Building a computer as a company, Apple soon discovered, was far more laborious than knocking together a machine in a garage. “The Apple III was designed by committee,” Randy Wigginton complained. “Apple felt that was the way a proper company should design a computer. Everybody had certain ideas about what the Apple III should do and unfortunately all of them were included.” The general plan was for a computer that contained all the features that were missing from the Apple II and to stretch the powers of the 6502 microprocessor since more powerful processors were not available at low prices. It was to have a larger memory, a built-in disk drive, a better operating system, a display of eighty columns that would be suitable for word processing and spreadsheet calculations, an upper- and lower-case keyboard, a keypad, improved color, and a faster microprocessor. It was also supposed to run all the programs developed for the Apple II and so become instantly useful in scores of different applications.
A fearsome pressure built up inside the company and helped promote stomach-contracting schedules. Some of this stemmed from marketing projections that repeatedly forecast imminent declines in the sales of Apple IIs. Wendell Sander, the chief hardware engineer on the Apple III, said, “We kept wondering when the bubble was going to burst on the Apple II. We could have done with more professionalism from the marketing side.” Pressure also sprang from the commitments to ship the Apple III that were made in the prospectus prepared for the public stock offering. None of this was helped by Jobs who, a few months before the computer was announced, doled out some glossy posters carrying the line THE DECISION YOU’RE MAKING NOW HELPED SHIP 50,000 APPLE IIIS IN 1980. The combination of pressures was sufficient to squelch the cries of anguish and dam the stream of frantic memos that circulated among the people under the most strain. “It was the classic story,” said Jef Raskin, “of people at the bottom saying, ‘Things aren’t working here. We’re in trouble.’ Then the next level up would say, ‘We’re in some trouble with this,’ and the level above would say, ‘We’re getting around the trouble,’ and the people at the top would say, ‘It will be okay. Let’s ship.’”
The rush to ship the computer resulted in an all-out scramble that was reflected most keenly in the publications department where the technical writers were again sandwiched between the changes being made on the laboratory bench and the implacable demands of the marketing department. The writers did not see the Apple III until nine weeks before it was announced, and the deadlines offered so little slack that the procedures for reviewing the manuals and the computer were all but ignored. Drafts of the completed manuals were sent to the engineering, marketing, and new-product-review departments on the same day they were delivered to the production department for paste-up. There the programmers worked two-hour shifts helping the graphic artists lay out the pages.
Meanwhile, Apple was also learning that there was nothing like software development to illustrate how quickly a year could slip by. Though the Apple III was supposed to run all the programs written for the Apple II, the improvements and modifications made adjustment of the Apple II software a complicated and tiresome venture. The programmers had to accommodate all the changes in the hardware: The computers started differently, the keyboards and disk drives were laid out differently, and the memory had been expanded. The programmers were also submerged beneath the sheer weight of the programming, which was ten times as much as for the Apple II.
Though the burden had increased Apple decided to try to develop as much software as possible inside the company. Little attention was paid to working closely with outside software houses, and there was a distinct effort to tighten up on the distribution of technical information about the intimate secrets of the Apple III. This made it almost impossible for independent software companies to develop programs for the computer. Two weeks before the announcement, a prototype machine was delivered to Visicorp accompanied by a request for a demonstration program of Visicalc. It was a year after the announcement before Apple’s programmers had finished modifying the Pascal language so it would work on the computer and thus give independent software a way to write programs other than with BASIC or assembly language.
The Apple III was announced with great fanfare at the National Computer Conference in Anaheim in the summer of 1980. Apple rented Disneyland for an evening, distributed twenty thousand free tickets, and hired a fleet of red double-decker buses to ferry guests to the amusement park. The splash didn’t deceive anybody in Cupertino. Sherry Livingston recalled, “They blew the Apple III and they knew it when they announced it.” Once the public promises had been made, Apple was hoist with its own petard. The pressure to ship started backbiting between the competing interests of the engineering, marketing, manufacturing, and corporate sides of the company.
Problems with the design, some of which resulted from creeping elegance, made it impossible to squeeze the computer into its case. This resulted in a second, clumsy board which had to be piggybacked on the main printed circuit board. In addition Apple didn’t pay much attention to testing quality. In the garage Jobs and Wozniak had performed their own crude, yet competent, tests, but as Apple grew no department had been formed to monitor the quality of parts. Wendell Sander said, “We didn’t have any way of comparing the quality of components. We didn’t have enough component-evaluation engineers to test the choice of connector. We listened to the salesmen and believed what they said.” A chip from National Semiconductor, which was supposed to provide the computer with a clock, usually failed after about three hours, and though Jobs savagely berated the chief executive of the semiconductor company, that didn’t solve the problem.
Lines on the printed circuit boards were too close together and that led to shorts. “We screamed that it shouldn’t be shipped without new boards,” said Rick Auricchio, “but the marketing people said it wouldn’t be a problem. The engineers said it would be.” The production team had its own gripes. Screws were so positioned that they pierced cables inside the computer. A heavy metal case was used because of uncertainty about some FCC regulations, but this made it unwieldy for many of the smaller women working on the assembly line. “It finished up as a mechanical nightmare,” said Roy Mollard, the production man. “The engineers washed their hands of it and said it was a manufacturing problem.” The connector between the two printed circuit boards didn’t have enough plating and kept shorting; chips slipped from sockets and the cables to the keyboard were too short. As a test and to help fasten the chips into the sockets, the engineers suggested that the computer be dropped three inches. The shock of the fall, the engineers said, was guaranteed to coax the computer to life. The manufacturing men devised a more scientific test to see if everything worked: They started hitting the computer with rubber hammers.
By then the damage had been done. The Apple III was bollixed up at almost every stage of its development. What was shipped was unreliable and prone to failure. Visicalc was included in the early shipments because no other piece of software was ready. The Apple software that accompanied the computer was untested. The manuals looked shoddy and were accompanied by twenty pages of corrections. Word began to seep out when buyers discovered that the computer was full of startling surprises, SYSTEM FAILURE flashed in an aberrant manner across the screen. Damaging newspaper articles began to appear which wreathed the machine in a funereal cloud. Apple stopped advertising the computer, subjected the machines to arduous tests, redesigned the circuit board, readied some software, allowed early customers to swap their machines for ones that worked, and reintroduced the machine (with an expanded memory) a year later. What eventually became a sound, reliable workhorse and a capable business computer was ruined by the disastrous introduction and Jobs’s optimistic poster became an embarrassing reminder of what might have been. For in the three years following its introduction only sixty-five thousand Apple IIIs were sold.
Jobs, who hopped away from the Apple III once the look and shape of the computer had been settled, was always more interested in the development of Lisa. Work had started on Lisa before the Apple III and from the start it was seen as a bolder, more ambitious project. In October 1978, or almost five years before the computer was shipped in any volume—at a price of around $10,000—Jobs had visualized what he wanted it to look like. He knew he wanted a computer that incorporated the disk drives and screen and also had a detachable keyboard. He knew too that he wanted it built around a sixteen-bit microprocessor rather than the eight-bit device that sat at the center of the Apple II. And he also had an inkling that word processing and a spread-sheet program like Visicalc would have to be included. A preliminary paper drawn up to accommodate most of these ideas bore out a colleague’s observation that Jobs “decided what he wanted Lisa to look like before he was sure what technology would be in the machine.” The original estimates for the computer called for shipment in January 1980, a retail price of $2,000, and a manufacturing cost of $600.
A small group formed to work on Lisa was quartered in Apple’s one-time home—the office suite behind the Good Earth building—and began to grope toward a target that was, to say the least, hazy. For close to eighteen months the project floundered. Occasionally it was interrupted by hiccups and spurts, by the arrival of new managers or by politicking. But there wasn’t all that much contact between the planners and the laboratory, or even between the software and hardware engineers. The general questions of who would use the machine and how it would blend with Apple’s lines of distribution were, for the most part, sidestepped. Left to their own devices, the hardware engineers built a prototype around an eight-bit chip, the Intel 8086, that turned out to be slow and disappointing. Others started to investigate the possibility of using a competing eight-bit chip, the Motorola 68000 (the successor to the eight-bit 6800 that Wozniak had used during development of the Apple I).
Another prototype was developed under the guidance of Ken Rothmueller, a one-time engineer in Hewlett-Packard’s Instruments Division. His computer was calculated to win the hearts of the sort of people who ran data-processing departments in large companies and those with a technical bent. It had a green screen controlled in the same way as the screens on the Apple II and III, a conventional typewriter keyboard, and an overwhelming gray formality—and it by no means matched Jobs’s aggressive spirit. The cynics said that it was a dull, solid machine Hewlett-Packard might have introduced.
Progress wasn’t helped by the carping and clash of ideas that developed between Rothmueller and John Couch, then the head of the software department. Each man had, at various times, worked for the other at Hewlett-Packard, and at Apple they both reported to the same person. It was a battle for control of the computer: a struggle for dominance between hardware and software. But the critical importance of the software was emphasized by work that was conducted, not by Apple, but at Xerox Corporation’s Palo Alto Research Center (PARC).
Xerox Corporation indelibly altered not only Jobs’s picture of the future but also the tone and nature of the computers that he was later to say would last Apple through the eighties. It enlarged his ideas and provided the specter of a competitor that, in the laboratory, was working on ideas that were far more dramatic than those being considered at Apple.
Set on a gently rising hill south of Stanford University, the research center had been built by Xerox as an incubator where young, bright sparks could dream up some grand new ideas that would have the spectacular impact of the company’s copiers. It had opened in 1969, but researchers hadn’t hatched any golden egg by the last month of the seventies when a group from Apple arrived to inspect the results of their work on personal computers. Xerox had spent more than $100 million at PARC to fund research on computers, semiconductor-chip design, and laser printers. That sum was more than double Apple’s total sales in 1979 but PARC had managed to demonstrate the wide gap that exists between the laboratory bench and the shop window.
Xerox’s substantial financial stake in Apple certainly smoothed the way for the expeditions of programmers and engineers that set out from Cupertino. But the curious didn’t need the mind of Sherlock Holmes to figure out what was going on at PARC. The center was a showcase for Xerox and visits by outsiders were part and parcel of daily life. Even without knowing the details of how Xerox’s prototype desk-top computers were designed, anybody who kept in touch with the field was certainly aware of some of the broader trends. A few well-placed telephone calls, a piece of cocktail-party chatter, or some interrogation of the bright high-school students Xerox was using as guinea pigs could shed light on the obscurer points. Computer journals had carried papers that reported on aspects of the Xerox research. A special 1977 issue of
Scientific American
contained an article by Alan Kay, an airy spirit and one of Xerox’s principal scientists, that described the work in Palo Alto and amounted to an enthusiastic prescription for personal computers that were easy to use.
More than a decade of research by scientists like Douglas Engelbart at the Stanford Research Institute, by child psychologists, and at Norwegian universities had all, to varying degrees, influenced the work at Xerox. Indeed some of the most important principles had been published in the mid-sixties and had been displayed by SRI as early as 1968 with a demonstration of a system called NLS. Its chief thrust was to find ways to help people with no technical training control computers. In a way it was an academic extension of the general effort of the hobbyists: to make computers personal and then remove, or at least conceal, the mysterious and intimidating elements. Xerox’s prototypes bore the traces of people who believed that computers were an entirely new medium, that they were far more than stolid, passive devices suitable for crunching numbers and editing prose. Some of the PARC researchers possessed the imaginative vision of flexible machines that would eventually combine the sensory charms of color television, stereophonic music, and finger paints. Like others before them, they said their eventual goal was to pop a computer into a case the size of a notebook or build a machine that businessmen could slip into their briefcases but use to communicate with other computers and people anywhere in the world.