Dealers of Lightning Read online

  Yet Elkind must have sensed something vaguely eccentric about the position he was being offered. "Why not just let Bob manage the lab?" he asked.

  "I need someone with more scientific training than just in the com­puter field," Pake replied smoothly. "The computer side is only twenty people now, but it will grow. Once the Computer Science Lab gets to thirty or forty scientists we'll need managers who can meld it with the physical and information sciences." Elkind, he implied, was just that sort of manager. Melding the labs? Bob Taylor's agenda seemed to lay entirely in obliterating the other labs.

  "George made it sound very exciting," Elkind recalled. "Here was a lab that was going to be supported well and the motivation was solid. It seemed like a very, very talented group of people had already come there."

  He came west to indulge CSL in its ritual of subjecting prospective new members to serial interviews with the entire staff. Taylor had orig­inally established this system of vetting recruits to the team, along with the rule that a candidate was required to win approval by a near- unanimous vote.

  "The system meant Joe Blow would have a huge advantage coming in," he explained, "because a whole bunch of people would have committed themselves to his success." Subjecting their future boss to the same all- day process was a bit on the bizarre side, but since they were already in place and Elkind was not, they went through with it. By all accounts Elkind acquitted himself well. "He certainly had good paper credentials from BBN," recalled Jim Mitchell. "But I didn't know him so I went on the basis of the interview, and he gave good interview."

  Pake was cheered, if a bit surprised, to see that Elkind and Taylor "seemed to interact pretty well. I would suspect that Taylor, who never had a low opinion of himself, felt that he really didn't need this other guy. But there was no problem about that, not at first."

  In any event with construction of MAXC well under way, CSLs "vec­tor" needed no fine-tuning. If Elkind cared, or even noticed, drat most of his new staff bypassed his office and did their brainstorming next door with Bob Taylor, he did not show it. He did, however, recruit a cadre of his own, raiding BBN for Daniel Bobrow and Warren Teitelman, two tal­ented Lisp programmers he encouraged to continue their work on artifi­cial intelligence. Bobrow, who would remain deeply loyal to Elkind throughout the strife to come, almost immediately detected something unsettling about CSLs ambiance. One day he confronted Taylor about it.

  "I asked him why he recruited somebody to be his own boss," he recalled. Taylor repeated his mantra about needing someone with better credentials to head CSL so he could be more free to exert his influence over both computer labs, SSL included. Bobrow was not completely sat­isfied. "I thought he felt that, just as at ARPA he was the power behind a lot of thrones, he could be the power behind a lot of thrones here, too," he observed. The unanswered question, however, was what might hap­pen if he ever got a hankering to sit on the throne himself.

  If PARC's computer engineers thought the design and construction of MAXC would place the PDP-10 controversy behind them, they were mistaken. At Xerox headquarters the contretemps earned PARC a reputation for insolence it would never entirely shake. Reinforced by a thousand further affronts over time, this would evolve into a major handicap in its relations with headquarters in Stamford. At first, how­ever, it simply gave Xerox a pretext to pay closer attention to what the research center was up to. The agent of this unwritten policy was a man named Don Pendery.

  A Xerox corporate planning executive, Pendery chaired a headquar­ters task force—one of innumerable such bodies—devoted to monitor­ing technological changes that might affect the company's business plan. In pursuit of the answers, he made frequent contact with the people at PARC, who tended to regard his concerns as short-sighted and parochial. Alan Kay, the center's self-defined futurist-in-residence, took particular umbrage at Pendery's approach, which treated the future largely as a Pandora's Box of threats to the bottom line. Kay and his colleagues preferred to consider it more as a harbinger of limitless opportunity. To scan the horizon only for hints of Xerox's future, they thought, forced the company to ask the wrong questions and ensured that whatever answers came would be misinterpreted or ignored.

  Pendery "really didn't understand what we were talking about," Kay recalled. Instead he was "interested in 'trends' and 'what was the future going to be like' and how Xerox could 'defend against it.'"

  In the course of one frustrating encounter Kay blurted out the line des­tined to become his (and PARC's) unofficial credo. "Look," he said, "the best way to predict the future is to invent it!"

  But PARC had come face to face with a force of nature, the corpo­rate instinct for self-preservation. While Kay urged upon Xerox the virtues of patience and trust in scientific serendipity, Pendery pressed for a definition of its vision that could be reduced to paper and pre­sented in a boardroom. Finally he got it. In mid-1971 George Pake sent up to headquarters a half-inch-thick folder containing seven docu­ments, each written by an individual PARC scientist—scarcely sixty pages altogether. Someone had cheekily labeled it "PARC Papers for Pendery and Planning Purposes." In lab shorthand they were hence­forth known as the "Pendery Papers."

  Not since Vannevar Bush had forecast how we might think in his essay for The Atlantic had such a comprehensive vision of technology and the future been set down in writing. The Pendery Papers were at once a sur­vey of the most promising technologies on the horizon and a road map for PARC's ten-year exploratory journey. Some of the forecasts overshot their marks. Kay, for instance, anticipated (perhaps wishfully) portable flat-screen displays at nominal cost by 1980. Jim Mitchell, writing on future office systems, envisioned error-free and infinitely customizable software, transmitted from vender to buyer over network connections, running flawlessly on a full spectrum of incompatible machines (as of this writing still a hazy dream). But on the whole the package stands with Bush's as a remarkable feat of scientific prognostication. Mitchells office of the future was one in which uncompleted memos, letters, and reports would exist solely on computer, to be printed out only when a final hard copy was needed. ("Much of the current 'paper push­ing' in today's offices will be replaced by people spending a large portion of their time using a computer via some personal terminal.") He forecast the propagation of electronic mail and divined its unique ability to allow people to "communicate and manipulate information simultaneously, without the necessity of physical proximity." The floppy disk would replace the file cabinet as the principal repository of documents and information.

  Dick Shoup, reporting on integrated circuit technology, anticipated the development of "smart" appliances such as toasters and alarm clocks equipped with simple but powerful chips. John Urbach's paper on "archival memory" described digital photo-optical media resembling today's CD- ROMs and compact audio discs. "There seems little reason to store sound in analog form," he wrote, observing that acoustical information is easily reduced to bursts of digital bits—thus consigning the LP record to the dustbin more than fifteen years before it actually met such a fate.

  To be fair, many of the startling innovations posited by the Pendery authors were ringers: They were only modest extrapolations from tech­nologies well-known throughout the research community, if not among the broad public. Mitchell's description of tomorrow's text-editing and office systems drew heavily from Doug Engelbart's 1968 demonstra­tion. Shoup's survey of integrated circuitry scarcely ventured much beyond devices that were already on the market or known to be under active development. Still, futurists have no obligation to venture solely into the realm of magic and crystal balls; sometimes a clear vision of what lies around the next bend will do. As things stood, the Pendery Papers were important for PARC and Xerox in three ways.

  The first was that they implicitly embraced the immense but still widely unappreciated power of Moore's Law. The term appeared nowhere in the Pendery Papers, but its significance permeated every page. The implication of Moore's article had been that technologies impractical in 1965 wou
ld be commonplace within a decade or two. In the Pendery Papers PARC informed Xerox that the devices on the drawing board today would be marketable in ten years, so it was time to get ready.

  "This was their version of the old hunters saying, 'Never aim at the ass end of a duck,'" remarked George White, Jack Goldman's assistant, who served on Pendery's task force. "PARC was telling us that if you want to invest in research at Palo Alto you've got to get way ahead. Otherwise, by the time the ripening and maturing process from your research comes through events will have overtaken you."

  PARC further understood that Moore's Law would pack its greatest wattage in the visual interaction between computer and man. Virtually every paper touched on this topic and some dwelled on it at length (Kay's was devoted entirely to display technology). It was as though the lab had finally absorbed the lesson Bob Taylor had been pressing on it for more than a year: The computer is a communications device in which the dis­play is the whole point.

  The third benefit of the Pendery Papers inured to PARC alone. "It was a matter of setting the primary focus for the lab," recalled Peter Deutsch. "Even though in our guts nobody believed that you would be able to put a portable computer on every desk ten years from now, that was what was said by the industry trend and the curves of various things. You'd be able to put something equivalent to MAXC on everybody's desk in ten years."

  One might argue that the Pendery Papers were another example, like Strassmann's veto of the PDP-10, of how a hectoring from headquarters proved itself to be a blessing in disguise. They named their file of white papers after their tormentor from the home office, but they wrote it for themselves. With dazzling audacity Mitchell, Kay, Urbach, and the oth­ers had fixed on their destination. Now it was up to all of them, working together, to blaze the path that would take them there.

  PART II

  Inventors

  CHAPTER 9

  The Refugee

  If anyone symbolized the gulf separating the inventors of the future in Palo Alto from the Xerox development drones back East, that person was Gary Starkweather.

  Starkweather was highly trained in an arcane subspecialty of physics, but he did not look like anyone's idea of a master physicist. With his stocky frame and friendly, guileless features, he more resembled your neighborhood phone lineman. But to his colleagues at PARC he was a special catch. He was the scientist outcast, the man who got branded a renegade by his bosses at Webster simply for proving that the novel technology of lasers could be used to "paint" an image onto a xerographic drum with greater speed and precision than ordinary white light.

  Instead of garnering praise and encouragement he was ordered to abandon his research and threatened with the loss of his lab assistants. His bosses hinted that his future at Xerox would be bleak if he failed to redirect his energies back to the pressing issues oflenses and white light. 'We had almost reached the point of maximum disconnect," he recalled, when it was finally recognized that the only place for him was that mad­house out in Palo Alto.

  And there at PARC he invented the laser printer, the success of which contradicts the canard that Xerox never earned a dime from the Palo Alto Research Center. It is one of the ironies of the story that despite Jack Goldman's tireless efforts to keep PARC insulated from Webster's copier-duplicator mentality, the most profitable product PARC ever produced sprang from the mind of a Webster man.

  Not that they ever thought of him that way at PARC. "Gary Stark­weather had been thrown out of Webster," Alan Kay remarked with manifest approval. "We considered him one of us."

  For all his considerable skills at manipulating light, Gary Stark­weather's career in optics began more or less on a whim. In 1960, hav­ing just received his bachelor's degree in physics from Michigan State University, he faced a limited spectrum of career options. "The choices were I could go into nuclear power, which was a hot thing in 1960, or I could go into optics. And I looked at nuclear and said, I don't think so. I wasn't sure how people would live with the problems, because when nuclear fails, it fails big. So I went into optics." It was a lucky choice. Just a year or so into his master's studies at the University of Rochester, the entire field blew wide open.

  At Hughes Research Laboratory in Malibu, Theodore Maiman had coiled an electronic tube around a cylinder of pink ruby polished at either end to a mirrored sheen. He touched off a flash of electrons within the coil, exciting the ruby into firing an instantaneous burst of single- wavelength red light from one end. The science of optics was never the same.

  Before the laser's appearance, light was a crude implement. Optical scientists could knock it about with lenses and mirrors and sort it into its constituent wavelengths with prisms. But these processes bore all the delicacy of surgery performed with a jackhammer. By contrast, the laser cut like a scalpel.

  White light generated thermally—by bulbs and electric arcs— comprises all the colors of the spectrum, oscillating at different wave­lengths and consisting of photons generated out of phase with one another. Under such conditions light inevitably scatters and diffuses over distance, like ocean waves spending themselves on the beach. Maiman's ruby device, however, emitted a beam immune to the scattering effect. It had spatial coherence (all the light in the beam was the same wavelength) and temporal coherence, meaning that its photons were in phase. The laser could be "tuned," like a radio antenna, to be so bright and fine that a beam shined from the Earth could visibly illuminate a spot on the moon.

  Optical scientists welcomed the new technology as a tool for making the theoretical concrete. Hypotheses of the existence of certain photo­electric effects and other phenomena could now be tested in the lab. At the University of Rochester Gary Starkweather abandoned his original masters topic in classical optics, refocused his attention on lasers, and received his degree for a thesis exploring holography, the laser-aided cre­ation of three-dimensional images. With great anticipation he brought his knowledge back to Xerox's Webster lab, where he had worked his way through school, only to be instructed to stop talking like a madman.

  For a company whose vast corporate fortune depended on the manip­ulation of fight, Xerox remained resolutely behind the curve in exploiting Ted Maiman's discovery. Everywhere Starkweather turned at Webster he saw projects coming to naught because they employed light sources too feeble. Whenever he pointed out that the laser packed 10,000 times the brightness of a conventional light source he encountered sneers, espe­cially when he suggested that the new devices might play a role in xero­graphic imaging. Lasers were difficult to handle and burned out faster than a rick of dry timber, his colleagues responded. Brisding with elec­trodes and emitting bursts of blinding light, they seemed about as safe to put into an office machine as nuclear warheads. And they were expensive—$2,500 to $25,000 for a single unit.

  For the next few years Starkweather had no choice but to experiment on his own. His instincts told him that a beam so precise could be modu­lated—that is, altered in intensity—to carry information, just like radio waves or the pulses on a phone line. Suppose one could educate a light beam to reliably transmit digital bits: These could then be translated into marks on a blank sheet, a feat that would allow one to consign to paper the thoughts and images created inside a machine.

  Enlisting the help of a couple of lab assistants, he built a clumsy proto­type, hitching a laser apparatus to an old seven-page-a-minute copier no one used anymore. Whenever he could steal an hour or two early in the morning or late at night he would run some equally clumsy tests by bom­barding an unused xerographic drum with laser beams. Eventually he learned how to scan an original image and turn out a duplicate. True, his first samples were crude and pale, not at all ready for prime time. Still, they were scarcely any worse than the faded, scrawled "10-22-38 Asto­ria" Chester Carlson had reproduced on a coarse apparatus in his kitchen. From Carlson’s crude and pale sample, Starkweather kept reminding himself, an awesome new industry had sprung. Who was to say that his might not do the same?

  Nevertheless
, Starkweather got scarcely more respect than Carlson had at the start of his own researches. "The theoreticians gave me every excuse," he recalled. "All hogwash. They told me the beam would be moving so rapidly the photoreceptor would never see it. They talked about 'photoconductor fatigue' and asked, How will you modulate? They thought there was no practical value in it. "We got copiers we need to ship, you need to work on the lenses for that . . . Painting laser beams, these things are expensive, they never last very long and they look like a ham radio set. It's a completely useless application. If you paint at 200 dots per inch that's a million bits of data, where will you ever get a million bits of information?' In 1968 that was probably a valid question. But it wasn't a valid question if you looked at where the technology might go."

  Over months and years of trying, fueled by the inner conviction that drives natural inventors, he fashioned experiments that answered every objection. He could modulate the beam by varying the power input and scan it by the clever application of a set of mirrors. He was proudest of disproving the old bugaboo about "photoconductor fatigue." This referred to a hypothetical property of the selenium coating of the copier's xerographic drum, the electrostatic charge of which must be neutralized by light in order for the duplicating process to work.