Dynabook and the Legacy of Xerox PARC

When Alan Kay was a teenager, he was skeptical about adults. He wasn’t high on traditional public education, either. These feelings stayed with him as he matured into adulthood. (In fact, he says, he never felt like he truly grew up.) Thanks in part to his two early dissatisfactions, Kay became the first computer scientist on record to envision a portable computer. If Kay hadn’t done it, somebody else doubtless would. Still, credit goes to him for launching the quest for portable computing. This chapter tells the story of the earliest days of what would eventually grow into a gigantic industry.

In the fall of 1968, Kay, a 28-year-old computer science graduate student at the University of Utah, flew to Boston to visit with Seymour Papert, a professor at the Massachusetts Institute of Technology. Papert was a disciple of educational theorist Jean Piaget, who believed that children could learn better by doing than by listening to lectures from teachers. A year earlier, Papert had designed one of the earliest computer learning programs for children, called Logo. During Kay’s visit, they toured a classroom where Logo was being tried out by children on large computer terminals wired to a distant mainframe computer. After the classroom tour, Kay and Papert had a long, stirring conversation about the potential for the computer as an educational tool. “This was a transformational experience for me,” recalls Kay. “I thought the computer could be the next big thing since the printing press.”

On the flight back to Utah, Kay sketched an illustration of a boy and a girl playing a game they had designed themselves on a small notebook-sized computer. This was the birth moment for his vision of a computer that three years later he called Dynabook. Kay’s idea was that a small portable computer loaded with educational software could serve as a powerful tool for learning and self-expression: It could be an intellectual amplifier for children. At the time, most computers were massive devices that were used primarily by scientists, weapon builders, and corporate data crunchers—so it was quite a stretch of the imagination to see them as personal mobile companions for kids. Also, the notion of empowering children and freeing them from the domination of the adult world was a radical one. “I felt that adults don’t do a good job of bringing up children, so children don’t grow up with the richest ideas in their culture,” he says. “Instead, they grow up influenced by mundane ideas, and that’s hugely dangerous. I thought about bypassing adults with the Dynabook.”

Kay himself had lived an unconventional childhood. His father was a college professor who changed jobs frequently. Kay spent some of his early years on a farm in Massachusetts where his grandfather had a collection of more than five thousand books. He was a loner, spending hours by himself in the family library. In his early teens, when the family lived in New York City, he was a frequent visitor to the New York Public Library in Jamaica, Queens. He also enjoyed repairing machines and tinkering with radios. He first got his hands on computers when he joined the Air Force and learned software programming. After his service, he went on to study math and molecular biology at the University of Colorado.

It was as a Ph.D. student that Kay came into his own as an inventor. In Utah, he came upon the work of Ivan Sutherland, who a few years earlier had written a computer program called SketchPad, which was the first software with a graphical user interface rather than just text and numbers on a screen. He saw a demonstration at the Rand Corporation think tank in California where a Rand scientist wrote with a stylus on an electronic tablet and saw the words appear on a computer terminal screen. Also, he had seen a movie showing computing inventions made by Douglas Engelbart and his team of researchers at SRI International, including the computer mouse, e-mail, and bit-mapped computer display screens. “Englebart is cosmic,” Kay told me. With all of these experiences as inspirations, Kay and Ed Cheadle, an aerospace engineer in Salt Lake City, were working on a prototype desktop computer, which they called the FLEX Machine.

So when Kay returned from his visit with Papert he already had some of the ideas that would go into the Dynabook concept. Over the next few days he sketched out designs for a mobile computer—some flat tablets and others like steno pads, with lids that held a flat display. He recalled from his reading that Aldus Manutius, a sixteenth-century Italian printer, had created the concept of the modern book as something that would be portable enough to fit in a saddlebag on a horse. Using a book as a model, Kay taped together a cardboard mockup of what the Dynabook computer might look like, and filled it with led shotgun pellets until he decided that he had reached the limit of what people would be willing to carry around. The optimal weight he decided on: two pounds.

Some of the earliest computer pioneers had thought of portability as an element of computing. In his prescient essay, “As We May Think,” published in the July 1945 edition of the Atlantic Monthly, Vannevar Bush, head of the U.S. Office of Scientific Research and Development, envisioned a personal computer, called Memex, which he thought of as an “enlarged intimate supplement to his memory” where the individual would store books, photographs, documents, and communications. Bush pictured a desklike machine including screens, a keyboard, buttons, and levers that could “presumably be operated from a distance.” Toward the end of his essay, he suggested that someday information and impressions might be directly transmitted from the human brain into machines. In other words, the person and the computer would be melded together—the ultimate in portable computing.

In 1968, at the same time that Kay was dreaming up the ideas that led to Dynabook, Bill Gates and his buddy, Paul Allen, later his cofounder of Microsoft, were taping into huge mainframe computers via timesharing accounts at Seattle’s Lakeside School. Gates recalls carrying home a General Electric data communications terminal and attaching it to the family telephone using an acoustic coupler so he could dial into the distant mainframe. “It was kind of portable computing,” he says. Gates wrote a Monopoly game program at home that way. Still, what he had at his fingertips in 1968 was a far cry from a truly mobile computer.

Kay never did make a Dynabook. The digital electronics technologies just weren’t advanced enough in the late 1960s and 1970s to shrink a computer down to notebook size. Instead, in his next career stop, as a researcher at Xerox’s famed Palo Alto Research Center (PARC) in California, he contributed some of the thinking that went into the Alto machine, which is widely considered the world’s first personal computer. Kay called it “the interim Dynabook.” He teamed up with a handful of PARC scientists, including Charles Thacker and Butler Lampson, to bring together the knowledge and funding necessary to develop the Alto. The beige Alto, conceived in 1973 and produced a year later, was designed to sit on a desktop and to be operated with a mouse as well as a keyboard. While his colleagues concentrated on preparing the hardware for Alto, Kay designed a software language, called Smalltalk, for programming the computer. It included a graphic user interface—which made it possible for the user to navigate the computer using icons, menus, and overlapping windows containing various programs. Again, Kay had children and education in mind. He knew that children are naturally very visually oriented, and, by using a graphic interface, he hoped to engage them more deeply in computing than a simple text interface could achieve. Using a mouse, they would click on icons and drag them around on the interface, engaging what he called their “caveman brain” and practicing eye-hand coordination. “The big idea was that they would essentially be having a conversation with the computer, and, through the interface, we’d put what the computer was thinking on the screen,” Kay says.

In the early 1970s, in the Bay Area, personal empowerment and communalism were in the air. The Free Speech Movement at the University of California at Berkeley, the anti–Vietnam War movement, free love, equal rights for women and minorities, rock music, psychedelic drugs—all of these influences mixed together and created an environment where revolutionary ideas flourished. And the Alto was one of them. What could be more antiestablishment than a computer that could give any individual the same computing capabilities as The Man? PARC was financed by a corporation, and some of the researchers were involved in the U.S. Department of Defense’s ARPA program (which brought us the Internet), so it wasn’t exactly a hotbed of social revolution. Still, it was very much of its time and place. “PARC wasn’t so much the counterculture, but it was influenced by it in the most positive way,” says Kay.

Though Xerox’s Alto was never commercialized, it inspired many a Silicon Valley entrepreneur. Among them, most famously, was Apple cofounder Steve Jobs, who visited PARC in 1979. There, watching a demonstration of the Alto, Jobs was excited by the possibilities of the computer mouse and the Alto’s graphical user interface. A few years later, Alto’s interface became the inspiration for the breakthrough Lisa and Macintosh computers from Apple, and, later, to Microsoft’s Windows operating system.
In 1976, a handful of PARC scientists had squeezed the Alto’s electronics into a smaller package called the NoteTaker. Though only a handful of prototype machines were built, this was arguably the first portable computer. Inspired by Kay’s Dynabook, the NoteTaker was developed by a team that included Douglas Fairbairn, Larry Tesler, and Adele Goldberg. The device looked like a portable sewing machine. To use it, you would lay it on its side and fold out the bottom, a keyboard, to reveal a built-in display monitor and the opening for a floppy disk drive.

Kay’s Dynabook concept had many disciples. Throughout the coming years, a succession of entrepreneurs, inventors, engineers, and designers labored to make his dream of portable computing a reality. “The idea was influential. It was a great science fiction story that gave people something to aim for,” says Steve Sakoman, the original engineer behind Apple’s infamous Newton handheld computer.

One of the first people to set the goal of building a commercial portable computer was Blair Newman. A Harvard Business School graduate who had worked briefly for Howard Hughes in his Las Vegas underground bunkers, Newman had found his way to Silicon Valley in the 1970s and established himself as a certified Big Idea man. “He came up with an idea twice a day you might want to start a company around,” recalls William M. “Trip” Hawkins, an early Apple employee and friend of Newman’s who later went on to found computer game giant Electronic Arts.

After graduating from Stanford Business School in 1977, Hawkins met Newman. Hawkins was working as a freelance researcher for a tech market research outfit, Creative Strategies, for which he wrote one of the first major national market research studies on the personal computer industry. Newman was one of his sources. After the study was published in early 1978, Apple’s Steve Jobs read it and, according to Hawkins, was upset that Hawkins had reported that Tandy’s RadioShack chain had sold more computers at that point than Apple had. Newman gave Hawkins’s phone number to Jobs. Hawkins must have impressed Jobs, because, in the summer of 1978, Jobs hired him as one of his early marketing analysts. He also brought on Newman as a consultant.

Almost immediately, Newman started plotting to start a new company. He had several ideas, including building a mobile computer that would come equipped with a package of business applications—one for number crunching that would later be called a spreadsheet. (This was a year before Dan Bricklin and Bob Frankston at Software Arts would produce VisiCalc, the first commercial spreadsheet program, for the Apple II computer.) Newman recruited Hawkins into his schemes. At the time, Hawkins and Newman shared a tiny ranch house in Palo Alto, just off the Stanford campus. “Blair drove roommates crazy, but he moved in with me until he drove me crazy and I kicked him out,” recalls Hawkins. “He’d be up all hours of the day and night and he’d talk until you shut off the lights.” In the apartment and at a Good Earth restaurant near the Apple campus in nearby Cupertino, the two stirred up a business plan. The idea was that this would be the first personal computer designed specifically for business use. Until then, PCs were for hobbyists and students. Their computer would be sold through retail stores—which would have been another first for a business computer. They didn’t think of offering the idea to Apple, since it was just establishing itself with desktop PCs for consumers. (Apple had sold only 1,000 computers at that point.) Hawkins wrote a 15-page business plan and, at Newman’s insistence, went alone to peddle it to one of the most important venture capitalists in Silicon Valley, William H. Draper III of Sutter Hill Ventures. “He [Newman] saw me as the front man, and the right socioeconomic model to give the business credibility,” says Hawkins.

Hawkins had an MBA, but still he was none too sophisticated. He was only 24 years old and had less than a year of tech-industry work under his belt at a time when most of the people who were approaching venture capitalists in the Valley had much stronger résumés. For his meeting with Draper, Hawkins wore an inexpensive suit he had bought after graduation at Filene’s Basement but had been too lazy to alter—so it had a pin in the waistband to hold his pants up properly. The Sutter Hill offices were in a high-rise office building to the south of the Stanford campus, and stuck out amid the sprawl of tech office parks. Even though he had a business degree, Hawkins didn’t feel like he belonged in a corporate office and was nervous as he rode up in the elevator and intimidated by the plush furnishings in Sutter Hill’s waiting room. Finally, he was ushered into Draper’s office, which had a window with a view of the Stanford campus. Draper was a distinguished-looking middle-aged man in a dark, tailored suit.

The way Hawkins recalls this meeting, Draper grilled him casually for a few minutes about his business plan. Then, as Hawkins remembers the scene, Draper dressed him down: “He called me a complete idiot. He told me there was no way businesses would buy a computer from a retail store.” Then Draper excused himself and left the office, returning in a couple of minutes with a younger colleague, who attacked the business plan even more aggressively than Draper had. “He was like a Doberman,” Hawkins says. “I had never been in a meeting before where there was so much skepticism about an idea. I felt like my cheap suit got shredded.”

Many years later, Draper didn’t remember this encounter, and he said he wouldn’t have been so critical of a young, would-be entrepreneur. Still, he said, he would have been skeptical of a plan involving mobile computing. “Others may have anticipated the mobility of computers back then, but I was not one of them,” he admitted.

After the Draper session, a dejected Hawkins sought advice from his boss at Apple, Armas “Mike” Markkula Jr., who was then running marketing. Markkula had heard through the grapevine what Hawkins and Newman were up to and he didn’t like it. “He told me it was a really big mistake. I had the opportunity for a good career at Apple. I should drop this thing and get back to work” Hawkins recalls. “So I did.” The outcome wasn’t so good for Newman. He was fired.

Hawkins went on to eventually become Apple’s director of strategy and marketing before quitting in 1982 to launch Electronic Arts (EA). He left EA in 1991 to start 3DO Corp., a futile attempt to knock off Nintendo and Sega as the top video game consoles, and later got into mobile computing for real for the first time with Digital Chocolate, a mobile games publisher. Newman continued to have groundbreaking ideas. His MetaView Corporation designed a computer-television hybrid in the late 1980s that presaged TiVo’s digital video recorders, which came a decade later. Newman ultimately fell into a downward spiral of depression. A habitué of the Well, an early online community, he in 1990 wrote a computer program that deleted every word he had ever written on the Well’s pages after he got in a dispute with some other members. A few weeks later, he committed suicide. His friends said he killed himself first virtually, and then really.

While Hawkins and Newman were searching for a foothold in portable computing, other entrepreneurs were comparing notes and inventions at the Homebrew Computer Club. The club is famous for being the place where Apple founders Steve Jobs and Steve Wozniak first showed off the Apple I desktop prototype, but it was also a seedbed for portable computing. The club was started in the garage of electronics enthusiast Gordon French in Menlo Park, California, on March 5, 1975. It attracted a host of early computer mavens and swelled in popularity to the point where, in the late 1970s, it held meetings in the auditorium at the Stanford Linear Accelerator Center in the hills above Stanford University. Ultimately, 23 companies were formed by club members. Two of the regulars were Lee Felsenstein, codesigner of one of the first commercial personal computers, the SOL-20, and Adam Osborne, a former Shell Oil chemical engineer who had become a technical book publisher. One of the books he published, An Introduction to Microcomputers, inspired many of Silicon Valley’s PC-industry pioneers. Felsenstein often moderated the meetings, and Osborne brought boxes of books, which he sold during breaks between sessions. These two combined forces to design and build the Osborne 1, the world’s first commercial mobile computer.

Felsenstein had come of age in the 1960s as a New Left radical and a member of the Free Speech movement at the University of California at Berkeley. In fact, he was one of those arrested at the famous “Sproul Hall Sit-In” in 1964, and later wrote articles for a peppery underground newspaper, the Berkeley Barb. Throughout the 1970s, he designed computers and components for a handful of Bay Area companies while volunteering with community computing organizations on the side. He was familiar with Kay’s Dynabook idea, but, in the earliest days at Homebrew, “Nobody was saying, ‘I’m building a mobile computer.’ It was hard enough to build one that was immobile,” he recalls.

Felsenstein and Osborne hooked up at the West Coast Computer Faire in San Francisco’s Brooks Hall in April 1980. They already knew each other. Osborne had hired Felsenstein to do some manuscript editing for him. Osborne had sold his publishing business to McGraw-Hill in 1979 but was still involved in it. At the Computer Faire, he was hawking books at a booth. Osborne called out to Felsenstein when the engineer sauntered past. He wanted to arrange a meeting to talk over computer product ideas. That was an odd request from a publisher, but Felsenstein was running out of money from his last electronics consulting project, and figured whatever Osborne had to say was worth a listen. Back at home, Felsenstein dusted off some product ideas from his files for the meeting with Osborne. But, when they met, he quickly learned that Osborne already had something specific in mind: a portable computer.
Osborne sketched out what he wanted on a piece of graph paper. Like the Xerox NoteTaker, the computer would be carried like a portable sewing machine and would be set up by turning it on its side and opening the bottom—a keyboard—to reveal a small cathode ray tube screen. The machine would run the then-popular CP/M operating system from Digital Research, in Pacific Grove, California; would have two floppy disks; and would have a five-inch screen. IBM already had a machine in the market that it called a portable computer, the IBM Model 5100, but, at 30 pounds and with a separate data storage unit, it was really a desktop machine. So Osborne’s would be the first truly portable computer to be sold.

It’s not clear exactly where Osborne’s concept came from. People who knew him believe he was inspired by Xerox PARC’s NoteTaker prototype, but Alan Kay says Osborne was “too narcissistic” to give credit to anybody else. Osborne died in 2003, and in his 1984 book, Hypergrowth: The Rise and Fall of Osborne Computer Corporation, he didn’t spell out why he decided to produce a portable. He wrote that he chose the C/PM operating system because he thought it had the best chance of becoming an industrywide standard. He wanted the computer to be inexpensive so the market would be large. And he wanted the machine to be small. He wrote: “I decided to build the smallest possible package, given the components then available; and thus was born the idea of the ‘portable computer.’” Felsenstein loved Osborne’s idea. “Adam said there was a hole the size of a truck in the market and he had decided to fill it,” recalls Felsenstein. “I said, “Why the hell not?” They soon signed a contract. Felsenstein would get $3,500 and 25 percent of the company.

Osborne Computer, the company they formed, introduced the Osborne 1 less than a year later, at the next West Coast Computer Faire. But they packed a lot into those few months. Felsenstein, who was still a freelancer, worked in a shared barnlike space on Berkeley’s Parker Street, an industrial area of the city. He had a drafting board set up in a corner, where, from his window, he could see the Palmolive soap plant billowing smoke in the distance. He designed the first commercial portable computer with paper and pencil, and used a slide rule to do his calculations.

His main challenge was fitting everything into a relatively small space. Osborne had told him to size the computer so it could fit under the seat of a commercial airliner. Felsenstein had earlier designed the first PC modem. But that device was too big for a portable computer. So he saw he had to start from scratch and design a smaller one. To save space on the main circuit board of the computer, he decided to combine the memory for the display on the main memory chip. For the brains of the computer, he chose the Z80 microprocessor from Zilog over Intel’s 8085 processor because it was already running a handful of machines equipped with the CP/M operating system, which was the leading PC operating system of the day.

As a youngster growing up in Philadelphia, Felsenstein had learned electronics by tinkering. He enjoyed playing around with radios, and it was his lucky day when a friend of his father’s bequeathed him a half-completed but paid-up correspondence course in radio and TV repair. His father, a commercial artist, had another friend who had a collection of radio and TV chassis that he handed over to Felsenstein, who scavenged them for parts to use on future repair projects. By the time Lee Felsenstein enrolled at Berkeley, he had a head start on many of his contemporaries. It was through a cooperative work study program that he got the full immersion in electronics—at the NASA Flight Research Center. He dropped out of school temporarily in 1967, after participating in a series of antiwar demonstrations, and kicked around the Valley working for electronics firms before returning to school to finish his degree in 1972. His passion was volunteering to help out in Bay Area community computing projects. He laid out a blueprint for a computer, which he called the Tom Swift Terminal, which he designed from the ground up to be added onto by others in the community. His thinking had been influenced by Ivan Illich’s 1973 book, Tools for Conviviality, which advocated designing technology products so people could learn about them by tinkering and modifying them. Ultimately, Felsenstein launched his career as a freelance engineer.

When it came time to design the Osborne 1, Felsenstein’s ability to tinker and invent on the fly had made it possible for him to create a machine like none he had ever seen before. “The Osborne 1 was a step into the void, a leap of faith,” he says. These days, many computers are designed and built to order for the top PC brands by electronics manufacturing firms in Asia. Felsenstein believes that’s antithetical to innovation. “All you get is small variations on what everybody else is making,” he says. “If you want a design to have any originality, you have no business turning the design over to a manufacturer.”

While Felsenstein was busy with engineering, Osborne, a promoter at heart, was the company’s cheerleader and evangelist. He had a British father and a Polish mother and had spent much of his childhood in India before going to college in the United Kingdom. Though he had been in the United States since 1961, he had retained his British accent. “Adam was like a character out of a Gilbert and Sullivan musical,” recalls Felsenstein. “He was ramrod straight and absolutely confident in himself. He played well on Americans’ tendency to be subservient to people with British accents.”

Gradually, Osborne hired a small crew of engineers and manufacturing people, and contracted with a Silicon Valley company, Sorcim, to write the software. He gave Felsenstein a deadline of the end of January 1981 to have the electronics design ready. The engineer brought it in a couple of weeks late, but missing the deadline was no big deal. The industrial design was done by Housh Ghorbani, a recent émigré from Iran, who worked on it in his Fremont, California, living room. “It was done on a pell mell schedule, so we didn’t have a chance to step back and ask, ‘Are we doing this right?’” says Felsenstein.

It didn’t seem to matter at first. When Osborne showed off the first prototypes at the West Coast Computer Faire in San Francisco in April 1981, his booth was mobbed. This was four months before IBM would announce its first PC. In Brooks Hall, people took turns opening and closing the machines and hefting them. The Osborne 1 was far from perfect. These first models had metal cases and weighed almost 30 pounds. Felsenstein recalls that when he carried one of the machines just four city blocks to the hall, he felt as if his arm was going to come out of its socket. Also, the screens, tiny to begin with, had a wobble caused by magnetic interference from the power transformer. Still, the crowd loved them. Afterward, when Osborne and Felsenstein were walking out of the building, Osborne clapped his partner on the back and babbled excitedly. “He told me, ‘The product is the company. We’ll build it up and sell it off and do a whole string of these!’” Felsenstein says.

Some of the defects were fixed by the time the first Osborne 1 computers shipped out of manufacturing two months later. Theses machines had plastic covers, so they came in at 23 pounds. Still, that was too heavy for easy carrying, and they were called “luggables,” rather than portable computers, by the technology press. The Osborne display was still only five inches across, and the data storage consisted of two single-sided floppy disks, which didn’t have enough capacity to be of much use to a businessperson. The price was $1,795, which was reasonable for the time. Sales were remarkable initially, with more than ten thousand selling per month at the peak. But the phenomenon was short lived. Poor manufacturing practices led to inventory shortages and quality problems. And, within a year, Kaypro released its Kaypro II, which had a nine-inch display and twice as much data storage. To Osborne employees, the Kaypro seemed like a junky computer—partly because its metal case tended to get dented. So they didn’t take it seriously enough. Before long, though, Kaypro sales overtook Osborne’s.

Adam Osborne compounded his problems in 1983 by preannouncing a successor to the Osborne 1, called the Osborne Executive, many months before it was ready to be shipped. That announcement slowed sales of the Osborne 1 to a trickle. His blunder was so monumental that it gave rise to a term that’s still used by some people in the computer industry to warn against preannouncing products: the Osborne effect. The company grew to 1,200 people at its peak, but it was so disorganized, Felsenstein says, that, “half of the people were correcting the mistakes of the other half.” Osborne stepped down as CEO in 1982 and brought in a soft drink executive to fill the role. But the company spiraled downward and filed for protection from bankruptcy in September 1983. A news photographer snapped a picture of Osborne rushing out of the company’s offices holding a briefcase in front of his face. Felsenstein resigned at a meeting of the last remaining 30 employees in the company’s cavernous offices in Hayward, California.

That was the peak of fame for both men. Osborne later started a company called Paperback Software that sold applications such as a database program and a spreadsheet for as little as $99. The spreadsheet got him in trouble with Lotus, whose Lotus 1-2-3 program it imitated. In 1987, Lotus sued Paperback Software for copyright infringement, and, in 1990, a federal judge ruled in favor of Lotus. The important ruling established the legal principle that software companies could protect the “look and feel” of their programs. So Osborne had created three companies—in books, computers, and software—that helped shape the PC industry, but he died in obscurity in India in 2003 at age 64. Felsenstein went on to design computer hardware, including one of the first “wearable” computers, and to work for Interval Research, the electronics research outfit set up by Microsoft cofounder Paul Allen. Still later, he cofounded The Fonly Institute, a consulting organization dedicated to developing low-cost, locally operated computing systems in developing nations. He had never abandoned his belief in the value of community computing.

Felsenstein learned valuable lessons at Osborne: “The major one is that appearances are not reality. People could be domestic animals, and follow somebody who appeared to know what he was talking about. And I learned that if you grow a company around a product, the company still doesn’t know how to put a product into production, and you have a big problem.”

With Osborne a bust, the next significant advance in portable computing came from one of Kay’s colleagues at Xerox. There was a lot of frustration going around at Xerox PARC in the late 1970s. The lab’s scientists there produced one wonderful technology advance after another, including the laser printer, the graphical user interface, the Alto desktop computer, the NoteTaker portable, PC networking, and object-oriented software programming. Yet Xerox brought almost none of it to market. That was particularly galling for John Ellenby, a British economist and computer scientist who had came to PARC in 1974 with the mission of moving the Alto into volume manufacturing so the machines could be tested in real-world situations. It was Ellenby who produced the first truly portable computer.

Ellenby had high hopes when he arrived at PARC from Edinburgh University and Ferranti in the United Kingdom, which was an early maker of mainframe computers. He loved the Alto, but felt it needed to be made more manufacturable, more easily serviceable, and smaller. The first machines actually sat under your desk, so they weren’t properly desktop computers. He nicknamed his personal Alto “Gzunda,” a play on words, since the computer “goes under” the desk, and is like a gzunda, which was slang for a chamber pot in England.

He was still optimistic about Xerox’s prospects in computing in 1977. That was the year he got the assignment of demonstrating the full suite of PARC products-in-waiting, collectively called the Office of the Future, at a Xerox executive retreat at a country club in Boca Raton, Florida. His partner on this project was Tim Mott. The event was called Futures Day, and it was attended by all of the top Xerox executives. The ballroom in the country club was darkened as the executives and their wives walked while huge audio speakers blasted the theme to Star Wars. Ellenby and Mott showed off the Alto II, an upgrade of the Alto, and a color laser printer code-named Pimlico. Afterward, the executives and their wives were allowed to step up to the front and try out the new machines. The women loved them, Ellenby recalls. Many of them were former executive assistants. But the men seemed less interested. “It was a huge problem for Xerox in those days,” Ellenby recalls. “Their machines were recommended and used by women, but all of their executives were men.”

Still, based largely on the recommendations of the women, a number of Alto IIs made their way into Xerox executives’ offices and even into some offices of Xerox customers. Ellenby believed that the only way to produce excellent computers and software was to put early models into the hands of users. A vital part of his job, he thought, was calling on people who were using the prototype machines and getting their feedback. His breakthrough as an inventor came during just such a visit in 1979. The customer, whom Ellenby will only identify as a person who worked in the Executive Office Building, next to the White House, had tried the Alto II, but said he rarely used it because he was usually on the road. What he’d really prefer, he said, was a portable computer small enough to fit in one half of a standard briefcase. That meant the computer would need to have a flat panel display rather than a bulky cathode ray tube, which was in all computers up until that time.

This was a life-shaping moment for Ellenby, and he savors his memory of it. The two men were having lunch in the White House dining room, an elegant place manned by chefs wearing toques. “What he really said,” Ellenby recalls, “is, ‘I want a real fucking computer in half your case,” indicating the briefcase that Ellenby had brought with him to the meeting. From then on, Ellenby referred to the portable computer he hoped to build one day as the RFC, but, when people asked him what RFC stood for, he said “real functioning computer.”

Ellenby wasn’t long for PARC. He had proposed a new version of the Alto, called Alto III, that would be an all-in-one computer like the Apple II, which came out at that time. But the Xerox bosses shot the proposal down. The company was under assault from a handful of aggressive Japanese copier makers, and it was cutting back on lab spending that wasn’t focused on its imaging business. All that computer technology that PARC had been inventing would continue to sit on the shelf. So Ellenby decided it was time to move on.

He and Tim Mott planned on leaving together to create a company and build a portable computer. It would be a thin machine with a multitasking operating system, meaning it could run more than one application at once. It would also come with a set of applications, including word processor, spreadsheet, and e-mail. Unfortunately, Mott never got to join the company. Xerox brass made it clear that they’d sue if Mott departed with Ellenby. But, in a short time, a handful of PARC and ex-PARC colleagues joined him, including Glenn Edens, who would run development, and David Paulsen, who would manage hardware engineering. Carol Hankins joined from Xerox to run software development. Ellenby’s wife, Gillian, a sculptor, had picked a name: GRiD Systems. The couple had a three-room bungalow in back of their house on Harker Avenue in Palo Alto—a mushroom-colored place that was shaded by an enormous oak tree. This is where GRiD would get its start in December 1979.

Despite the modesty of the facilities, Ellenby lined up some very impressive angel investors. A friend and mentor, Gene Amdahl, founder of the mainframe maker Amdahl Corporation, helped Ellenby land Robert Noyce, a founder of Intel, and Silicon Valley superlawyer Larry Sonsini. Ellenby remembered Sonsini showing up at an early company party in the backyard with a six-pack of beer and two pizzas, which he carried in the trunk of his Porsche.

It was clear to Ellenby from the start that industrial design would play an important role in the machine, which was ultimately called the GRiD Compass. For that, he tapped Bill Moggridge, a British designer who had just moved his family to Palo Alto and was setting up a branch office of his design firm in the Bay Area. The two had met a few months earlier at a party in London. Moggridge, who later went on to cofound a leading Silicon Valley design firm, IDEO, recalls their first encounter. “The thing I loved immediately about him was he was intensely curious. He was just bubbling away,” recalls Moggridge. “Most clients I talked to were only interested in their own thing. John was curious about all the different devices and the design process.” Ellenby wanted to hire Moggridge as his vice president of design, but since Moggridge didn’t want to leave his firm, he brought him on as a consultant. Moggridge recalled that when he and his family arrived in the Bay Area from England, Gillian Ellenby was there to meet them at the airport in a huge Dodge van.

Moggridge and Ellenby got down to the serious business of figuring out what GRiD’s portable computer would look like. Ellenby wanted a model he could take around with him to show venture capitalists and potential customers. His concept idea was that this computer would be used by executives. At the time, personal computers hadn’t yet become an office tool. Plus, most executives didn’t type. They relied on their secretaries to do their typing for them. But Ellenby thought that if they made the GRiD machine compelling enough, they could actually change behavior—convincing corporate bosses to begin using personal computers.

This was a new challenge for Moggridge. He had designed computers before, but mostly they were as big as a table. “As a designer, the place you start a project by looking at the current state of the art,” Moggridge says. “But, in this case, we were trying something completely new, so the reference point couldn’t be early examples. Instead, it’s what people will do with the device and what components are available. You have the freedom to create something that’s precedent setting.” The initial design dimensions were 8½ by 11 and three inches thick. Moggridge decided to use a clamshell design, so people would tilt up a lid containing a flat-panel display to reveal a keyboard. He had never seen a computer with a clamshell design before, because there had never been one before. Richard Grant, the leader of the model shop in his London office, made a detailed model in a week. The display was painted to simulate what it would look like. It was black on the outside; with a green, white, and blue keyboard; a little touch screen next to the display; and calculator buttons built in.

The next step was figuring out how to fit all of the components of a computer into such a small space. Today, modems are about the size of a matchbox, but, back in 1980, they were closer to a shoebox size. Ellenby and his colleagues approached a modem company, Racal-Vadic, and said he wanted something along the lines of a cigarette pack. Their first reaction: it couldn’t be done. But, it turned out that some of Racal-Vadic’s engineers had been thinking of ways to reduce the dimensions of their products. They worked with Ellenby’s engineers to shrink their modem drastically. Similar improvements were made with other components.

During the early days of GRiD, Ellenby underwent a personality transformation. Normally a gregarious partier, he toned down a lot and he slept very little. “He went into a monklike state,” recalls Moggridge.
While those difficult engineering tasks were underway, Moggridge ran some experiments to see how heavy the machine could be and how much of impact it should be capable of absorbing. He made up a set of blocks and asked colleagues to carry them around in their briefcases, steadily adding more until they reached the limit of what they could comfortably carry. The limit was eight pounds, and that became the weight target for the Compass. To set durability standards, Moggridge rented an impact recorder and sent it to himself via Federal Express. He found that when the recorder was dropped, the maximum gravitational force (g-force) that was recorded was 60 g’s. He would design the Compass so it would withstand that level of impact. Moggridge and the GRiD team decided to use magnesium for the machine’s enclosure because it was light and strong and could disperse heat from the electronic components inside. Nobody had ever diecast large flat sections of magnesium before, but GRiD’s vice president of manufacturing, Paul Hammel, figured out how to do it. Since then, many laptops and cell phones have included magnesium parts.

A year passed before Moggridge could complete the industrial design. He had to wait for the basic engineering of the electronics to be finished. Ellenby’s engineers had done such a good job of shrinking components that the machine could come in at just two inches thick—though it was broader than the original model. Moggridge added little legs that folded down in the back to tilt the computer at a comfortable ergonomic angle when it was being typed on. And he gave it a full-size keyboard. The display was a glowing orange color. “When it was closed, it looked powerful—almost monolithic. When you opened it, it was friendly and inviting and personal,” says Moggridge.

Tensions arose between Moggridge and the engineers. Moggridge wanted to create a perfect visual and tactile design, while the engineers were most concerned about their ability to cool the machine and manufacture it rather than the way it looked. For instance, there were latches that held the display in place when the lid was open. Moggridge wanted people to be able to find and use them easily, so he designed release buttons with a complex grip texture made up of lots of miniature pyramids. The engineers thought they were too complex and would be too expensive to engineer and tool. They called them “paranoids,” rather than pyramids, because they felt Moggridge was overly concerned about making the latches easily releasable.

Moggridge and Ellenby had differences of opinion as well. Ellenby wanted to make the machines look and feel a bit like a piece of military equipment. (In fact, ultimately, many of them were bought by the military, and several were used on NASA’s space shuttles.) Potential customers had told him they wanted serious-looking machines. Investment bankers wanted portables they could take into a deal negotiation meeting that would impress the people on the other side of the table. And they needed devices that were tough enough to be slung into the overhead compartments on airliners. Moggridge favored a more friendly appearance. Not surprisingly, since Ellenby was the boss, he got his way, mostly. “I’m still not comfortable with mobile products that look too jovial,” Ellenby explains. “I don’t think we want to have computers looking like Hello Kitty”—the cheerful Japanese fictional character that’s printed on children’s clothing.

One of the most compelling features of the Compass for many people was the orange glowing display screen. Since computers didn’t have flat-panel displays at the time, GRiD had to find a company that could supply them. They settled on Sharp, the Japanese electronics company, which was a pioneer in display technology, and convinced Sharp’s executives that there would be a large market for such displays. Sharp invested in tooling up a factory production line. It was a risky bet, but it gave Sharp a quick start in the laptop display business, where it has been a dominant player ever since. Everything was brand new then. The basic infrastructure of the PC industry had not been built yet. GRiD was the first American company to visit Sharp and most of the other Japanese electronics companies of the day. So it pioneered what was to become an essential element of the PC industry: U.S. computer companies would build their machines from Japanese components. At the time, few Japanese spoke English and visa versa, so the meetings were long and awkward.

This was an exciting time for the Japanese, as well. A year after their contract was signed, the general manager of Sharp’s engineering center asked to be allowed to visit the building where GRiD got its start. It was night and had been raining when Ellenby showed the executive the tiny cottage behind his home in Palo Alto. Raindrops were still dripping from the giant spreading oak tree as they walked through the yard. The Sharp executive didn’t speak much English, and he said little at the time, but, later, one of his lieutenants passed along a message to Ellenby from his boss. “He told his staff to tell me, ‘I visited your idea and the rain dropped on my head even though it had stopped raining,” Ellenby recalls. “This was his haiku. The raindrop reminded him of the freshness of our idea. He now had more faith in his decision to build the product line.”

Though the display and some of the other features of the Compass worked as anticipated, the design had one major flaw: the preproduction machine was way too hot. Ellenby had set a deadline for shipping the computer in early 1982 and didn’t want to delay it by six months so they could complete design of a custom-made chip that would create less heat. Edens disagreed with him. He thought the machines might be hot enough to actually burn the fingers of people who touched them. But Ellenby overruled him, relying on the magnesium case to keep the heat at a manageable level. GRiD shipped a few of the machines on schedule. Years later, Ellenby talked to some U.S. soldiers who told him the early machines were so hot that once, when they were deployed in the desert, they tried to fry eggs on one.

The Compass was launched on April 5, 1982, one year after the Osborne 1 had hit the market. It had all sorts of innovative features, including e-mail, an operating system that allowed the computer to run more than one program at once, and the ability to connect via telephone lines to a server computer called GRiD Central where documents and data could be stored. In this way, it anticipated the computer networks that became ubiquitous in the 1990s. (In an April 5, 1982 story in BusinessWeek, Apple’s Steve Jobs criticized the GRiD Compass for being connected to a network: “I don’t want to be a link in a chain, but something very independent.” This was one case where the maestro just didn’t get it. He didn’t understand how important networking would eventually be for personal computing.) On the negative side, the early display measured only six inches diagonally and it cost a whopping $8,100. In spite of those factors, there was a lot of excitement when the Compass became available. It was much smaller and lighter than the luggables that were being produced at the time.

Shortly after the Compass was released, Ellenby sat on a panel at a tech industry conference in Monterey, California, with Adam Osborne. Ellenby recalls that Osborne proudly hefted his Osborne 1 onto the table in front of him and nearly vanished behind it. Ellenby already had his Compass sitting on the table. With the lid down, it was hidden from the audience by his name plaque. He lifted the plaque to show them how slim his machine was. “They were stunned. They all rushed up to look at it,” Ellenby recalls. “I can remember Adam’s face. He was shocked.”

For Ellenby and his crew, it was a time to celebrate. They had designed a remarkable machine. On Fridays the company would sponsor official beer bashes, and, often, they’d decamp to a tavern in Mountain View, the St. James Infirmary, where a 25-foot-tall, well-endowed wooden statue of Wonder Woman stood behind and above the bar. After innumerable beers, recalls Ted Clark, later an executive at Hewlett-Packard, who worked at GRiD at the time, GRiD-ites would try to climb up Wonder Woman and put GRiD bumper stickers on some of her more pronounced body parts.

Impressive as the Compass was, though, sales were disappointing. The price doubtlessly had a negative impact, but its biggest problem was IBM. Big Blue had introduced its personal computer in August 1981. Since IBM was the dominant maker of computers at the time, its PC had instant credibility with businesspeople. Many corporations had standardized on its technology, and they wanted PCs that ran the same operating from Microsoft that powered the IBM PC. The Compass had its own operating system. It was vastly superior to Microsoft’s DOS, in part because of its networking ability. But, no matter. GRiD’s software wasn’t the industry standard.

Ellenby had had an inkling that something big was coming from IBM long before it arrived. Robert Noyce, one of GRiD’s angel investors, was running Intel’s research and development (R&D) at the time. IBM had chosen Intel’s 8088 microprocessor as the brains of its PC. Noyce wasn’t allowed to disclose proprietary information about the IBM machine, but still he guided Ellenby away from using Motorola’s then-popular 68000 processor, suggesting Intel processors instead. GRiD chose the Intel 8086 chip. Since the Compass had a chip in the same family as IBM’s, it was easier to create a version of DOS for use in its machine. Microsoft’s Bill Gates agreed to prepare a special version of DOS for GRiD. Still, it wasn’t until 1985 that GRiD came out with a fully PC-compatible version of its machine that ran off-the-shelf versions of DOS.

It wasn’t enough to make the company a success. The market for an $8,000 notebook computer was fairly small. Compaq had introduced its luggable in November 1982, which was a functional clone of the IBM PC, and GRiD couldn’t compete with its price of $2,995. Ellenby tried to make a deal with Toshiba that would have handed off manufacturing to the Japanese company and reduced costs significantly, but his board wouldn’t go along with it. At the time, almost all computer companies did their own manufacturing. GRiD was gradually running out of money. It was acquired by Tandy, then the world’s largest PC company, in 1988. Tandy’s computer business was later bought by PC maker AST, which was still later bought by Korea’s Samsung.

Years later, looking back , Ellenby is proud of what GRiD had accomplished. The company produced the first true laptop computer, and other companies paid royalties for using GRiD’s patented folding design for years. It created a network computer a decade and a half before that idea caught fire in the post–Internet world. The Compass was the first computer to come with an integrated package of software applications, including word processor and spreadsheet. And Ellenby conceived of the idea of renting software to computer users, a concept that became mainstream 25 years later.

After he left GRiD, Ellenby continued to innovate. One night in 1991, when he and his son, Thomas, were sailing off the coast of Mexico, they made a discovery that led ultimately, years later, to the commercial launch of another invention. Thomas was coming off watch and told his father that there was a ship off in the distance he should keep an eye on. Up on deck, Ellenby found that he couldn’t pick out the lights of the ship in the darkness. Thomas mused that it would be handy if somebody built a pair of binoculars with a global positioning system in them. You could place an electronic marker on an object you saw through the binoculars, and then anybody else who used the binoculars would be directed to the same object. Later, Ellenby’s other son, Peter, got involved. Ultimately, after much thinking and experimenting, the three came up with a set of technologies and a service that essentially turned the physical world into an interface to information about it. (Think of the world as a giant computer screen, and the cellphone as a computer mouse.) If the technology is embedded in a mobile device, and the user points the device at a building that has been preprogrammed into a database, he or she will see information about the building on the screen. Microsoft’s Bill Gates had long talked about his vision of information at your fingertips via mobile computing. Thanks to the Ellenbys, that vision was realized, quite literally, when their company, GeoVector, launched their mobile pointing-based information service in Japan in 2007.

Moggridge’s journey alongside Ellenby launched him into a whole new way of thinking about computer design. He recalls that when he got his hands on an early working model of the Compass, he was proud of what he had done. Later, though, he took it home and played around with it, he realized that much of the experience of using a personal computer related to the software rather than the hardware design. He’d had nothing to do with that. Sitting in his house in Palo Alto, he had a revelation. “I was sucked down into the world of software,” he recalls. “I realized that if I was going to be serious about designing for people, I’d have to master software interactions, too.” Later, after Moggridge cofounded IDEO with the noted designer David Kelley, they made software an important part of their practice. In 2007, Moggridge published a book, Designing Interactions, which traces the development of user interface software design to a position of prominence in computers, mobile phones, and Web sites. So, Moggridge had learned what Kay seemed to understand from the beginning: that software would become just as important as the hardware as portable computers evolved and improved. That insight would shape many of the innovations to come.

Copyright Steve Hamm, 2008