Read an Excerpt
Chapter 12: Finance: The Rejection of the Alto
Archie McCardell had embarked upon the Hughes strategy effort by noting that, among Xerox's starkest weaknesses, were its failure to "manage the diverse technology needs of the future" and "to demonstrate the ability to market diverse products in the office." In other words, Xerox had yet to transcend the copier business. Peter McColough had not disagreed. Nor could he have, given the abysmal performance of Scientific Data Systems under Xerox management. But McCardell's late 1973 assessment went well beyond Xerox's problems with SDS - more than four years and a billion dollars after diversifying into computers, Xerox had failed to integrate digital capabilities into the mainstream of its organization. And none of the new endeavors stood more isolated than PARC.
"They are newcomers to the world of Xerox. Too new to know the traditions," wrote a company publicist about PARC's computer scientists for a 1971 edition of Xerox World. "Too new to know the lore of the 914, and the laboratory on Rochester's Hollenbeck Street, and the people who made Haloid into Xerox. Many of these newcomers have been part of Xerox for less than a year, most for less than two. Yet, talking among them, watching them work, one has the feeling that something akin to that Haloid energy is here in California."
Energy, enthusiasm, dedication-yes. But the organizational comparison of Xerox's two most significant scientific efforts ended there. Unlike Haloid's technicians of the 1950s, PARC's scientists were asked to research ideas, not develop products. The Haloid engineers had worked in the same town as other companyemployees; PARC sat a continent away from Rochester and Stamford. Most important, those who refined xerography for Haloid did so as partners with sales, manufacturing, and finance in Joe Wilson's 91 or nothing struggle for a viable office copier business. By contrast, the digital wizards in Palo Alto defined an 44 architecture of information" with neither active participation nor broad commitment from the business side of Xerox.
"PARC was floating around in free space," says George White.
"PARC was a head. But a head to which body? Who was going to pick up from whatever was done at PARC to do all of the rest of the hard work to make a business out of it?"
The answer, according to Jack Goldman's original 1969 scheme, was to have been SDS. Goldman's plan, however, reflected assumptions drawn from his own experience rather than from the history of Xerox's computer division. He had spent his career at Ford and Xerox, industrial mammoths that could afford to allocate significant resources to research and development. SDS was different. In 1969, it was less than one-tenth the size of Xerox and not yet a decade old. Its prosperity depended on existing.not future-technology. And it operated in an industry where its vendors, the hardware components suppliers, instead of itself or its competitors, underwrote long-term research advances. SDS lacked the capability to convert laboratory-proven inventions into products. Indeed, surfeited by their impressive early success, SDS's senior executives openly scorned research-surprising Goldman, for example, when they tried to scuttle his proposal for PARC.
Scientists more familiar with the computer industry expected such biases from SDS. Bob Taylor, Butler Lampson, and others from the ARPA research community had struggled to persuade a reluctant SDS to commercialize timesharing when that technology was the latest blossom of interactive computing. Many computer researchers bound for PARC had heard the story of Taylor's confrontation with SDS's chairman Max Palevsky over the future of timesharing. They also considered the timesharing systems SDS eventually produced mediocre, and they questioned the relevance of a mainframe computer manufacturer to the office of the future.
Thus, Goldman's dream that PARC and SDS would combine cheerfully to identify and exploit advanced computer concepts was at odds with reality. SDS belittled the very idea of PARC and lacked the capacity to develop inventions into products; many of PARC's computer scientists scoffed at SDS's talents and dreaded the prospect of working on behalf of the computer division. Far from the alliance Goldman expected, PARC and SDS antagonized each other from the outset.
Their mutual hostility erupted within a year of PARCs creation. In the spring of 197 1, Bob Taylor's Computer Science Laboratory decided to build its own timesharing system instead of using an SDS "Sigma" computer to conduct research. The action infuriated SDS., How would potential customers react, asked SDS managers, to news that Xerox's own computer scientists had rejected the Sigma? How could Xerox afford such adverse publicity, they harped, in light of SDS's worsening financial position? The SDS men shouted foul; they demanded an explanation from PARC's director, George Pake.
Ever since Bob Taylor had warned him that supporting SDS would be a "mistake," Pake had known about the bad blood between many of his researchers and the management of Xerox's computer division. Pake's natural style was conciliatory; he wanted to cooperate with SDS. Nevertheless, an investigation into the matter satisfied him that the Computer Science Laboratory had sound reasons for rejecting a Sigma in favor of constructing their own "home built" timesharing system.
To proceed with its research, the Computer Science Laboratory required access to ARPA software previously developed on a Digital Equipment Corporation computer named the "PDP-10." The PDP-10, however, did compete directly against SDS's Sigma; for the Computer Science Laboratory to buy one clearly would have damaged SDS. By the same token, forcing a Sigma on the lab would have occasioned either years of effort to make the Sigma compatible with the PDP-10 so that it could run the required ARPA software or an even longer period to reinvent the software itself. Both outcomes would have wasted too much time and energy. The "home built" alternative, Pake concluded, was a thoughtful compromise, one that SDS did not appreciate.
"It is ironic," he wrote, "that a question has been raised by SDS over the way in which PARC scientists propose to use noncompetitive mini-computers - plus months of hard work-to get around any PDP-10 embarrassment for SDS. It is unthinkable to me that Xerox sets me the task of hiring creative, imaginative, top-rank researchers and then expects me to insist that they handcuff themselves with inappropriate equipment or that they fritter away their talents in make-work tasks that pointlessly tie SDS equipment into the research apparatus."
In his mind, the hue and cry over the "home built" decision had less to do with the best equipment for research than the willingness and ability of SDS to cooperate with PARC's charter. Instead of supporting PARC in a drive toward the information technologies of the future, SDS pressured Xerox headquarters to subordinate PARC's resources to the shorter-term needs of the computer operating division. To Pake, the current upset merely replayed the unacceptable objections posed by SDS ever since the inception of Jack Goldman's research proposal. He thought the bickering ought to stop so that PARC could get on with providing the technology Peter McColough expected to dominate Xerox in the 1980s.
That technology, Pake noted in his May 1971 memorandum to management, "is projected to be information systems we do not yet know how to build, that will include technologies we have yet to discover. For this, Xerox will require creative people exploring the frontiers of information science with the most effective tools available for the job."
Speaking of those scientists and their mission, Pake declared, "I will hire them for their competence and their judgment how best to do that research, and, until they prove my judgment wrong, I will do my best to provide them with the kind of first-rate technical support it is reasonable to expect in Xerox research laboratories. If that is the wrong way to build a first-rate corporate research center for Xerox, then I am the wrong man for the job."
Once Pake put his company badge on the line, SDS backed off and never again threatened PARC's budget or purpose. Nevertheless, the episode had continuing repercussions. In the midst of the controversy, Pake had argued that SDS desperately needed to organize a development function capable of turning PARC's inventions into products. Otherwise SDS - and Xerox-would remain critically disconnected from whatever PARC created. Unfortunately, however, Xerox waited several years before taking steps to overcome this serious flaw in its organization. And when, in 1975, the company finally did establish a development group for PARC, SDS had all but expired as an operation, leaving the new development group without any natural partners of its own in manufacturing, marketing, or sales.
The "home built" controversy also fanned the condescending attitude of many PARC researchers. When the Computer Science Laboratory completed its timesharing machine in the summer of 1972, one of the lab's resident wags christened the computer "MAXC." Literally, the acronym stood for "Multi-Access Xerox Computer." As any of PARC's denizens would volunteer with a smile, however, the C was silent. In a sarcastic swipe, PARC named the system SDS never wanted built after SDS's founder, Max Palevsky. A funny enough pun, but at SDS - deep into a third consecutive year of heavy red ink-no one laughed....
