was starting from scratch," he said. "I had to ask around to find out who are the good people, what are the big issues and so on. But I did wony because I was not a computer scientist."

  He did, however, know one person who boasted a first-rate familiar­ity with the Young Turks of the new discipline: Bob Taylor.

  Pake had met Taylor in 1964, back when Washington University under­took an unusual rescue operation for Wes Clark, the MIT computer pio­neer. Among Clark's idiosyncrasies was a visceral antipathy to the concept of time
-sharing. ("I'm one of the oldest continuing floating objectors in the business," he once told an interviewer.) Time-sharing, he believed, encouraged institutions like universities to lust after grander and costlier machines tiiat were by their nature inefficient for the small-scale work students and professors typically did. Their only virtue was that they could be paid off by overcharging every user for his or her time-slice of the entire behemoth, no matter how much of it the user actually employed. Thus was computing rendered more remote and intimidating than ever—a backwards trend exemplified in Clark's view by the arche­typal system at MIT: "That of a very large International Business Machine in a tightly sealed Computation Center: The computer not as tool, but as demigod." What Clark found even more troubling was that subdividing the main processor, as time-sharing did, rendered impossible the sort of display-based research that Ivan Sutherland had achieved so spectacularly on the TX-2. No user of a time-shared computer could ever monopolize the processor long enough to drive a coherent visual display as Sutherland had. (Clark allowed the TX-2 to be shared, but only serially—you signed up for a block of time on it, but during that period the entire machine was yours.) Time-sharers were limited to communi­cating with their machines via teletype, because the sluggish rate at which people typed was what gave the system the necessary opportuni­ties to shift its attention from one customer to another between key­strokes.