Dealers of Lightning

by Michael Hiltzik

  But even the most experienced lecturers could get themselves man­handled. The featured speaker at one memorable Dealer was Alan Newell, a distinguished professor at Carnegie-Mellon University, or CMU, who was not only friend but mentor to a good half-dozen of the engineers in the room. Newell literally had written the textbook on com­puter architectures. On this occasion the agenda called for him to solve a tricky programming problem in front of a video camera so his students back in Pittsburgh would be able to study his thought process step by step, as though debugging lines of code. Within the first few steps, how­ever, he unwittingly committed a rudimentary mistake. What the stu­dents got on tape instead was a roomful of smart-assed engineers pepper­ing the increasingly flustered Newell with bluntly phrased suggestions about how to recover from his blunder.

  Only once could anyone recall the group s being specifically ordered to go easy on a guest. In early 1973 Pake decided to hire Harold Hall, an avuncular research executive who had worked at ARPA and Ford, to be the long-awaited replacement for Bill Gunning as SSL chief. Hall was not exempt from the ritual of the mass interview, but Pake did not want him roughhoused, either.

  "Taylor obviously had been told that he had to make sure Harold got a nice respectful reception," recalled the CSL engineer Chuck Geschke. "So rather than have him come in right at the beginning of Dealer, Bob first gave us a little lecture on appropriate modes of behavior and how most executives in the Xerox Corporation wouldn't be accustomed to what normally went on in Dealer." Jim Morris, an acid-tongued trans­plant from CMU, was sitting in the back. "Suddenly," Geschke recalled, "Morris said, 'Wait a minute! I get it! You're trying to tell us that you're just about to send a piece of china into the bull shop!'"

  But such special handling was rare. The pitiless judgments dis­pensed at Dealer derived from the ethos of the engineer, who is taught that an answer can be right or wrong, "one" or "zero," but not anything in between. It was felt that if you were wrong you were done no favor in being told you were right, or half-right, or had made a decent try. "There was nothing personal about it," said Ornstein. "We didn't want to be coddled or have our time wasted."

  That is not to say that the system was entirely objective. One who thought the lab occasionally used the brutish spirit of Dealer to enforce its own prejudgments was Bob Metcalfe, who arrived at CSL in 1972 with the reassuring credentials of a Harvard and MIT educa­tion. Metcalfe was acerbic and free-speaking, a man who never met an ego he couldn't pierce. At Dealer his radar often detected the unmis­takable "ping" of people pulling rank.

  "I'm being cynical now, but if you were from Berkeley or MIT or, especially, CMU, you'd give your talk, you'd get some questions, you'd get congratulated, and you'd get a job offer," he said. "But if you were some poor schmuck from the University of Arizona, they'd grill you and it was all over. In other words, if the department head at CMU said you were cool, that was good enough for them."

  Others did not overlook the converse of Taylor's effort to promote a group sensibility at CSL. If there were no walls within the lab, there were certainly barriers erected against the outside. "It was almost a cult-like thing," remembered Lynn Conway, an SSL engineer whose background included work on an IBM supercomputer. "I'm not easily attracted to cults and it always made me a little uncomfortable. Taylor's a very powerful personality. Here he was in the background with these gunslingers out front and the groupies in back."

  Taylor's chief gunslinger was Butler Lampson. His combination of a razor-sharp intellect with peerless debating skills raised the bar for new ideas to an intimidating height. It was not impossible to win an argument with Lampson, but it was not at all rare for him to win one even when he was wrong. Even as practiced a navigator of Ideaspace as Alan Kay could be backed to the wall when one of his flights of fancy came up against Lampson's rigorous command of pragmatic engineer­ing. Routed in the battles, Kay sometimes had to retreat and regroup for another run at the fence. "I can't ever remember winning an argu­ment with Butler on the same day," he said later. "I could win quite a few on the second day. His mind worked about twice as fast as anyone else's."

  Lampson was fiercely intellectual, an inveterate kibitzer whose finely realized insights and designs, often recorded on the run on scraps of yel­low paper, became indispensable ingredients of more PARC inventions than anyone has bothered to count. He could also be ferociously tem­peramental, a fearsome screamer and tantrum-thrower when thwarted or contradicted. Once Warren Teitelman managed to goad Lampson into firing a glass ashtray at him. "Butler tended to intimidate people," recalled the outspoken Teitelman. "He made it very difficult for those who didn't think quite as fast as he did or weren't quite as smart."

  On this occasion "Butler was doing one of these 'That's ridiculous!' things, and I just replied, 'Why? Because you say it's ridiculous?' and he heaved the ashtray at me," Teitelman recalled. The ashtray shat­tered harmlessly on the wall behind him, but Teitelman understood the lesson. "He was the 400-pound gorilla in that lab. You had to be real careful." Teitelman's friends suggested he attend future Dealers wearing a hard hat.

  Outside Dealer with its deliberate intellectual gunplay, PARC in this period was a model of casual collegiality. The place retained the ambiance of a college campus, which was unsurprising. Most of the staff, after all, were fresh out of grad school (some were still working toward their advanced degrees while working full-time at PARC). Unmarried or with young families, their social spheres would not extend much beyond their laboratory colleagues until much later, when those families began to grow and exercised their own gravitational pull. For now, driven by the thrill of pursuing a common vision, they would work together all day and late into the night.

  To let off steam there were family picnics and a Softball team Rick Jones organized to play in a Palo Alto community league. In the spa­cious open yard behind Building 34, to which the computer and sys­tems science labs relocated in early 1972, was strung a volleyball net for daily lunch time matches.

  For the extended family of the Computer Science Lab, Bob Taylor served as a sort of social director. On weekends there might be touch football (quarterback: Bob Taylor) or marathon sessions of "Diplomacy," a board game whose framework of negotiation, alliance, and betrayal fed the host's appetite for intrigue, at his Palo Alto house. "That was great fun, when you had nothing to do for a whole eight or ten hours on a Sat­urday or Sunday," one participant recalled.

  This was the sunny side of Taylor's personality. When he was playing the role of paterfamilias, as opposed to sneering at the physicists or dis­puting a football ref s call with an opponent's shirt grasped in his fist, one could appreciate the 95 per cent of the time he could be "an absolutely charming person," as Jones recalled, without thinking of the other 5 per­ cent when he was a rude and arrogant beast. Even his beleaguered supe­riors could laugh at his foibles and persnickety habits, as they did one Halloween when half of CSL came dressed as Bob Taylor, in nearly iden­tical plaid slacks, blue blazers, and white turtleneck sweaters, then sat together at a table in the cafeteria with pipes in one hand and Dr Peppers in the other.

  "There isn't an organization newly begun where you don't find those honeymoon years where there's a special bond among people," reflected Jeffers, who recognized the phenomenon from the Peace Corps. "It was true there, it was true in PARC. It's true in anything that's new. It's a great period. Everyone should be a part of something at the beginning."

  This atmosphere of professional and personal fellowship was a power­ful factor behind some of the center's earliest projects, including MAXC. They called the process of informal collaboration by the name "Tom Sawyering." Like Tom with his paintbrush and whitewash, someone would set forth his idea or project—whether it was in a formal meeting or a hallway bull session was unimportant—to mobilize a few intrigued colleagues in an attempt to make it happen. If you saw a glimmer of how to implement a new operation in microcode, you would gather a few expert coders in a room and have at the problem until ever
y whiteboard in the place was filled with boxes and arrows and symbols as arcane as Nordic runes. If you had a big project with a lot of soldering to be done, everyone who knew how to wield a soldering gun strapped on his holster.

  If an idea worked, the team stuck together for the next three or six months to complete the job; if not, everyone simply dispersed like free electrons in search of a new creative valence. Thacker viewed this sys­tem as "a continuous form of peer review. Projects that were exciting and challenging received something much more important than finan­cial and administrative support. They received help and participation ... As a result, quality work flourished, less interesting work tended to wither."

  In this spirit Systems Science Lab engineers wrote code for Computer Science Lab hardware, CSL designers helped SSL build prototypes, and the General Science Lab's physicists chipped in with valuable insights into material properties and electrical behavior (as when Dave Biegelsen told Starkweather how to use sound waves to modulate a light beam and got his offhand suggestion incorporated into the world's first laser printer).

  At one point Tom Sawyering even begot an audacious extracurricular project. This was the so-called "Bose Conspiracy," which was hatched at a poker game at Rick Jones's house. Jones, Kay, Thacker, Dick Shoup, Chuck Geschke, and a couple of others had fallen into a discussion of the merits of stereo speakers. Kay was a particular fan of the state-of-the-art Bose 901s, which came with their own electronic equalizer and cost $1,100 the set (in the pre-oil shock dollars of the early 1970s). He was also the only one in the group who owned a pair, having acquired them on his PARC budget as part of a real-time music synthesizer his group was developing.

  "You know," someone said as cards riffled in the background, "there's no reason why we couldn't make the electronics work just as well. And for a lot less money, too."

  Appropriating a basement room in Building 34, the group took apart Kay's speakers and painstakingly analyzed the design. They bought cone speakers from the same Kentucky factory that supplied them to Bose, and on a shrieking diamond-toothed radial saw in Jones's garage they cut and shaped the sound baffles out of high-density particle board. (The marathon session left Kay covered with an inch-thick coating of sawdust and Jones with a lifelong case of tinnitus.) Then they apportioned the assembly tasks—one conspirator handled the soldering, another installed the speaker cones, and so on—the same way they had distributed the tasks on MAXC, which happened to be running contentedly in its own air-conditioned room a few doors away. All told, they manufactured more than forty pairs at $125 each. The buyers among their PARC colleagues could customize the units with their choice of grille cloth but were oth­erwise challenged to tell the knockoffs apart from the real thing. No one could.

  "It was so typical of PARC," Kay recalled. "If you didn't know how something was done, you just rolled your own."

  The realization that something extraordinary was germinating on a Palo Alto hillside soon started permeating the world of computer sci­ence, thanks in part to the researchers' eagerness to give demos to friends visiting from Stanford, Berkeley, or Carnegie-Mellon. The names on the employee roster added further luster. Gathering Lampson, Kay, and Deutsch under one roof would have been enough on its own to make PARC a byword; but the center employed a dozen others with reputations nearly as luminous. Kay was fond of proclaiming that of the top hundred computer scientists in the country, fifty-six worked at PARC.

  Or sometimes he was quoted saying fifty-eight, or seventy-eight. Kay's formulation has appeared in a hundred different versions, none of which is correct in a mathematical sense (PARC never employed as many as seventy-eight computer scientists). But all are accurate metaphorically. PARC had become the premier draw for the country's best computer sci­entists, like Disneyland for seven-year-olds. Under the circumstances it was easy to imagine that almost every talented young scientist or engi­neer in the land was already inside.

  "People were accusing us of monopolizing the field," recalled Jack Goldman. One day at a formal luncheon he was cornered by Jerome Wiesner, the president of MIT. "Wiesner accused me of destroying the ability of universities to teach computing because we were grabbing all the good people."

  Delighted as he was by the complaint, Goldman recognized that the key to PARC's success was not the head count of researchers but their exceptional gifts. He found it hard to keep away from his pampered child. Arriving in Palo Alto in the evening on a company plane, sometimes with his wife along ("My only inhibition to her coming along was it stifled my ability to play poker with the guys"), he would drive directly to the lab to drink in the atmosphere.

  "The lights would all be lit and dozens of people around, even it if was nine or ten at night," he recalled. "Often they were playing computer games. Now, just remember, in those days computer games were not what they are today. This was a new thing. These guys were literally inventing computer games and learning how to use the machine."

  Yet there was a downside to the cheery insularity and game-playing that Goldman so enjoyed witnessing at PARC. For one thing, the cen­ter's attitude problem was growing worse. Xerox headquarters discov­ered this to its dismay the day that attitude got laid out for public view in the pages of a rather unsavory magazine.

  CHAPTER 11

  Spacewar

  One day early in December 1972, Rick Jones and Gloria Warner drove to the San Francisco airport to meet George Pake's plane from New York. Normally they would not have made the effort. The established routine whenever Pake returned from a visit to Xerox headquarters was for Warner to send a car for him. This time she canceled the arrangement. The moment Pake saw his two assistants waiting at the gate, he got a bad feeling.

  "What's the matter?" he asked.

  "George, you better have a look at this," Jones said. He handed over a tabloid-sized biweekly magazine he had bought that morning at a newsstand across from the Stanford campus. Pake's glance took in the cover and its unfamiliar banner: Rolling Stone.

  "What is this?" Pake asked.

  "Start on page fifty," Jones replied.

  Pake opened the magazine to a feature article entitled "Spacewar: Fanatic Life and Symbolic Death among the Computer Bums." Its language was loose and profane, its attitude toward computer science individualistic and anti-corporate, and among its leading characters were the not particularly presentable scientists of Xerox Palo Alto Research Center, shown lounging about in their sandals and T-shirts. The date on the cover was December the seventh. If Pake happened to notice it was Pearl Harbor Day he would have thought it grimly appropriate.

  "As we were driving back from the airport," Jones remembered, "all I could hear was George sitting in the back seat, leafing through the article and going, 'Oh, no . . . Oh, no . . . Oh, no!'"

  The piece that was to cause Xerox and PARC so much distress over the following few weeks had in a sense been underwritten by Xerox money. Rolling Stone was then five years old. Its founder, a Berkeley dropout named Jann Wenner who had started the magazine on a shoe­string, had recently turned up backing from a decidedly mainstream source: Max Palevsky, who had left the Xerox board that May. Always in search of entree to the snazzier milieus of countercultural life, Palevsky had placed some of his gains from the sale of SDS at Wenner's disposal and taken for himself the title of Rolling Stone's chairman of the board.

  By this time, Rolling Stone had matured well beyond its origins as a fresh voice in rock journalism and had turned into a purveyor of off­beat but incisive reporting on a wide range of issues, including presi­dential politics and economic policy. But its audience was still essen­tially a college age crowd, as tuned in to the music of Hendrix, Joplin, and the Grateful Dead as to the writing of Hunter S. Thompson.

  Rick Jones had never heard of it before that morning, when Gloria Warner knocked on his door to report that a friend had just called her from San Francisco to say PARC had been written up.

  "What the hell is Rolling Stone?" he asked.

  "Its some druggie magazine,"
she reported.

  Jones swallowed hard. "We'd better get a look at it."

  Together they drove to an off-campus newsstand where they found the magazine prominently displayed. Before they had read to the end of "Spacewar" they knew they had a major crisis on their hands.

  With Bob Taylor's apparent permission, but to the complete igno­rance of anyone else in PARC management, the writer Stewart Brand had apparently been ranging freely through the Computer Science Lab for weeks. Brand was a technology fancier whose recent sale of the Whole Earth Catalog, his popular offbeat guidebook, had left him with the money and time to conduct a personal grand tour of the Bay Area's leading computer research facilities. (A few years later he would resurface as a founder of The Well, a pioneering on-line computer ser­vice.) At the outset, he said later, some old friends at Doug Engelbart's lab put him in touch with Bill English at PARC. But it was Taylor, he recalled, who actually arranged for him to walk into the lab past the lone receptionist who counted, for the moment, as PARC's entire secu­rity force.

  "Spacewar" was Brand's travel report. From its dramatic opening scene, an imaginary battle among players of the eponymous interactive spaceship-and-torpedo computer game invented at MIT in 1962, the article captured the adolescent ferment at the heart of the computer cul­ture. Echoing the phantasmagoric tone of hacker favorite E. E. "Doc" Smiths cosmic swashbucklers ("Beams, rods, and lances of energy flamed and flared . . ."), "Spacewar" painted its subjects as dashing young figures engaged in dynamic battle with a sinister state.