Dealers of Lightning Page 9

  Or occasionally too free. One morning in early 1971 the weekly meet­ing was addressed by Jack Goldman, who was in the habit of paying fre­quent visits to his new incubator. Goldman's talk—perhaps inspired by PARC's recent close call—had to do with the need to start generating for­mal reports and white papers to reassure Stamford that the money being spent out west was buying genuine intellectual achievement.

  Someone Thomburg did not recognize interrupted Goldman with a suggestion. "He said, 'Well, if you ask me, Jack'—the rest of us never called him anything but 'Dr. Goldman'—'If you ask me, Jack, what we should do is build a computer-based query system where we can tag the different levels of the report, so somebody who just wants an executive summary could get that and someone who wants more could get the full report.' He was basically talking about a hypertext-like environment. We were all sitting there thinking this is pretty good stuff, and Goldman was up front, chomping on his cigar, saying 'Yeah, that's a good idea.'"

  The moment the meeting broke up Thornburg saw his friend Bob Bauer shoot out into the hallway. Curious, he followed, and finally found Bauer leaned up against a wall, laughing so hard he could hardly catch his breath.

  "What's so funny?" Thornburg asked.

  "You know that guy who said, 'Well, if you ask me, Jack?'"

  "Yeah, who is he?"

  "He doesn't work here," Bauer said. "He just came over from Stan­ford to have lunch with somebody in computer science. They said, 'We got a meeting, stick around,' so he followed them in. Goldman is prob­ably going to want to give him a bonus or something, and the guy doesn't even work for Xerox!"

  But the paradise of collegiality was more mythical than real, or at least it was destined to be short-lived. Jack Goldman had not acceded to Pake's desire to have physicists and other traditional scientists on the premises because he subscribed to any notion of marrying the old sci­ence to the new. Instead he saw it as a way to rapidly ramp up PARC's head count by kidnapping available talent from Webster while recruit­ing the computer and systems experts he was counting on to make PARC's reputation.

  "The idea was that when you brought new people in you wanted them to have someone to talk to," Goldman said. "So we sort of seeded the two scientific departments"—that is, physics and optics—"with people from the company, a few of the shining lights from Rochester who were desirous of moving. They were very good guys and I suffered a certain amount of criticism for taking them away from Webster and essentially lowering the average IQ of the Webster group." But he sensed that the physicists and computer scientists would end up in a profound philo­sophical and scientific tug of war. If Pake believed he could paper over such an elemental conflict, Goldman thought, he was mistaken.

  The hiring freeze ended after a couple of months. In that period the downsizing in the research industries had sharply intensified, in part because the Mansfield Amendment restricting Pentagon spending to specifically military research had begun to bite nationwide. Pake and Squires resumed recruiting with the same cautious deliberation as before.

  In a superb buyers' market for research and engineering talent, PARC's lavish budget and open-ended charter stood alone among corporate enti­ties. Other industrial research centers might enjoy generous funding or comparably liberal charters, but none had both the open checkbook and apparent immunity from product development pressures enjoyed by PARC. From the point of view of the nation's outstanding computer research scientists, Xerox—outside of a handful of top universities—was the only game in town. "All the super-bright guys who had swell ideas were tickled pink to go work for George Pake and Jack Goldman," recalled George M. White, the research executive on Goldman's staff.* "Nobody was going to float money and start a company for them, like

  *'No relation to George White, the PARC researcher.

  they would today. At PARC they could get a good budget and a good lab and independence, all of which Pake and Goldman provided.

  Adding to PARC's charm was its premium pay scale. This was partially the result of shrewd entreaties to corporate management by Pake, who feared that Xerox policy requiring PARC's salary scale to match Web­ster's, dollar for dollar, would allow the most prestigious universities to outbid PARC for the best talent.

  Pake urged Jack Goldman to secure PARC a dispensation on the grounds that computer scientists were a different breed from the physi­cists and chemists of Webster. For one thing they were comparatively scarce. In 1970 a mere handful of academic institutions offered graduate programs in computer science. The congressional restrictions on ARPA grants foretold that the number would stay small and the inventory of first-class graduates thin. At length Goldman secured a differential for computer science Ph.D.s of 15 to 20 percent over Webster scale. That helped PARC secure the best recruits, but had the predictable side effect of generating resentment among the General Science Lab's physicists and optical scientists, who were excluded.

  Henceforth PARC could offer people with advanced computer science degrees or working experience in the field starting salaries between $30,000 and $35,000—excellent pay for Ph.D.s at the time, although the range remained wide and often depended on a recruit's worldliness and bargaining skills. Some from Stanford University, which was known for its stinginess -with salaried professionals, got low-balled. One recalled accept­ing a full-time PARC salary of $24,000, which was at the low end of pay for principal researchers in the computer lab but a big step up from the $16,000 he had earned at the Stanford Artificial Intelligence Lab.As though to foreshadow the shoals ahead, it was Taylor's pay that caused Pake the biggest headache. Despite his lack of formal credentials Pake had rostered him on the payroll as associate manager of the Com­puter Science Lab and, more formally, as area manager of computer graphics within CSL. The starting salary for an area manager in PARC's order of batde was $44,000, off the scale for a non-Ph.D. anywhere at Xerox. "I had to fight with Goldman and he had to fight with headquar­ters to get that, because Taylor had only a master's degree in psychology and it didn't look right," Pake said. "Of course, I could understand the bureaucratic problems with that myself."

  It was not only Taylor's lack of credentials that made his salary a sore point. Within months of his arrival at PARC his personality started to grate on the other lab managers, who understandably took exception to his attitude that PARC's sole raison d'etre was to pursue computer research and that anything spent on the hard sciences was by defini­tion money down a rathole.

  A man who would never tolerate personal attacks at his ARPA confer­ences, Taylor seemed to tack treacherously close to the ad hominem at PARC management meetings. The other lab managers were particularly appalled by his treatment of Gunning, a warm and charming individual who had spent almost as many years in the electronics industry as Taylor had spent on Earth. "He treated Gunning with the utmost condescen­sion," Jones recalled. "It really created a lot of strife. Bill would say some­thing and Bob would come out with, 'That's stupid!' or, 'I'm just wasting my time in here!' It was very unprofessional, and not at all the general atmosphere everyone was used to."

  Before the year was out a delegation of several middle managers marched into Pake's office to demand Taylor be fired for his behavior. On this occasion, Pake demurred. He was no more charmed by Taylor than they, but he was more acutely aware of the man's uncommon value to the organization. The Taylor who had spun a web of carefully nurtured loyalties among the nation's best young computer researchers seemed an entirely different character from the one who so charmlessly provoked his peers and superiors. The bottom line was that no one could match his ability to lure research talent to PARC; virtually everyone hired thus far into the Computer Science Lab was someone who knew and respected him personally. Pake felt there were many more gifted scientists yet to be snagged.

  He was right about that. Toward the end of 1970 Taylor called in some of his old chits to stage a pair of dazzling heists.

  The first was a raid on the only laboratory on the West Coast—possibly the country—whose wo
rk on interactive computing met his stern stan­dards. The lab belonged to the legendary engineer Douglas C. Engel- bart, an adamantine visionary who held court out of a small tiiink tank called SRI, or the Stanford Research Institute, a couple of miles north of Palo Alto in the community of Menlo Park.

  There Engelbart had established his "Augmentation Research Center." The name derived from his conviction that the computer was not only capable of assisting the human thought process, but reinventing it on a higher plane. The "augmentation of human intellect," as he defined it, meant that the computers ability to store, classify, and retrieve informa­tion would someday alter the very way people thought, wrote, and figured.

  Engelbart s vision refined and expanded a concept memorably set forth by Dr. Vannevar Bush, an MIT engineering dean and wartime science advisor to Franklin D. Roosevelt. In 1945 Bush had turned his attention to the scientific advances produced in the name of war and to how they might serve the peace. The result was a small masterpiece of scientific augury entitled "As We May Think," which appeared in the July 1945 issue of The Atlantic Monthly.

  "As We May Think" remains one of the few genuinely seminal docu­ments of the computer age. Even today it stands out as a work of meticulous scientific and social analysis. The contemporary reader is struck by its pragmatism and farsightedness, expressed without a hint of platitude or utopianism, those common afflictions of writing about the future. Bush was not interested in drawing magical pictures in the air; he was busy scrutinizing the new technologies of the postwar world to see how they might relieve society's pressing burdens.

  His essay dealt chiefly with technology's ability to manage informa­tion. Bush discerned the birth of what would come to be called the "information glut" and projected it forward to a cacophonous posterity. "Publication has been extended far beyond our present ability to make real use of the record," he wrote. "The summation of human experi­ence is being expanded at a prodigious rate, and the means we use for threading through the consequent maze to the momentarily important item is the same as was used in the days of square-rigged ships."

  Himself the inventor of a successful analog computer, Bush under­stood that computer technology might help society draw sense out of the chaos. He sketched out something called the "memex," which he described as "a device in which an individual stores all his books, records, and communications, and which is mechanized so that it may be consulted with exceeding speed and flexibility." The mechanism of consultation would be "associative indexing . . . whereby any item may be caused at will to select immediately and automatically another. This is the essential feature of the memex."

  Doug Engelbart first encountered Bush's memex in a magazine arti­cle he found in an a Red Cross library in Manila, where he was await­ing transport home from his World War II service. He succumbed to the author's vision of a world of interlinked data as though to a sor­cerer's spell. By the time he left Berkeley a few years later with a Ph.D. in engineering, he had decided that his mission in life would be, in effect, to turn the memex into reality.

  In the event, he went far beyond anything Bush himself had imagined. At SRI he propagated from Bush's rough blueprint a full-blown system of interactive hardware and software aimed at managing, manipulating, and communicating text and video images. The achievement was all the more remarkable given that it involved an uphill battle against nearly universal skepticism. More than once Engelbart's thinly financed project narrowly eluded extermination. Gradually, however, he acquired a sizable coterie of young engineers and scientists who felt their lives altered by their first meetings with the charismatic Doug Engelbart and who regarded his vision with an almost religious awe. "He not only made sense," recalled Bill Duvall, one of the early disciples. "It was like someone turning on a light. Love at first sight is perhaps the wrong term to use, but it was as close to that as you can get."

  One other individual entranced by Engelbart's work was Bob Taylor. At NASA in 1963 Taylor had saved Engelbart's lab by scrounging enough money to overcome a budget crisis. After moving on to ARPA he turned the trickle of funding into a flood. By the end of the decade the Augmentation Research Center, fueled by ARPA's half-million- dollar annual grant and occupying one entire wing of SRI's Menlo Park headquarters, reigned as the think tank's dominant research program.

  What it produced was nothing short of astonishing. Obsessed with developing new ways for man and computer to interact, Engelbart linked video terminals to mainframes by cable and communicated with the machines via televised images. To allow the user to move the insertion point, or cursor, from place to place in a block of text instantaneously, he outfitted a hollowed-out block of wood with two small wheels fixed at right angles so it could be rolled smoothly over a flat surface. The wheels communicated their motion to potentiometers whose signals in turn were translated by the computer into the placement of the cursor on the screen. From this crude device would spring an entire culture. "No one is quite sure why it got named a 'mouse,'" Engelbart said years later. "None of us thought that the name would have stayed with it, out in the world." The entire interactive system—mouse, screen, computer, soft­ware, and underlying philosophy—was known by the acronym "NLS," for "oNLine System."

  Until 1968 Engelbart and his aides labored in relative obscurity, their work known only within the insular fraternity of government grant- makers and computer theorists. That year he requested ninety minutes to demonstrate NLS at the Fall Joint Computer Conference of two lead­ing engineering societies, scheduled for San Francisco in December. The result was one of the most famous events in computing history.

  The mouse, making its first public appearance, was the least of it. Engelbart and his sixteen assistants stretched existing electronic tech­nology nearly to the breaking point. He recalled later: "We built special electronics that picked up the control inputs from my mouse, keyset, and keyboard and piped them down to SRI [that is, Stanford Research Institute] over a telephone hookup. We leased two microwave lines up from our laboratory, roughly thirty miles. It took two additional antennas on the roof at SRI, four more on a truck on Skyline Boulevard, and two on the roof of the conference center. It cost money . . . The nice people at ARPA and NASA, who were funding us, effectively had to say, 'Don't tell me!'"

  The effort was worth every penny. The audience was riveted, as Engelbart in his subdued drone described and demonstrated a fully operational system of interactive video conferencing, multimedia dis­plays, and split-screen technology.

  At one point half of a twenty-foot-tall projection screen was occupied by a live video image of Engelbart on stage, the other half by text trans­mitted live from Menlo Park (it was a shopping list including apples, oranges, bean soup, and French bread). Minutes later the screen carried a live video image of a hand rolling the unusual "mouse" around a desk­top while a superimposed computer display showed how the cursor simultaneously and obediently followed its path.

  The piece de resistance was Engelbart s implementation of the memex. The screen showed how a user could select a single word in a text docu­ment and be instantly transported to the relevant portion of a second document—the essence of hypertext, found today, some thirty years later, on every World Wide Web page and countless word-processed docu­ments. At the conclusion of the bravura performance Doug Engelbart, previously a prophet without honor, was rewarded with a standing ovation.

  In 1971 Taylor, whose ARPA funds had helped pay for that demo, was intent on somehow importing Engelbarts interactive vision into PARC. The only question was how to do it without also importing Doug Engel­bart. The problem was that the master’s inspirational dreams were insep­arable from his inflexible and self-righteous disposition. One admirer called him "a prophet of biblical dimensions," a role he fit down to his physical appearance. Tall and craggy, with deep-set eyes and a hawk-like nose, he might have been carved from a slab of antediluvian granite. Soft- spoken but intransigent, his years of battling unbelievers had convinced him that he was fated to remain the solita
ry leader of a devoted cadre.

  By the time Taylor was poised to strike, that cadre was showing signs of serious discontent. As the novelty of his ideas wore off (to be fair, this was a process that could take several years), some disciples started to discern the drawbacks of working for so uncompromising a boss— particularly one whose tendency to oracular pronouncements required a stratum of top assistants to periodically sit the rest of the staff down and explain what Doug had in mind.

  Engelbart's self-defined mission was not to produce a product, or even a prototype; it was an open-ended search for knowledge. Conse­quently, no project in his lab ever seemed to come to an end. Whenever one approached a milestone he would abruptly redefine it, condemning the lab to months or years of further work. The finish line was con­stantly receding, like the oasis in a desert mirage. Said one long-term member of his lab: "We were like rats running in his maze."

  The first to defect was William K. English, a brilliant engineer who had been with Engelbart almost from the start, joining him in 1962 as his hardware ace and all-around major-domo. Wiry and deliberative, Bill English had been the invisible guiding hand behind the 1968 demo. He was ferociously loyal to his boss but bridled at the lab s perpetual lack of closure. Some also believed he was fed up with Engelbarts way of monopolizing credit for the lab s accomplishments.