Dealers of Lightning Read online

  Nothing like this was simple in 1972 because of the cost of memory. Nor was it enough for Rider’s machine to generate only the bland stan­dardized ASCII text of conventional line printers. The RCG had to incorporate a large number of custom typefaces that were to be drawn by hand, converted into digital bits, and stored somewhere in memory until needed, as if on an electronic shelf.

  This meant an exponential increase in the complexity of the task. ASCII characters were all the same size and each fit into the same squared-off shape. The only formatting a conventional document nor­mally required was a command instructing the printer when to move to the next line. By contrast, the custom-designed characters PARC desired to print would be proportionately spaced: some fat, some thin, some reaching above the print line, some dangling below; some roman, some italic, some BOLD.

  Finally, the character generator had to adapt to the Model 7000’s sys­tem of feeding in pages wide-edge-first, which moved paper through the machine at a faster rate. For copiers this posed no problem—one simply aligned the originals along the same axis. For a printer, however, it was a horror. The image coming from the computer would somehow have to be rotated before it could be printed out. Instead of printing a page in prim linear order like a typewriter, SLOT would have to reproduce the characters in vertical slices, somehow keeping its place on twenty or thirty lines of print per page.

  Rider ultimately came to see the proliferation of complications as a blessing in disguise. "It forced you to think about the problem of printing in a much more generalized fashion, so the solution turned out to be much more robust." Despite its name, the research character generator was less about delivering images character-by-character than about trans­mitting digitized images in whatever form the computer dictated. Like so much PARC developed in those first few years, this turned out to be the answer to a multitude of questions no one was yet even asking.

  Starkweather and Rider worked together on coordinating the SLOT and character generator until early 1972, when they were stymied not by a technical obstacle but one entirely man-made. This was the relo­cation of more than twenty of PARC's seventy scientists up the hill to a building newly rented from the Singer Company and known as Build­ing 34 (because its address was 3406 Hillview). The Computer Science Lab, including Rider, got bundled off to the new quarters while every­one else, including Starkweather, temporarily stayed behind on Porter. The move separated the two by a kilometer of real estate—too far to string an overhead line and, with the four-lane Foothill Highway in the way, impossible to link via a ground cable.

  "The administrators said, 'Don't worry. You'll be back together in another year,'" Starkweather recalled. "I said, 'Great, what are we sup­posed to do in the meantime?'"

  But one Sunday afternoon shortly after the move Starkweather got a brainstorm while sitting at home. He immediately jumped in his car, drove to Porter Drive, and mounted a stairwell to the roof. Just as he had thought, he could take line-of-sight aim from where he stood to the rooftop of Building 34. He might not be able to span the distance by cable or wire—but he could do it by laser beam.

  The next day he ordered four telescopes from Edmund's for about $300 apiece. He and Rider replaced the eyepieces of two with low-power lasers and the others with sensitive photo detectors. They bolted one laser scope and one detector on each roof, aiming each at its complement across the way, to create a visible light data link. The circuit worked flaw­lessly in almost any weather, even fog, although minor adjustments were often necessary after a rainstorm, when the weight of accumulated water made the roofs sag slightly.

  "When SLOT was running I'd send a pulse of light up the hill to signal the character generator to send a line of data down to the detector on my roof, which would send it down to this laser and then to the printer,"

  Starkweather recalled. "After all, we were only encoding ones and zeros. It was like sending binary data on a long wire made out of light, instead of copper."

  The only real problem arose from the arrangement's elemental spookiness. One morning after a foggy night Rick Jones was summoned from his office to field a complaint from a peeved Palo Alto police officer. It seemed that a local motorist startled by a ghostly red beam crossing over­head had run herself off Foothill Highway into a ditch the night before. Whatever PARC was up to, it had created a traffic hazard and would have to stop.

  Jones placated the officer and brought the issue to Starkweather, who averted further mishaps by coarsening the focus just a bit. From then on the beam would be too broad to be seen even in the fog, but not so much that it could not be refocused to adequate tolerance at the receptor end.

  "That way we were able to keep the experiment going for a year, until we could move everybody up the hill," Starkweather recalled. "Outside of that and a couple of birds that got hit with a bright red flash, we never had a single problem."

  Starkweather's SLOT and Rider's character generator were two of the four legs of the complete interactive office environment PARC was cre­ating on the fly. In the same period Thacker, McCreight, and Lampson were building the Alto; Alan Kay and his Learning Research Group were designing a graphical user interface aimed at making computers intu­itively simple to use; and Bob Metcalfe and David Boggs were designing a network—the Ethernet—to tie all the other components together. "We had in mind that you ought to be able to create a file on the Alto and ship it via the Ethernet to a print server [that is, a communal computer man­aging everyone's print orders], which would convert it to a raster and print it out," Rider recalled. When it was finally implemented, the whole array would be known by the rather inelegant acronym EARS, which stood for "Ethernet-Alto-RCG-SLOT."

  Of the four, the laser printer was closest to being marketable, repre­senting as it did a fairly straightforward modification of a standard Xerox copier. Yet its road to commercialization would be a long and "gory" one, as Jack Goldman later remarked—presaging other battles to come in the war to bring PARC's inventions to market.

  The first stumble occurred in 1972, even before EARS's other compo­nents were operational. That year the Lawrence Livermore National Laboratory, an institution always primed to promote new technologies, publicly requested bids for five laser printers. Only on the surface was this a public solicitation, for Livermore knew that PARC alone had devel­oped the applicable technology.

  Jack Goldman was eager to fill the bid, figuring that the Livermore contract would guarantee instant celebrity for PARC's first marketable product. Unexpectedly, he was overruled by James O'Neill, a former Ford Motor Company finance man who was in charge of Xerox's engi­neering and manufacturing group.

  Goldman was furious. "I raised a fuss with him," he recalled. "I said, 'Why are you turning it down?' He said, 'I'm turning it down because we'll lose money. The reliability of the Xerox 7000 can't stand the copy volume Livermore will be turning out. We'll be sending so many repairmen out there we'll lose $150,000 over the life of the contract.'"

  O'Neill had it all wrong, Goldman argued. PARC had shown that the machine's reliability improved by more than tenfold when it operated in laser mode, because laser printing circumvented the moving parts most prone to failure. "We had a lot of experience in the reliability of this thing," he said. 'We had turned out millions of copies already in the lab, where everyone was using it."

  Yet the two executives' disagreement was more than a technical misun­derstanding. It reflected a fundamental clash of marketing values. O'Neill saw little point in committing Xerox to selling a machine for which there was no immediate prospect of high-volume production or marketing backup. The company would not sell Livermore a prototype copier; why sell it a prototype laser printer?

  Goldman's rejoinder was that there was a world of difference between introducing a new version of an old copier and launching an entirely new technology; the only way to accomplish the latter was to feed the appetite of "early adopters"—clients willing to take a chance on unfamiliar prod­ucts just to see what they might do. But h
e lost the argument.

  "You have to take certain chances if you're going to introduce a new product," he said later. "O'Neill refused to let us fill that order, and look what he sacrificed. That machine would have had the world by the tail."

  Instead the laser printer spent another two years in product planning limbo, at which point Goldman had to intercede again—this time more successfully—to save it from extinction.

  That happened in 1974 when Xerox's product review committee, on which corporate staff planners were overrepresented and engineers almost non-existent, debated which kind of computer printer Xerox should bring to market. At the eleventh hour Goldman discovered that the committee planned to recommend a Webster-designed machine known as the "Superprinter," which used CRTs, or cathode ray tubes— thousands of times dimmer than a laser—to project an image onto a pho­toreceptor.

  "A bunch of horse's asses who didn't know anything about technology were making the decision," Goldman recollected. The Superprinter, he contended, was hopelessly unequal to the demands of high-speed print­ing. "Here laser printing had already been developed by Starkweather, and the guys back in Rochester were thinking in terms of CRTs, which was absolutely a backward way of doing it."

  This time Goldman did more than argue. Commandeering a com­pany plane, he hustled two key committee members onto it—Don Pendery, the planning vice president, and his boss, a staff vice presi­dent named Bill Souders—for a hastily arranged demo of laser print­ing at PARC.

  "It was Monday night. I said, 'We're going out tonight and coming back tomorrow night in time for Wednesday morning's meeting.' And we made believers out of them. The guys at Palo Alto did a masterful job of presenting it. Everything worked without a hitch. These two guys looked at it and said, 'Hey, this is really the way to go.' And we were able to override the proposal from Webster."

  Still, it was a Xerox-style victory, Pyrrhic at best. Although the com­mittee accepted laser technology, it rejected Goldman's appeal to build laser-adapted Model 7000 copiers, as Starkweather had done. This would have allowed the company to market a laser printer within a year. The panel decided instead to wait until the launch of Xerox's next generation of high-speed copiers, the 9000 series—which was not scheduled for another three years.

  It was a perilous delay. The plan to commercialize the laser printer would be killed and resurrected three times in that period, saved only by the obstinacy of an executive named Jack Lewis, who ran the company's printing division and ignored the orders from higher-ups to deep-six the project. Finally launched in 1977 as the 9700 printer, Gary Stark­weather's laser device fulfilled its inventor's faith by becoming one of Xerox's best-selling products of all time.

  Even so, for the white-light copier engineers of Webster the laser printer never shed the frightening aspect of an alien technology.

  "Years afterwards I went back there," Starkweather said. "I ran into my old boss, the one who had tried to keep me from leaving. His last words to me were, 'Are you still playing around with that laser stuff?'

  "By then the laser printer was a $2 billion-a-year business."

  CHAPTER 10

  Beating the Dealer

  Chris Jeffers took a deep breath before walking into the big corner conference room on Porter Drive. Alan Kay, who had recruited his childhood friend to join PARC as a sort of amanuensis and chief of staff, had guided him through the rigorous interview process and as far as this last hurdle, the delivering of a technical presentation to his future colleagues sitting in a sort of ple­nary session. Waiting for Jeffers inside the room were about twenty sci­entists and engineers, all lounging improbably on beanbag chairs upholstered in a ghastly mustard-yellow fabric. The weekly meeting about to convene had come to be known simply as "Dealer." It was already a PARC institution.

  Bob Taylor liked to tell people that his style of managing CSL com­bined the best features of all the research labs he had ever known. But its structure sprouted largely from a small kernel: the management princi­ples developed at ARPA. Taylors predecessors had bequeathed him the axiom that the best way to manage research was to select the best people in a given field and set them loose. Scientists with the lofty skills ARPA demanded, Ivan Sutherland said, "are people who have ideas you can either back or not, but they are quite difficult to influence. You can maybe convince them that something’s of interest and importance, but you cannot tell them what to do."

  On the other hand, you can find a way for them to tell each other. The uncompromising give-and-take of Taylors ARPA contractor meet­ings lent itself to reproduction at PARC in the form of "Dealer."

  The name derived from the book Beat the Dealer, by Edward O. Thorp, an MIT math professor who had developed a sure-fire system for winning at blackjack—"beating the dealer"—by counting the high- and low-value cards dealt out in hands. (This truly effective system would make the unassuming Ed Thorp the godfather of professional blackjack card-counting.)

  Taylor was not much of a blackjack buff. What interested him about Beat the Dealer was its compelling metaphor of a doughty individual fielding the challenge of a group of trained and determined adver­saries. In casino blackjack the dealer plays against everyone at the table. In Taylor's variant a single researcher would propose an idea or project, then stand alone to defend it against dissection by his peers.

  Dealer was soon institutionalized as the beating heart of CSL's profes­sional organism, a time when the entire lab would gather in a room fur­nished with the beanbag chairs drat Peter Deutsch and his wife, Barbara, had discovered at a friend's shop in Berkeley. The meetings, which were usually on a Tuesday (although the designated day changed from time to time), were scheduled more or less at lunchtime and generally lasted an hour. Attendance was mandatory for all of Taylor's subordinates, the only lab rule he rigidly enforced, and the other labs were welcome to attend, at least at first. Later, as PARC expanded and the crowd at Dealer threat­ened to become unmanageable, non-CSL personnel became welcome only upon invitation or special dispensation. (Kay, though an SSL member, owned a permanent pass.)

  Taylor would open each session with ten to fifteen minutes of house­keeping items before yielding the floor to that week's designated dealer. At that point the game transmuted into something more like poker. It was the dealer's prerogative to set not only the topic of discussion, but the rules of debate.

  "I wanted to have conditions where someone could get up to the table and set rules as czar," Taylor recalled. "You could say, no interruptions; or interrupt whenever you want. Or I'll only debate x, y, or z; or only right-handers can argue." The discussion topics were similarly unconstrained. Certainly they tended toward issues of importance to the lab, but that category was broadly defined. Bob Flegal, a CSL graphics expert, once demonstrated for his colleagues how to take a bicycle apart and lubricate the parts, and Ed Fiala was famous for a memorable presentation on how programming algorithms resemble kitchen recipes.

  Outsiders arriving with influential backing got extra latitude, as hap­pened when Kay surprised Jeffers with the news that he would be making a speech at the next Dealer. Jeffers, the farthest thing from a trained computer scientist, had spent the previous few years first as a Peace Corps volunteer in Nepal and then as an official in the agency's Washington office. He told Kay there was no way he could cook up an appropriate presentation to the digital elect of PARC.

  Kay advised, "Just talk about something you know."

  "So I gave a speech about the sociolinguisties of Nepalese language and culture, and we had a good time with that," Jeffers recalled with relief. "Actually, I felt quite at home."

  This was also part of Taylor's scheme. Once accepted into the lab, you were immune to the petty harassments common to university depart­ments. "You were part of the extended family," related John Shoch, a member of Kay's lab. "No one ever asked, "Who the hell are you and what are you doing here?'" The alternative, Taylor believed, was for one-upmanship to hobble the unfettered exchange of ideas. "If someone tried to p
ush their personality rather than their argument, they'd find that it wouldn't work."

  But the argument had best be carefully thought out. Anyone trying to slip an unsound concept past this group was sure to be stopped short by an explosive "Bullshit!" from Thacker or "Nonsense!" from the beetle- browed ARPANET veteran Severo Ornstein. Then would follow a cas­cade of angry denunciations: "You don't know what you're talking about!" "That'll never work!" "That's the stupidest idea I've ever heard!" Lamp- son might add a warp-speed chapter-and-verse deconstruction of the speaker's sorry reasoning. If the chastened dealer was lucky (and still standing), the discussion might finally turn to how he might improve on his poor first effort.

  The criticisms could be particularly ruthless when Dealer turned to the qualifications of a job candidate. Scientific prodigies who had spent half their lives defending abstruse research before hostile faculty committees were easily unnerved by this small group slouched in their beanbags, rudely firing off comments of annihilating incisiveness. Newcomers almost always came away from Dealer profoundly unsettled.