The Innovators
They made an odd but powerful partnership: Woz was an angelic naïf who looked like a panda, Jobs a demon-driven mesmerizer who looked like a whippet. Gates had bullied Allen into giving him more than half of their partnership. In the case of Apple, it was Wozniak’s father, an engineer who respected engineers and disdained marketers and managers, who insisted that his son, who had made the designs, be given more than 50 percent of the partnership. He confronted Jobs when he came by the Wozniak house: “You don’t deserve shit. You haven’t produced anything.” Jobs began to cry and told Steve Wozniak that he was willing to call off the partnership. “If we’re not 50-50,” Jobs said, “you can have the whole thing.” Wozniak, however, understood what Jobs contributed to their partnership, and it was worth at least 50 percent. If he had been on his own, Wozniak might not have progressed beyond handing out free schematics.
After they demonstrated the computer at a Homebrew meeting, Jobs was approached by Paul Terrell, the owner of a small chain of computer stores called The Byte Shop. After they talked, Terrell said, “Keep in touch,” handing Jobs his card. The next day Jobs walked into his store barefoot and announced, “I’m keeping in touch.” By the time Jobs had finished his pitch, Terrell had agreed to order fifty of what became known as the Apple I computer. But he wanted them fully assembled, not just printed boards with a pile of components. It was another step in the evolution of personal computers. They would not be just for solder-gun-wielding hobbyists anymore.
Jobs understood this trend. When it came time to build the Apple II, he did not spend much time studying microprocessor specs. Instead he went to Macy’s at the Stanford mall and studied the Cuisinart. He decided that the next personal computer should be like an appliance: all fit together with a sleek case and no assembly required. From the power supply to the software, from the keyboard to the monitor, everything should be tightly integrated. “My vision was to create the first fully packaged computer,” he explained. “We were no longer aiming for the handful of hobbyists who liked to assemble their own computers, who knew how to buy transformers and keyboards. For every one of them there were a thousand people who would want the machine to be ready to run.”
By early 1977 a few other hobbyist computer companies had bubbled up from the Homebrew and other such cauldrons. Lee Felsenstein, the club’s master of ceremonies, had launched Processor Technology and come out with a computer called Sol. Other companies included Cromemco, Vector Graphic, Southwest Technical Products, Commodore, and IMSAI. But the Apple II was the first personal computer to be simple and fully integrated, from the hardware to the software. It went on sale in June 1977 for $1,298, and within three years 100,000 of them were sold.
The rise of Apple marked a decline of hobbyist culture. For decades young innovators such as Kilby and Noyce had been introduced to electronics by knowing how to distinguish different transistors, resistors, capacitors, and diodes, then wire-wrapping or soldering them into breadboards to create circuits that became ham radios, rocket controllers, amplifiers, and oscilloscopes. But in 1971 microprocessors began making complex circuit boards obsolete, and Japanese electronics companies began mass-producing products that were cheaper than homemade ones. Sales of do-it-yourself kits withered away. Hardware hackers such as Wozniak ceded primacy to software coders such as Gates. With the Apple II and then, more notably, the Macintosh in 1984, Apple pioneered the practice of creating machines that users were not supposed to open and fiddle with their innards.
The Apple II also established a doctrine that would become a religious creed for Steve Jobs: his company’s hardware was tightly integrated with its operating system software. He was a perfectionist who liked to control the user experience end to end. He didn’t want to let you buy an Apple machine and run someone else’s clunky operating system on it, nor buy Apple’s operating system and put it on someone else’s junky hardware.
That integrated model did not become standard practice. The launch of the Apple II woke up the big computer companies, most notably IBM, and prompted an alternative to emerge. IBM—more specifically IBM as it was outmaneuvered by Bill Gates—would embrace an approach in which the personal computer’s hardware and its operating system were made by different companies. As a result, software would become king, and, except at Apple, most computer hardware would become a commodity.
DAN BRICKLIN AND VISICALC
For personal computers to be useful, and for practical people to justify buying them, they had to become tools rather than merely toys. Even the Apple II might have been a passing fad, once the excitement of the hobbyists receded, if users had not been able to apply it to a practical task. Thus there arose a demand for what became known as application software, programs that could apply a personal computer’s processing power to a specific chore.
The most influential pioneer in that field was Dan Bricklin, who conceived the first financial spreadsheet program, VisiCalc.83 Bricklin was an MIT computer science graduate who spent a few years developing word-processing software at Digital Equipment Corporation and then enrolled at Harvard Business School. Sitting in a lecture one day in the spring of 1978, he watched as the professor created the columns and rows for a financial model on the blackboard. When he found an error or wanted to modify a value in one cell, the professor had to use his eraser and change the values in many of the other cells.84
Bricklin had seen Doug Engelbart demonstrate his oNLine System, made famous at the Mother of All Demos, which featured a graphical display and a mouse for pointing and clicking. Bricklin began envisioning an electronic spreadsheet that would use a mouse and simple point-drag-and-click interface. That summer, while riding a bike on Martha’s Vineyard, he decided to turn the idea into a product. He was well suited for such an endeavor. He was a software engineer with the instincts of a product person; he had a feel for what users would want. His parents were entrepreneurs, and he was excited by the prospect of starting a business. And he was a good team player, who knew how to find the right partners. “I had the right combination of experience and knowledge to develop software that met a need people had,” he observed.85
So he teamed up with a friend he had met at MIT, Bob Frankston, another software engineer whose father was an entrepreneur. “The ability for Dan and me to work as a team was crucial,” Frankston said. Although Bricklin could have written the program alone, instead he sketched it out and had Frankston develop it. “It gave him the freedom to focus on what the program should do rather than how to do it,” Frankston said of their collaboration.86
The first decision they made was to develop the program for use on a personal computer rather than on a DEC business computer. They chose the Apple II because Wozniak had made its architecture open and transparent enough that the functions needed by software developers were easily accessible.
They created the prototype over a weekend on an Apple II they borrowed from someone who would, in effect, become a third collaborator, Dan Fylstra. A recent graduate of Harvard Business School, Fylstra had launched a software publishing company, which focused on games such as chess, that he ran out of his Cambridge apartment. In order for a software industry to develop separately from the hardware industry, it was necessary to have publishers who knew how to promote and distribute products.
Because both Bricklin and Frankston had good business sense and a feel for consumer desires, they focused on making VisiCalc a product, not just a program. They used friends and professors as focus groups to make sure the interface was intuitive and easy to use. “The goal was to give the user a conceptual model that was unsurprising,” Frankston explained. “It was called the principle of least surprise. We were illusionists synthesizing an experience.”87
Among those who helped turn VisiCalc into a business phenomenon was Ben Rosen, then an analyst with Morgan Stanley who later turned his influential newsletter and conferences into a business of his own and then started a venture capital firm in Manhattan. In May 1979 Fylstra demonstrated an early version of VisiCalc at Ro
sen’s Personal Computer Forum in his hometown of New Orleans. In his newsletter, Rosen enthused, “VisiCalc comes alive visually. . . . In minutes, people who have never used a computer are writing and using programs.” He ended with a prediction that came true: “VisiCalc could someday become the software tail that wags (and sells) the personal computer dog.”
VisiCalc catapulted the Apple II to triumph, because for a year there were no versions for other personal computers. “That’s what really drove the Apple II to the success it achieved,” Jobs later said.88 It was quickly followed by word-processing software, such as Apple Writer and EasyWriter. Thus did VisiCalc not only stimulate the market for personal computers, but it helped to create an entire new profit-driven industry, that of publishing proprietary application software.
THE IBM OPERATING SYSTEM
During the 1970s IBM dominated the mainframe market with its 360 series. But it was beaten by DEC and Wang in the market for refrigerator-size minicomputers, and it looked like it might be left behind in personal computers as well. “IBM bringing out a personal computer would be like teaching an elephant to tap dance,” one expert declared.89
The company’s top management seemed to agree. So they considered instead just licensing the Atari 800 home computer and slapping IBM’s name on it. But when that option was debated at a July 1980 meeting, IBM’s CEO Frank Carey dismissed it. Surely the world’s greatest computer company could create a personal computer of its own, he said. Doing anything new at the company, he complained, seemed to require three hundred people working three years.
That is when Bill Lowe, who was the director of IBM’s development lab in Boca Raton, Florida, piped up. “No, sir, you’re wrong,” he stated. “We can get a project out in a year.”90 His cockiness got him assigned the task of overseeing the project, code-named Acorn, to create an IBM personal computer.
Lowe’s new team was led by Don Estridge, who chose Jack Sams, a gentle southerner who was a twenty-year veteran of IBM, to be in charge of piecing together the software. Given the one-year deadline, Sams knew that he would have to license software from outside vendors rather than have it written in-house. So on July 21, 1980, he placed a call to Bill Gates and asked to see him right away. When Gates invited him to fly to Seattle the following week, Sams replied that he was already heading for the airport and wanted to see Gates the next day. Sensing a big fish hungry to be hooked, Gates was thrilled.
A few weeks earlier, Gates had recruited his Harvard dorm mate Steve Ballmer to Microsoft as the business manager, and he asked Ballmer to join him at the IBM meeting. “You’re the only other guy here who can wear a suit,” Gates pointed out.91 When Sams arrived, Gates was also wearing a suit, but he did not quite fill it. “This young fellow came out to take us back, and I thought he was the office boy,” recalled Sams, who was dressed in the IBM standard blue suit and white shirt. But he and the rest of his team were soon dazzled by Gates’s brilliance.
At first the IBM folks wanted to talk about licensing Microsoft BASIC, but Gates turned the conversation into an intense discussion about where technology was heading. By the end of a few hours, they were talking about licensing all of the programming languages Microsoft had or could produce, including Fortran and COBOL in addition to BASIC. “We told IBM, ‘Okay, you can have everything we make,’ even though we hadn’t even made it yet,” Gates recalled.92
The IBM team returned a few weeks later. There was one essential piece of software, in addition to these programming languages, that IBM was missing. It needed an operating system, the software program that would serve as the foundation for all of the other programs. An operating system handles the basic instructions that other software uses, including such chores as deciding where data should be stored, how memory and processing resources should be allocated, and how applications software interacts with the computer’s hardware.
Microsoft did not yet make an operating system. It was instead working with one called CP/M (for Control Program for Microcomputers) that was owned by Gary Kildall, a childhood friend of Gates who had recently moved to Monterey, California. So with Sams sitting in his office, Gates picked up the phone and called Kildall. “I’m sending some guys down,” he said, describing what the IBM executives were seeking. “Treat them right, they’re important guys.”93
Kildall didn’t. Gates later referred to it as “the day Gary decided to go flying.” Instead of meeting the IBM visitors, Kildall chose to pilot his private plane, as he loved to do, and keep a previously scheduled appointment in San Francisco. He left it to his wife to meet with the four dark-suited men of the IBM team in the quirky Victorian house that served as Kildall’s company headquarters. When they presented her with a long nondisclosure agreement, she refused to sign it. After much haggling, the IBM folks walked out in disgust. “We popped out our letter that said please don’t tell anybody we’re here, and we don’t want to hear anything confidential, and she read it and said I can’t sign this,” Sams recalled. “We spent the whole day in Pacific Grove debating with them and with our attorneys and her attorneys and everybody else about whether or not she could even talk to us about talking to us, and then we left.” Kildall’s little company had just blown its chance to become the dominant player in computer software.94
Sams flew back to Seattle to see Gates and asked him to figure out another way to conjure up an operating system. Fortunately, Paul Allen knew someone in Seattle who could help: Tim Paterson, who worked for a small firm called Seattle Computer Products. A few months earlier Paterson had become frustrated that Kildall’s CP/M was not available for Intel’s newest microprocessors, so he adapted it into an operating system that he dubbed QDOS, for Quick and Dirty Operating System.95
By then Gates had come to the realization that one operating system, most likely the one chosen by IBM, would end up being the standard operating system that most personal computers would use. He also figured out that whoever owned that operating system would be in the catbird seat. So instead of sending the IBM folks to see Paterson, Gates and his team said that they would handle things on their own. Ballmer later recalled, “We just told IBM, ‘Look, we’ll go and get this operating system from this small local company, we’ll take care of it, we’ll fix it up.’ ”
Paterson’s firm was struggling to make ends meet, so Allen was able to negotiate a savvy deal with his friend. After initially acquiring just a nonexclusive license, Allen went back when an IBM deal looked likely and bought Paterson’s software outright, without telling him why. “We ended up working out a deal to buy the operating system from him, for whatever usage we wanted, for fifty thousand dollars,” Allen recalled.96 For that pittance Microsoft acquired the software that, after they spruced it up, would allow it to dominate the software industry for more than three decades.
But Gates almost balked. He was uncharacteristically worried that Microsoft, which was wildly overcommitted with other projects, might not have the capacity to gussy up QDOS into an IBM-worthy operating system. Microsoft had only forty ragtag employees, some of whom slept on the floor and took sponge baths in the morning, and it was led by a twenty-four-year-old who could still be mistaken for an office boy. On a Sunday at the end of September 1980, two months after IBM first came calling, Gates gathered his top team to make the go or no-go decision. It was Kay Nishi, a young computer entrepreneur from Japan with a Gatesian intensity, who was the most adamant. “Gotta do it! Gotta do it!” he squealed repeatedly as he bounced around the room. Gates decided he was right.97
Gates and Ballmer took an overnight flight to Boca Raton to negotiate the deal. Their 1980 revenues were $7.5 million, compared to IBM’s $30 billion, but Gates was gunning for an agreement that would allow Microsoft to keep ownership of an operating system that IBM would turn into a global standard. In its deal with Paterson’s company, Microsoft had bought DOS outright, “for whatever usage,” rather than merely licensing it. That was smart, but what was even smarter was not letting IBM force Microsoft to make the same ar
rangement.
When they landed at the Miami airport, they went to a bathroom to change into suits, and Gates realized he had forgotten a tie. In an unusual display of fastidiousness, he insisted that they stop at a Burdine’s department store on the drive to Boca in order to buy one. It did not have the full desired effect on the crisp-suited IBM executives waiting to greet him. One of the software engineers recalled that Gates looked like a “kid that had chased somebody around the block and stolen a suit off him and the suit was way too big for him. His collar stuck up and he looked like some punk kid, and I said, ‘Who the hell is this?’ ”98
Once Gates began his presentation, however, they quit focusing on his disheveled appearance. He wowed the IBM team with his mastery of details, both technical and legal, and projected calm confidence when insisting on terms. It was largely an act. When he arrived back in Seattle, Gates went into his office, lay on the floor, and agonized aloud to Ballmer about all of his doubts.
After a month of back-and-forth, a thirty-two-page deal was struck in early November 1980. “Steve and I knew that contract by heart,” Gates declared.99 “We didn’t get paid that much. The total was something like $186,000.” At least initially. But it had the two provisions that Gates knew would alter the balance of power in the computer industry. The first was that IBM’s license to use the operating system, which it would call PC-DOS, would be nonexclusive. Gates could license the same operating system to other personal computer makers under the name MS-DOS. Second, Microsoft would keep control of the source code. This meant that IBM couldn’t modify or evolve the software into something that became proprietary to its machines. Only Microsoft could make changes, and then it could license each new version to any company it wanted. “We knew there were going to be clones of the IBM PC,” Gates said. “We structured that original contract to allow them. It was a key point in our negotiations.”100