Dealers of Lightning

by Michael Hiltzik

  "Taylor sacrificed a lot of his career at Xerox so we wouldn't see a lot of the bullshit," Alan Kay remarked later. This sentiment was widely shared. It accounted for much of their loyalty to him (even among those with whom he had clashed), and even more for their fear that things would be immeasurably different once he was gone.

  His resignation marked the passing of an era at PARC that some peo­ple believed, perhaps correctly, had run out its string anyway. To Spencer and Pake, Taylor was not merely the strident exponent of CSL privilege; he was the defender of a stultified regime.

  "I knew there was a high risk we would lose people," Spencer recalled. "But by 1980 the research center had really come dead in the water. So you'd had this wonderful burst of imaginative things, a great burst of energy, and by that time it was dead. I had a feeling we were going backwards. Absolutely, the place needed a change.

  "If I had had my choice I would not have chosen to lose all of those computer scientists. But in retrospect, it may have been a good thing."

  Taylor meanwhile found a sponsor for a new lab. It was not Hewlett- Packard, as Pake and Spencer suspected, but Digital Equipment Corp., the maker of the PDP-10 minicomputer that had caused CSL's very first flap with Xerox so many long years before. "DEC called to ask if I would consult," Taylor said. "Then three of their real estate people showed up and asked: Where should we build the lab?"

  Selecting a site near downtown Palo Alto, on the far side of the Stan­ford campus from PARC, Taylor ultimately attracted fifteen CSL staff, among them Thacker and Lampson, to the DEC Systems Research Cen­ter, Robert W. Taylor, director. The emigration of so many top scientists to one place finally got Kearns's attention. He and his right-hand man,

  Bill Glavin, flew to DEC headquarters outside Boston to implore Ken Olson, its founder and CEO, to halt the raid on PARC. They even deliv­ered a rather fatuous warning that Xerox was a very large DEC customer ... for the moment.

  Nothing ever came of the threat. But as Kearns later recounted, a few years later Olson pulled him aside at a corporate chief executive s confer­ence to complain about the Palo Alto SRC's independent-minded engineers.

  "We're having some difficulty with the group," he griped, "now that we're trying to tie them more directly to the business strategy."

  Kearns chuckled. "Ken," he said, "that's how you got them in the first place."

  Spencer, meanwhile, was left with the task of restoring the morale of dozens of PARC engineers shocked and upset at Taylor's departure. Shortly after the ouster he joined the Learning Research Group at one of its last Pajaro Dunes retreats.

  The atmosphere of change could not have been stronger. Kay and his fecund imagination were gone. Adele Goldberg and Dave Robson had published the first commercial guide to Smalltalk, which Xerox had officially released to the public.

  PARC's original three laboratories had fissioned into six, including the Intelligent Systems Laboratory under John Seely Brown—spun off from SSL, which was recast as the System Concepts Laboratory under Goldberg. With the exception of the Optical Science Laboratory, which was still headed by John Urbach, all of the original PARC labs had new managers.

  Spencer took it as his duty to communicate to Alan Kay's old group how different the world had become. The PARC of Jack Goldman and George Pake, of researchers following their instincts into a new world without the least concern for corporate imperatives, was gone. Pake's original deadline—the ten-year grace period before Xerox would see results from PARC—had passed. In that time the center had given the company the laser printer, Ethernet, and the technology of the Star, but there was more to do. Henceforth the researchers would have to play a more direct role in helping the company exploit their knowledge. People like Taylor, with his worldview of scientific research and corporate profit as two somehow antagonistic forces, were now in the way.

  "I can still see myself sitting with Spencer on the steps looking out at the dunes, toward the water," recalled Diana Merry. "I was trying to get him to explain to me why he thought he had to push Taylor out. He said “Well, you know, he just wouldn’t play on the team.”

  "He said, 'Well, you know, he just wouldn't play on the team.'"EPILOGUE

  Epilogue

  Did Xerox Blow It?

  Xerox could have owned the entire computer industry today. Could have been, you know, a company ten-times it’s size. Could have been IBM of the nineties. Could have been the Microsoft of the nineties."

  The speaker is Steve Jobs, interviewed for a 1996 public television documentary on the history of the personal computer. The castigation is familiar fare. Almost everyone aware of PARC's achievements also has heard the corollary that Xerox, having invented the technology underlying present-day personal computing, committed the monu­mental blunder of letting it slip through its fingers.

  This image of Xerox as a company uniquely maladroit at exploiting a new technology offers the incontestable allure of a high-concept Holly­wood melodrama, in which the protagonists are the inventors of dazzling innovations and the villain is the dead hand of corporate bureaucracy. The record is certainly damning: Xerox had the Alto; IBM launched the Personal Computer. Xerox had the graphical user interface with mouse, icons, and overlapping windows; Apple and Microsoft launched the Mac­intosh and Windows. Xerox invented What-You-See-Is-What-You-Get word processing; Microsoft brazenly turned it into Microsoft Word and conquered the office market. Xerox invented the Ethernet; today the bat­tle for market share in the networking hardware industry is between Cisco Systems and 3Com. Even the laser printer is a tainted triumph. Thanks to the five years Xerox dithered in bringing it to market, IBM got there first, introducing its own model in 1975.

  Nor does the saga lack aspects of comic opera. There is Bob Potter beating John Ellenby in the battle of cost analysis, then losing the battle of the marketplace with an obsolescent machine; Harold Hall launching the company's most far-reaching new product initiative with a second­hand staff of ten; Jim O'Neill vetoing the sale of five printers to Lawrence Livermore National Laboratory because it might cost Xerox $150,000 over the contract term, in a period when the company's profits ran at more than $350 million a year.

  When a PARC invention survived Xerox decision-making and reached the market, the determining factor was usually a stubborn champion placing his career on the line. One thinks of the difficult gestation of the 9700 laser printer: Jack Goldman flying two executives to PARC for an eleventh-hour demonstration of its superiority over the CRT-based "Superprinter," followed by Jack Lewis's ignoring three successive direc­tives to kill the project. The factors governing the corporation's decision­making on dozens of occasions when its future hung in the balance were not technologies and opportunities, but personalities and politics.

  Yet to chalk up the mixed fate of PARC's technologies purely to Xerox's blundering, as has been done for many years, is misleading. It encourages others to believe that the commercializing of advanced new technologies is easy, provided only that one has the will to do so; and that a company's early domination of a high-tech market will reward it with an unassailable competitive advantage for decades to follow. It presupposes that a corpo­ration should invariably be able to recoup its investment in all its basic research—a mindset bound to lead not to more effective corporate- funded basic research but simply to less of it. And it overlooks Xerox's generous funding of PARC to this day—despite the extinction of SDS, which the research center had been created to serve.

  ***

  The notion that Xerox squandered a golden opportunity to monopolize the personal computer business—that it "fumbled the future," to para­phrase the title of a 1988 case study—rests on several very questionable assumptions. One is that any company can control so polymorphous an industry for very long. The fact is, the technologies of personal comput­ing bestow their commercial favors with great capriciousness.

  One need look no further than the roster of companies that at one time or another claimed a lead in computing to see how reso
lutely the technology foils its tamers. The IBM Personal Computer became the industry standard almost instantly upon its introduction in 1981; yet as of this writing IBM's market share in desktop personal computers is infinitesimal. Apple Computer's Macintosh, the most successful com­mercial expression of PARC's design principles, was launched in 1983 and by 1985 ranked as the world's most popular personal computer; far from parlaying that advantage into a lasting franchise, Apple commit­ted a series of management blunders that today leave its very existence in doubt. (At the very moment Steve Jobs was so self-confidently cri­tiquing Xerox's performance for PBS, his old company was sinking toward a single-digit market share for the first time since the launch of the Macintosh. It soon breached this dubious milestone.) AT&T, Sony, even Exxon all tried to grab a share of the PC business by throwing their marketing muscle behind the packaging of cutting-edge technol­ogy, and all suffered embarrassing flops.

  At this writing Microsoft remains the single most formidable force in the computer industry. This position it achieved thanks not to hard­ware but (as Butler Lampson, Alan Kay, and Bob Taylor predicted would happen) to the power of software. Yet Microsoft is not without challenges that may someday condemn it to the same fate as IBM, whose market position was once considered every bit as unassailable. These challenges include an antitrust attack by the government, the appearance of new technology platforms in which its dominion is less than absolute, and distracting and costly dalliances in entertainment "content" and cable television services. Tomorrow may bring new, unexpected threats. Whether Microsoft itself will be the "Microsoft of the nineties," in Steve Jobs’s phrase, will not be known until well after the turn of the millennium.

  Another dubious assumption is that Xerox, simply by dint of its size and marketing savvy, should have been well up to the demands of com­mercializing the Alto, Ethernet, and dozens of other orphaned PARC innovations. This argument is most often expressed as a question: "If little Apple could sell personal computers, why couldn't big Xerox?"

  The answer, of course, is that Apple was able to market the PC not in spite of its small size, but because of it.

  Commercializing a radical new technology often means betting the company on the outcome. In 1981 this had decidedly different meanings for Xerox and Apple. Xerox employed 125,000 workers, Apple forty. Virtually Xerox's entire workforce was focused on selling one type of product: the office copier. They represented decades of corporate invest­ment—hundreds of millions of dollars—in embedded training, technol­ogy, and customer service.

  This might not have been an impediment if the customer bases for copiers and computers were identical. But for the most part they were quite different. Through the 1980s, copiers typically were high-volume machines designed to be installed in central copy rooms inhabited by clerks and secretaries. They were ordered by the same managers who handled the purchases of typewriters and telephones. But computers were systems, designed to support corporate professionals a rung or two further up the ladder. Their purchasing came under the jurisdiction of a new category of professional information manager, a breed entirely unfa­miliar to the platoons of copier salesmen on Xerox's payroll.

  Apple, obviously, faced none of these issues. Young and agile, it was in a perfect position to build suitable factories and a computer- oriented sales force from the ground up—just as Xerox had been when it chose to commercialize another innovative and suspect technology, fifteen years earlier. In fact, the two companies' relationship perfectly illustrates the "creative destruction" model of industrial evolution pro­posed by Joseph Schumpeter in the 1930s, in which entrepreneurial opportunists snatch markets away from their anachronistic precursors.

  Computers, moreover, did not lend themselves to the pricing regime that many considered Xerox's most important invention: leasing the copiers and charging customers by the page. When 250 Xerox salesmen contemplated the Alto on Futures Day and wondered, "Where's the click?" they were asking a question fundamental to the difference between the old machines and the new. To sell computers, Xerox would not only have to build a new land of machine, but also a new system of compensating and motivating its more than 100,000 sales executives. The combined task of retooling plants, retraining factory workers and sales­men, and adjusting to the mentality of a new set of customers resembled that of turning a supertanker around in mid-ocean.

  History suggests that corporations are seldom able to remake them­selves as thoroughly as a Xerox trying to turn into a computer company. How many leading American corporations have survived multiple revo­lutions in industrial technology to stay atop the business pyramid for more than two or three decades? With the exception of perennial leaders such as General Electric and, possibly, Hewlett-Packard, the examples are scarce. The more usual pattern is that a company is granted one or two runs at the fence before it is finally overmatched by changes beyond its control—as a sort of corporate Darwinism ruthlessly ensures that openings will always exist for fresh entrepreneurial companies to arise and nurture the technologies that make every generation's future.

  Even under the best circumstances, the Xerox of the 1970s and 1980s was not a promising candidate to exploit a new technology, gov­erned as it was by the twin burdens of its culture and its business envi­ronment.

  Its culture was formed by betting on a long shot and seeing it come up trumps. For Xerox the spectacular, instantaneous success of xerography has always been something of a mixed curse. While it brought the com­pany fabulous wealth, it blinded management to the fact that sometimes the marketplace needs time to absorb new technologies.

  Xerox "fundamentally was cursed by the Chester Carlson vision," con­tends its former chief technology officer, Paul Strassmann. "This is the immaculate conception view that all you have to do is give us the right technology and the whole world will come to us. Unfortunately, when it happens like that it's a fluke."

  The first generations of post-914 executives were therefore decidedly unreceptive to a business, like computers, that might grow only slowly and for which Xerox would have to battle ferociously for every point of market share. They viewed the Star as the closest thing in computing to the sainted 914—a machine that could spring fully mature from the development pipeline, poised to sweep up the entire market and hold it against all challengers. Had the computer market remained receptive to $16,000 systems-oriented workstations, the Star might well have domi­nated. But the market, of course, transformed in the blink of an eye into something entirely different. Xerox had no machine to compete with stand-alone desktop workstations costing only $2,000.

  Xerox's second burden, too rarely acknowledged by critics of its han­dling of PARC, was the merciless business environment confronting it during this period. Only a few months after Peter McColough pro­claimed the company's commitment to the "architecture of information" his company was overwhelmed by war on multiple fronts—new foreign and domestic competitors, an antitrust investigation by the federal gov­ernment, and a deteriorating economy. Even a management team sea­soned by years of scratching and clawing in competitive markets would have found these challenges daunting; the serene monopolists of Xerox were left as disoriented as weekend sailors caught in an ocean squall.

  The distractions of the "lost decade" left more by the wayside than the "architecture of information" battle cry: Although Japanese com­petitors materialized in force in 1975, Xerox did not introduce a low- cost machine to rival theirs until four years later. It is not surprising that Peter McColough and David Kearns, embroiled in the fight of their lives simply to protect the copier franchise, had scarcely any patience for the revolutionary solutions being floated for the tough problems of technology transfer at PARC.

  In 1972, for example, Jack Goldman presented McColough with a novel plan to expedite the manufacture of laser printers by spinning off an entrepreneurial manufacturing venture. "The idea was to start a lit­tle company in the Palo Alto area," Goldman recalled. "Guys were ready to jump ship and build in garages, assuming Xerox
was willing to make available the xerographic engines. They would get venture financing. One-third would go to capital, one-third to the employees, and one-third to Xerox for providing the Xerox machines.

  "McColough would absolutely not hear of it at all. His attitude on such things back in the 1970s was that we could not encourage our people to go off like that when they had a good idea, otherwise we'd lose their loy­alty. Of course, they all went off anyway, people like Chuck Geschke and Bob Metcalfe who went to start their own companies, not to speak of Charles Simonyi who went off to Microsoft and brought Microsoft Word with him." Today the sort of spinoffs Goldman proposed are common­place at Xerox. But the company needed ten years—and a complete change of management—to learn that lesson.

  Of the many mistakes Xerox committed in handling PARC technol­ogy, some were clearly visible in advance, others only in hindsight. Crit­ics should never confuse the two. The former category includes such blunders as the unnecessary delay in marketing the laser printer and the repeated rejection of the Alto as an alternative to the nonprogram­mable electromechanical word processors manufactured in Dallas.