This heretofore unpublished document essentially provided a blueprint for how Apple could save itself from long-term debilitation - a course that, had it been taken, would have put Apple into the driver's seat into the 1990s and possibly beyond.

What would possess Bill Gates to offer such a helping hand to a competitor? Surely, it couldn't be out of the goodness of his heart. And of course it wasn't. The idea for the memorandum actually originated with a young man named Jeff Raikes, who had joined Microsoft in late 1981 at the age of 23 to become one of the company's first product marketing managers. Raikes, a Nebraska farm boy with all-American good looks, had joined Apple out of Stanford in 1980 and eventually risen to engineering manager but jumped ship to Microsoft a year later after he decided - correctly, as it turned out - that software would become more important than the hardware business Apple was focused on. Steve Jobs, who had asked Raikes to join his new Macintosh team, was furious when Raikes announced he was leaving. "Jobs read to me from the riot act," remembers Raikes, who went on to become one of nine members of Microsoft's powerful Executive Committee. "He said, 'Microsoft will go out of business.'" Right, Steve.

At Microsoft, Gates, too, had GUI on the brain in the early 1980s. MS-DOS was becoming the workhorse of IBM-compatibles, but Gates knew that software applications would become much more compelling when they could be presented graphically, in a manner that users would intuitively understand. Most people, Gates had told attendees of the Rosen Research Personal Computer Forum at Lake Geneva, Wisconsin, in May 1981, "want things to be user-friendly. They want a way of understanding how information is represented in their terms. Drawers, files, folders - whatever terminology you pick, it's got to somehow tie into something the user had used before."

In those days, in fact, it wasn't just Steve Jobs and Bill Gates thinking GUI, it was practically the whole industry. Digital Research was working on GUI-based software called GEM. Apple was relying on it for the ill-fated Lisa project, which evolved into the Macintosh. And another company called VisiCorp, which had soared to industry stardom on the success of its VisiCalc spreadsheet, shocked the computer world - and Bill Gates - when it demonstrated, at the fall 1982 Comdex show in Las Vegas, the VisiOn, a graphical user interface for powerful IBM-compatible PCs. Gates, after seeing the VisiOn demo in VisiCorp's booth, called a Microsoft technologist named Charles Simonyi, back in Bellevue, and told him to fly down to see the technology.

Simonyi, a Hungarian refugee and computer programming whiz, was perhaps Gates's equal in energy and intensity. He had joined Microsoft in 1980, following years of work as one of the élite scientists developing futuristic computer products at Xerox PARC in Palo Alto, California. Simonyi knew all about GUIs. In fact, he had written a wordprocessing program for one at Xerox. "The minute Charles came on, we said, 'OK, it's in our future to do the graphical interface. The question is when,'" recalls Gates in an interview for this book in his suite on the 29th floor of the Las Vegas Hilton during the fall 1996 Comdex show in that city.

In 1981, at Simonyi's insistence, Microsoft procured the latest in GUI technology: a Xerox Star computer, which was the first commercial product to use the new technology but failed because of its prohibitive cost of $100,000, including a printer, a data-storage server, and a network to hook all the gear together. "We wanted people at Microsoft to understand the future," recalls Simonyi, seated on an ergonomically correct black chair in a room adorned with modern art originals in his mansion overlooking Seattle's Lake Washington. "Bill knew from the outset that GUI was the future."

Until he saw the VisiOn in action, Gates had been preoccupied with building software applications to run on top of MS-DOS and the proliferation of smaller computer platforms on the market. Since it was not yet clear which of the platforms would survive, Gates - who had been an avid poker player during his days at Harvard - covered his bets by supporting everything he could. "Understand, you make a lot more money selling applications than you do operating systems," Gates says, sipping from a can of Coke as he reclines in a chair with his back to a panorama of Las Vegas sprawling below. "Operating systems, you get a couple percent of the price of the machine [or $40 for a $2,000 PC]. Applications, you can get hundreds of dollars.

"Remember, back then," Gates adds, "we weren't talking about Microsoft being an $8 billion-a-year company. We were hoping we would be a $200 million [per year] company. Well, if you could get a few million Macs to sell a year, we would have been triple the size we were then." Microsoft's revenues in 1982 totaled just $25 million.

Upon their return from the 1982 Comdex, Gates and Simonyi set to work on a graphically based operating system called Interface Manager, which later was renamed Windows when the first version finally shipped in November 1985. Windows, as originally designed, was clearly inferior to the Macintosh system because it featured a "tiled" look as opposed to the Mac's use of overlapping windows. If you opened three wordprocessing documents on Windows, for example, they would appear as tiles that occupied equal amounts of space on the screen, obscuring much of the text from view. On the Mac, however, those same three documents overlapped one another, just as if they were lying on a desk. Each could be maneuvered to another part of the screen so that more of the document could be seen. The Mac, in short, mimicked the way people really worked, which is why it was so attractive.

As impressive as VisiOn had looked in Las Vegas, the operating system proved so clunky and filled with bugs when it finally shipped a year later that it never gained any momentum and eventually died a quiet death. Gates, however, began spreading the word of Windows everywhere, even as he quietly moved to hedge his bets after taking a sneak peek at Steve Jobs's Macintosh in 1981. At the time, Microsoft's biggest-selling application was a spreadsheet called Multiplan, which competed against VisiCalc. With Multiplan selling like hotcakes, Gates saw no reason why Microsoft should not support this new machine called the Macintosh, too.

And oh, what a machine it was. Fueled by a Motorola 68000 chip that was far more powerful than Intel's chips, the graphical displays fairly danced across the Mac's screen. After seeing a demonstration of the Mac in Cupertino during October 1981, "our view was that it was exactly what we were looking for," remembers Jeff Harbers, who at the time was a Multiplan manager. In January 1982, Microsoft signed an agreement to develop applications for the Mac. Initially, Gates committed to delivering three programs for the Mac's launch: a spreadsheet, a business graphics program, and a database.

By the next year, though, Microsoft's Macintosh development effort evolved from a "cover our bets" strategy to one of "bet the farm." The change in emphasis came after a start-up company called Lotus Development Corporation in Massachusetts began shipping a new spreadsheet for IBM-compatibles called 1-2-3. It was faster and more powerful than either VisiCalc or Microsoft's Multiplan, and it went on to become the killer application that solidified the IBM-compatible as the world's preeminent desktop computing standard. Lotus 1-2-3 was also proving a killer to Microsoft's Multiplan, as Bill Gates recognized with horror as sales of his spreadsheet began to wither under the assault.

On October 25, 1983, Gates convened with his top lieutenants in a retreat at a Red Lion Inn near Bellevue to ponder what to do about 1-2-3. In that retreat, Gates and his strategists came up with the concept of a new spreadsheet that featured a GUI and would later be named Excel. "In our euphoria over the Mac and in our awe at 1-2-3, we decided we needed to focus on GUI applications," Raikes recalls. Initially, the plan was to offer Excel first on IBM-compatible PCs. But as the Macintosh grabbed so much industry attention in its January 1984 launch, Gates shifted gears and decided to put Excel onto the Mac first. "We bet on the Macintosh, hoping Windows would come in sooner rather than later," Raikes remembers.

It was a big wager, indeed. Gates committed fully one-third of Microsoft's programming resources to the Macintosh, putting Jeff Harbers in charge of the project. "We were complete Mac fanatics," remembers Harbers, who along with other engineers would joke, "I'm going to the beach" whenever they wanted to go into a locked, windowless room at Microsoft where a Mac prototype, or "SAND," had been stowed. Adds Gates, "We were in it together. We bet a lot of the future on the Mac."

From the outset, though, the spirit of cooperation was not reciprocated by Apple. "Steve was convinced that Bill would take ideas from the Mac and incorporate them into Windows," recalls Mike Murray, who was then Apple's Macintosh marketing manager and later became Microsoft's vice president for human resources and administration. "Steve would call Bill and say, 'Get down here right now.' We would go into a room at Bandley 1 [on the Apple campus], and Bill would go to a whiteboard and sketch out everything Microsoft was doing. He'd say, 'I shouldn't tell you this, but I'm going to tell you everything I'm doing.'" Gates would sketch out his Windows path, hop onto a plane, and go home.

Jobs had good reason to be paranoid. After all, Gates was on his way to becoming king of the IBM PC and made no bones about the fact that he wanted to push Windows as the software standard in the Intel world. Yet, Harbers remembers, "we felt we owed it to Apple to keep the Mac secret. Only Gates, Simonyi, and the Mac development team knew about it." Murray, who was friends with Gates's top lieutenant and Harvard classmate, Steve Ballmer, remembers getting frantic calls from the Microsoft chieftain. "One day, Bill called me and said, 'Mike, what are we supposed to do? Steve keeps yelling at us. I don't know whether to work on the Mac or not,'" Murray recalls. "I'd say, 'Bill, just keep the pedal to the metal. We need you. I'll manage Steve.'"

Microsoft Excel for the Macintosh would not be ready to ship until September 1985, after it was announced the previous May at that New York City press conference at Tavern on the Green. After the Mac's launch in 1984, Gates watched with relief as computer enthusiasts gobbled up the new machines. But anxiety set in for himself and the others at Microsoft when the Mac's sales tapered off in late 1984 and into 1985. "I remember having a meeting with Ballmer and the [Microsoft] Mac team," Gates says. "We were all saying, 'Jesus, you know, Apple may not do this well.' And Ballmer said, 'Well, we can help them. But we have to assume they're staying awake at night worrying about these same things.'"

One day in the first quarter of 1985, one of Microsoft's product managers, Chris Larson, made an offhand comment to Jeff Raikes as they commiserated about the declining fortunes of the Macintosh. "He said they should license the Mac operating system," Raikes remembers. A lightbulb flashed in Raikes's head, and he hurried to his computer to put the idea down on paper. "So I wrote a letter to Bill saying I really think Apple should license its operating system," Raikes remembers. "I said, 'They are competing against all of the R&D on the IBM platform.' My conclusion was that Apple should license the Mac. I sent the memo to Bill in May 1985." Gates took the memo and expanded it to include a list of potential clone manufacturers Apple could call on for help. Gates was careful, in compiling this list, to include manufacturers that could broaden the Macintosh market, not just cannibalize sales from Apple. Canon, for instance, was strong in Japan, while Apple at that time was not.

Before sending the memo, Gates put in calls to the senior executives he knew at both AT&T and Hewlett-Packard. "We talked to them about 'Well, if Apple approached you, would you be interested?' And those were the top two on our list," Gates says. Those companies, in fact, were interested. "If Apple really thought licensing was a complicated thing somehow, we were glad, because we understood licensing, to help out," Gates adds. "But the letter was very clear that we're saying we're not trying to make money off of licensing. If it's necessary, we'll facilitate it by being a middleman."

The memo went out, and Gates and Raikes waited. And waited. But after several days, there was no response. "We didn't hear from John, so Bill called him," Raikes says. "And Sculley said, 'Well, how do you do this? Do we sell system boards to the OEMs [original equipment manufacturers]?' They just didn't understand." Gates and Raikes had not worked out details of how a Macintosh licensing plan would be carried out, given Apple's lack of interest. But it quite likely would have followed the usual Microsoft model of licensing software to manufacturers in return for royalties.

Nor did Apple want to. Apple was always a religious company, and the religion of Macintosh made the subject of licensing Apple's most contentious and divisive issue ever. Jean-Louis Gassée and his engineers rightfully believed the Mac represented a quantum leap in technology, a watershed product every bit as significant as Woz's Apple II and IBM's first PC. No way, no how would Gassée see his precious Mac turned over to a ragtag army of copycats. This was Apple's crown jewel, and Gassée meant to defend it with his life. He was the guardian of the castle, the keeper of the flame. Never mind that he tooled around town in a Mercedes with license plates that read "open mac." By that, Gassée was only illustrating his support for opening the Mac to hard drives, plug-in circuit boards, and other useful accessories that Jobs had ordered kept out of the original Mac. The open mac license plates, however, should have read closed mac, because that is really how Gassée felt about licensing Mac technology to the clone market.

"Apple was so committed to being different from the rest," recalls Kevin Sullivan, whose arrival in 1987 as head of Apple's human resources department would signal a new era in the company's management. "There was a glee, almost. We were elegant. Jean-Louis called it 'the beautiful business we are in.'"

The furor of the licensing debate manifested itself early on, during a meeting of Sculley's executive staff one day in 1985 after Gates's memo had been received. A young man named Dan Eilers would present his case for why Apple should license the Mac, a scant three years after he had joined the company upon graduating from nearby Stanford University with a degree in economics. Eilers, Apple's director of investor relations and then aged 30, might have been young and inexperienced, but he was a business pragmatist detached from the religious fervor of the engineers. He knew Apple was in bad financial shape and simply thought the licensing plan would help out. Sculley invited Eilers to brief the executive staff on his plan.

Eilers was about to run into a buzz saw.

Eilers was no wild-eyed radical, though from the reaction he would receive you would have thought he was. A slight, soft-spoken man who liked to fly private planes and take long walks in the woods, Eilers was understandably nervous about the meeting. Although he had a fancy title, he was really a peon in the organizational structure. And here he was, about to give his very first presentation to the big brass, the members of Sculley's inner circle, which besides Gassée, Campbell, Spindler, and Coleman included Jay Elliott, head of human resources; Al Eisenstat, general counsel; and Dave Barram, chief financial officer. The company was still in crisis mode, just weeks after Jobs had been forced out, and executive staff was meeting with Sculley every day at 7:30 a.m. sharp to keep tabs on cash and inventories.

Apple headquarters at the time was atop a four-story building called De Anza 4, an easy stroll from the De Anza 7 building, to which it would soon move. Eilers took the elevator to the fourth floor and walked into the 7:30 a.m. executive staff meeting commencing in a tiny conference room, appropriately called the "Small Room." Seated around a rectangular table about 10 feet in length were Sculley, Gassée, Elliott, Eisenstat, Barram, and Campbell. Most were dressed casually, as usual, in khakis, slacks, or jeans. Gassée, who typically looked like a biker in his trademark black leather jackets and black leather pants, glowered as the young man stood up to begin a two-hour presentation.

"Apple should recognize as a distinct advantage that its operating system is superior to DOS," Eilers said, according to people familiar with the meeting. "And the best way to make that a standard would be to put it on the Intel platform."

As Eilers displayed slide after slide on an overhead projector to support his argument, Gassée's face reddened and his eyes bulged. He was upset and could contain himself no more. In his thick French accent, he began yelling and screaming that licensing could not be done, according to those familiar with the episode. The scheme was flawed for two reasons, Gassée asserted: One, he did not believe it was even technically feasible for the Mac to run on anything other than an Apple computer, because it was so closely intertwined with Motorola's microchip. Furthermore, Gassée argued, opening the Mac to the outside world would give competitors a license to steal sales away from Apple itself.

Gassée did have a point. Rejiggering the Mac to run on an Intel machine would have been a tall order. The basic problem was that the Mac had always been designed to tie together the software of a machine with its hardware innards. Microsoft's MS-DOS, on the other hand, was designed primarily to tie into the Intel chip. Practically everything else needed for the computer, such as keyboards and disk drives, could be found at the nearest components junk lot and then plugged in to support the MS-DOS/Intel standard. But the Mac's software and hardware were virtually inseparable. Take away the software, and the distinctive look and feel are lost as well. Take away the hardware, and the Mac doesn't run as smoothly. That's why the Mac not only was far easier to use than a Microsoft-run computer but ran much more smoothly as well.

One possibility would have been to get manufacturers to use the Macintosh technology as is, just as Bill Gates had recommended. Another would have been just to slap the top layer of the Mac's operating system on top of, say, MS-DOS, and let users get a taste of the Mac. The machine would not run as seamlessly as a Mac, because the hardware was not tied in to the software as closely. But it would at least offer users the appealing look of the Mac, with its graphical icons. That was not only a possibility; it was already being done by Digital Research.

Digital Research had been founded by a software entrepreneur named Gary Kildall, whose operating system, CP/M (Control Program for Microcomputers), had been an early rival of MS-DOS. The success of MS-DOS eventually killed off CP/M. By 1985, another former Xerox PARCer named Lee Lorenzen had successfully copied the look of the Macintosh so that it could run on top of MS-DOS with Digital Research's GEM software. GEM was essentially designed to run a graphical user interface of any kind, no matter which. If Windows was taking off, Lorenzen could retool GEM to look like Windows. Since the Macintosh was the best of the GUI bunch, he designed it to look like the Mac - actually, to "look and feel" like a Mac. It looked like a Mac, down to containing the same trash can icon for discarding unwanted files. And it felt like a Mac, with the same ability to use the mouse to move objects around on the screen. Those three words, "look and feel," would become the focus of an industrywide debate over software copyright protection.

At the time, Lorenzen remembers, IBM was negotiating with Digital Research to license GEM for use on all its MS-DOS-based machines. That's when the Apple lawyers showed up at Digital's door in Pacific Grove, a village set in an idyllic setting of pine trees and crashing ocean waves on California's Monterey Peninsula. They pointed out, in no uncertain terms, that Digital was illegally copying Apple's technology. Digital had thought it was in the clear by borrowing just the Mac's look, not the actual technology. But software copyright law was still a very murky area, and IBM, for one, wanted no part of any litigation. "IBM was ready to acquire GEM, but Apple showed up and threatened suit," recalls Lorenzen, who is now CEO of a small software developer, Altura Software Inc. "IBM chickened out." And that was the end of GEM. Apple stopped this particular threat in its tracks, but it missed another opportunity. It could just as easily have acquired the GEM technology itself, to proliferate the Mac's look all over the place.

Years later, Apple would attempt what Digital had done, and the results would hold the potential for breathtaking implications across the whole industry.

Gassée was also correct in worrying that Apple would cannibalize its sales by opening itself to voracious competition from cloners, many of whom would consist of two guys and a screwdriver in a garage, who could seriously undercut the mothership on price. Indeed, an all-out licensing plan would have required a fundamental change in Apple's whole business model. Gassée's biggest fear was that the company would have to undergo wrenching layoffs, perhaps on the order of half the workforce. And he had good reason to be afraid.

Apple was pulling in about $2 billion a year in revenues by selling roughly 700,000 computers at $3,000 each. If Apple were to license its Mac software to all 4 million of the Intel computers being sold per year at a premium rate of about $100 each, that would bring in about $400 million in sales. Assuming that sales of Apple computers would fall by half, as Gassée feared, Apple would shrink to a $1.4 billion company almost overnight. Bill Gates believes, however, that Apple could have structured its licensing in such a way as to protect itself. "They wouldn't have had to open it wide open," Gates says. "Let's just say they licensed HP, or just AT&T or somebody in Europe, you know, like Olivetti, or somebody in Japan, like Sony or whoever. It would have made all the difference. Momentum creates momentum. If you have volume, then people write apps. If people write apps, you get momentum."

In any case, no one ever said it would be easy. The end reward, as Eilers, Gates, and the other licensing advocates all argued, was the creation of a standard that, in the end, would provide more profit to Apple than to anybody else because it held the keys to a kingdom. Both Microsoft and Intel proved that theory true. With combined revenues only slightly higher than Apple's by the mid-1990s, Microsoft and Intel became so profitable that together they would account for fully half the entire PC industry's profits - an amazing feat in a $100 billion industry with thousands of competitors.

The great minds in the Small Room, however, were lost in small thought that day, far more concerned with the here and now than anything that might happen down the road. Gassée had done most of the arguing against Eilers's plan while the other executives sat and listened. It was clear Eilers had little support, because no one in the room, not even Sculley, rose to defend him. So when his presentation was over, Eilers simply scooped up his slides and papers and saw himself out. The licensing plan had died by lack of an endorsement. But this was not the end of the debate, not by a long shot.

It wasn't that Sculley didn't recognize the merits of licensing. After all, it was he who had put Eilers in charge of strategic investments to explore the possibilities of Apple forming various alliances. At Pepsi, he had learned the importance of forming strategic partnerships with outside companies, such as the legion of Pepsi bottlers. And he could certainly appreciate the significance of market share, since careers at Pepsi had been made or broken on as little as a tenth of a percentage point fluctuation between it and Coca-Cola. In point of fact, Sculley came to realize, the hostility directed at Eilers was really aimed at him.

"Dan was incredibly unpopular in engineering because they knew he was my agent," Sculley says. "Every time Dan would come in with an outside idea, not only would the idea be shot down, but he would be lucky to get out alive."

Sculley was having Eilers play around with some other wild ideas, such as Apple buying another company. Sculley had a gleam in his eye, especially, for Silicon Graphics and Sun Microsystems, small but thriving makers of big computer workstations, as well as Novell, a small company that was then pioneering a new form of software to link networks of computers together. Sculley saw great strategic opportunities in each of these companies, since all were focused on the big corporate market he wanted to crack. "But the engineers felt Apple didn't need anyone else," Sculley says. "Just because you had the title of anything did not mean they would do what you asked."

This attitude, which had originated in the Steve Jobs days, came to be known in Silicon Valley circles as NIH, or "not invented here." If it wasn't invented at Apple, the smartest place in the universe, Apple's engineers wanted no part of it.

Unbeknown to many people in the company at the time, Sculley had also set up a strategic sales group to study, among other things, the possibility of putting the Mac's look and feel - the top layer of the software, which the user sees - on top of other computers, just as Digital Research had done. This was a less radical step than Eilers's plan to license the Mac technology with all its bells and whistles to clone manufacturers. Allowing others to use just the "look and feel" was more like an outpatient alternative to open-heart surgery on the Mac. It would not be necessary to ditch Motorola altogether. Using the Trojan horse approach, the Mac interface could be sneaked into corporations on another company's computer. Once the workers saw for themselves how great it was, they would refuse to use anything else. Named to head that venture was Chuck Berger, an outdoorsman who loved to water-ski on northern California lakes. He and Eilers were kindred spirits and would become equally despised at Apple.

Berger, vice president of Apple's new strategic sales group, had been given the green light by Sculley to talk to as many manufacturers as he could about this particular scheme. Over a 12-month period beginning in 1985, Berger and Sculley's former technical aide, Mike Homer, who was named to assist Berger, crisscrossed the United States, drumming up outside interest in the plan. There was more than enough to keep them hopping. Dr. An Wang, the founder of Wang Laboratories outside Boston, wanted to put the Mac software on top of his company's wordprocessing machines. Digital Equipment, Wang's neighbor down the Massachusetts Turnpike in Maynard, planned to incorporate the Mac's look into a line of new desktop computers. AT&T was so interested in putting the Mac onto the company's Unix workstations that approvals had been made all the way up to Bob Allen, then AT&T's CEO. Silicon Graphics, which would go on to fame as creator of the digital special effects in 1990s movie blockbusters such as Jurassic Park, was also profoundly interested.

"All of these had either a handshake agreement or letters of intent," says an industry executive intimately familiar with the discussions. "John and Chuck flew to AT&T twice and had them in the bag."

Sculley, however, would go on to reject all deals on the table. Gassée was yelling and screaming again, and Sculley just could not bear to hear it. Like Eilers, Berger was left to fall on his own sword. As of late 1985, Sculley was on track with a plan that would inflate Apple's profit margins above 50 percent on forthcoming sales of the souped-up Mac Plus. In a series of executive staff meetings at which Berger presented his case for licensing, Gassée railed against going through with anything that would rob those profits.

"He made a strong stand that it was stupid to give up 55 percent margins for what would be, at best, 45 percent margins," says an executive close to all the discussions. "Jean-Louis said there would not be enough money left to fund the 'insanely great' technology and that the engineers would probably leave." Berger argued that it was clear that closed, or "proprietary," standards did not work. The best example of that, he said, was Sony's failure in the early '80s to set a standard in the videocassette recording industry with its Betamax machine. While Betamax was widely regarded as technically superior to the rival VHS machines, VHS was an open standard that other manufacturers could copy. Since Betamax was not, VHS went on to take over the VCR market.

Berger, in one of the meetings, also said, "Eventually someone will catch up with the [Mac's] GUI." Rolling his eyes in disgust, Gassée snapped back, "No one will ever catch up to the GUI." Gassée could not have been more blind if he had had a blindfold on.

Gassée may have been the most outspoken person at Apple against licensing, but he was certainly not alone. Indeed, when looking back on it all, Sculley says he is not so sure the board itself would have backed any kind of licensing scheme, even if he had pursued it full tilt. "Remember, at the time the board was interested in one thing: gross margin," he said to me in our first of several discussions for this book, sipping from a cup of black coffee as he mused on the situation a decade later in his lawyer's office in Palo Alto, California. By gross margin, Sculley was referring to the gross profit margin, measured as a percentage of sales, which serves as a key barometer of a manufacturer's profitability. "The engineers wanted innovation. You had to fuel the innovation and manage the profits. So you had to stay within this envelope."

Others in the industry empathize with Sculley's situation, given the time and circumstances. "There was no question they should have licensed the software. It was leadership technology in the marketplace," says retired IBM president Jack Kuehler. "[But] it would have required an unusual person to do that early on. And you would never know if it was the right thing to do, because detractors would shoot you down. If it had not worked out according to plan, it probably would have cost Sculley his job."

Even Gassée, shockingly enough, now admits he was flat-out wrong. "I am aware that I am known as the Great Satan on licensing," he says. "My mistake was, I got into a debate that I should not have gotten into. I thought, financially, it didn't make sense. I was never for or against licensing. I just did not see how it would make sense. But my approach was stupid. We were just fat cats living off a business that had no competition."

Just as Gassée was vowing to Berger that no one would overtake Apple's lead, Gates was hard at work on Windows 1.0, the prototype of a successor to MS-DOS that would grow to envelop the planet. Gates badly wanted it to have the same look as the Mac and already planned to include some Mac-like features in the graphics, including Mac-like control panels and Mac-like pulldown menus. Actually, Gates was also influenced in this approach by the graphical user interface work at Xerox PARC, as well as other early implementations of the technology such as VisiOn. But it was the Mac that became the first commercially successful version of this concept and the one he most wanted to emulate.

This incensed Sculley, who began contemplating a lawsuit. One day in the fall of 1985, an Apple lawyer named Jack Brown showed up on Microsoft's doorstep. It was a scene reminiscent of the one at Digital Research a few months earlier. Only this time Apple was not dealing with a pushover. "He [Brown] came in and made incredible threats about patents, copyrights, and trade secrets," Gates remembers, indignation still rising in his voice years later. "And he said he is a lawyer who has never lost a trade secret lawsuit. And we said, 'But Apple's being very careful not to give us any of their trade secrets.' Everything Apple gave us, Apple was being very careful about because Apple knew exactly what we were doing. We didn't need a license at all, in any way, and that is very clear."

Gates was hopping mad. He had not stolen anything from Apple, he insisted then and continues to insist now. The whole idea of GUIs had originated not with Apple, he points out, but with Xerox. "The father of the Mac is Xerox. The father of Windows is Xerox," Gates says. Charles Simonyi, Microsoft's in-house GUI maestro, compares the similarities between Windows and the Macintosh to those found in different automobile models. "When you decide to build an automobile, you're not going to change the steering wheel," Simonyi says. "They all have common ancestry. This was such a silly and pointless argument that they were falling into."

After Jack Brown's threats, Gates and Bill Neukom, Microsoft's chief counsel, arranged to fly down to Cupertino to meet with Sculley and his top legal gun, Al Eisenstat. In a phone conversation beforehand, Gates, according to Sculley, put a gun to his head. "If we're on a collision course, I want to know it because we'll stop all development on Mac products," Gates told Sculley. "I hope we can find a way to settle this thing. The Mac is important to us and to our sales." Gates denies ever making that threat, calling Sculley's statement "the most unfair characterization of anything I've ever heard."

Physically, Gates was hardly an imposing figure. Tall and thin, his hair was often tousled in those days, and, with his big glasses, he appeared to be little more than a teenager. But when it came to business, Gates was a Muhammad Ali, the dude you didn't want to mess with. And he most certainly would have had the gumption to carry out his threat, if in fact he made it. It was true that he needed Apple, but Apple needed him a lot more. At that time, Gates was rolling out Microsoft Excel, the spreadsheet program that would significantly increase the appeal of the Mac to business customers. The original Mac had no numeric keys to run a spreadsheet, much less the memory to do so. The forthcoming Mac Plus would. Along with other programs, including Basic and Multiplan, Gates controlled roughly two-thirds of all the software then available on the Mac. This was no guy to shove around.

Before Gates arrived for the meeting in the De Anza 4 boardroom on October 24, 1985, Sculley's executive staff pleaded with him not to cave in. But, mindful of Gates's power, Sculley was convinced that a war between Apple and Microsoft would seriously disrupt the company's resurgent momentum, stripping the Mac of its most important software ally at a critical juncture. In their meeting in Apple's boardroom, where Gates remembers enough sushi being brought in "for 50 people," he and Sculley haggled.

"I went to Sculley, and I said, 'We don't need a license. Steve and I talked explicitly about us doing graphical applications. You've seen Windows every step of the way,'" Gates remembers. "Sculley said, 'I understand what you're saying, but isn't there some concession you can make to us?' I said, 'OK, we'll do Excel first on the Mac and have a period of exclusivity.' And Sculley said, 'Well, what's going to happen if you don't perform on that?' And I said, 'Why don't you give us a license so this dispute doesn't come up again?'"

So Sculley instructed Eisenstat's legal team to draw up a contract allowing Microsoft to license the look and feel of the Mac - or "visual displays," as they were referred to in legal terms - but only in Windows 1.0. Gates and Neukom, however, refused to sign that agreement, believing Microsoft had the right to use those visual displays in its Macintosh applications as well as other present and future products. Neukom drafted a revised three-page contract and sent it to Eisenstat on November 14. Eisenstat shrugged and made a few minor changes, sending it on to Gates and Sculley, who signed it on November 22. With this contract, Apple thus agreed that Gates was free to come up with his own take on the graphical technology that had originated with Xerox, from which Microsoft had already obtained a license to certain GUI technology. "We were buying peace with Microsoft," Eisenstat recalls.

But Gates got an unexpected bonanza, after he successfully pushed Eisenstat to modify that agreement, in a way that would grant Microsoft a de facto license to copy the Mac at will. A phrase in the three-page contract, written by Microsoft, granted to Microsoft "a non-exclusive, worldwide, royalty-free, perpetual, nontransferable license to use these derivative works in present and future software programs, and to license them to and through third parties for use in their software programs." By agreeing to include the phrase "in present and future software programs," Apple had unwittingly given Gates carte blanche to use virtually any visual features borrowed from the Mac in Windows 1.0 and all future versions. The courts, in fact, in a case that would carry mammoth implications for the entire computer industry, would later interpret the phrase "in present and future software programs" to mean all Windows versions deriving from the one at issue in this agreement.

Sculley and Eisenstat had just given away the store. "If I knew then what I know now," says Eisenstat, sighing deeply as he recounts the blunder a decade later, "I would have said, 'Don't do it.'" Gassée and the other members of the executive staff had not wanted Sculley to give in to Gates in any way, much less by granting a license.

Gates himself downplays the significance of the licensing agreement, saying he had the right to pursue Windows anyway. What could have derailed Windows, Gates believes, is if Apple had followed the advice from him and Jeff Raikes to license the Mac widely. If the Mac had become the standard, Gates says, "We would have sold less Windows. But the key thing is, we could have sold more applications."

The matter of Apple cloning its beloved Macintosh had been shelved for now. Sculley was too busy tending to details of the turnaround to pay the subject more than passing attention. And as the first few years went by, it appeared that keeping the Mac in-house wasn't so dumb, after all. The industry slump of 1985 had decimated many PC manufacturers and crimped IBM's performance as well. Gassée and his engineers would point merrily to IBM and say, "Thank God we didn't listen to Eilers and Berger." IBM, however, was not exactly a textbook case on how to license. It had given up control of its computer when it had turned over the operating system to Microsoft and the microprocessor to Intel. The fortunes of Apple, in any event, had never looked brighter. So the timing could hardly have been worse when Berger presented his next case for licensing in 1987.

This time, Berger, whose strategic sales group had been renamed "business development," steered away from using the "L" word. He rejoined with his old colleague Mike Homer to explore the possibility of putting the Mac's software onto just a workstation, a much bigger version of the PC that was used heavily in data-intensive jobs such as engineering. Sculley saw great merits in putting the Mac onto a workstation. One of Apple's problems in landing big corporate accounts was that its computers did not "scale," or run the spectrum from bare-bones secretaries' desktops all the way up to a high-performance workstation that a rocket scientist could use. A big reason IBM's computers had become so popular in corporate America was that its computers did scale: all needed equipment could be bought from the same company, simplifying employee training and technical support. Apple had the desktops covered, all right, but it didn't have diddly for anything more powerful.

That's where a Mac workstation would enter the picture: If Apple were to put its software onto workstations manufactured by other companies, it might not be an Apple computer these corporations would be buying, but it would walk and talk like a Mac. That would make it a lot easier for a chief technology officer to recommend outfitting the whole corporation with Macintoshes, since the workers could all be trained the same. "John thought it would be a great marriage between us and a low-end workstation," says an executive familiar with the situation. Unlike the previous broadscale licensing effort, though, Sculley wanted this deal limited to just one outside company. Berger's natural inclination was to approach Sun Microsystems first, since that company was around the corner in Mountain View, California, and in the past had negotiated with Apple in a buyout deal.

Sun, which specialized in workstations, was run by an amateur hockey player named Scott McNealy. With his boyish face and mouthful of large teeth, McNealy resembled a large chipmunk when he smiled. Like most Silicon Valley executives, he hardly ever wore a suit. Even at big meetings with industry analysts, he would trot onstage in a uniform of pullover shirt and faded jeans. But McNealy was a tough, voracious competitor like Gates, a veritable pit bull whose company was starting to give fits to big rivals in the workstation market such as Hewlett-Packard and Digital. Scott McNealy was an up-and-comer, and, sure, he wouldn't mind getting into bed with Apple if he could come out ahead.

But McNealy's ego got in the way. He was insistent, for one thing, that any Mac/Sun computer use his company's new chip, called SPARC. This was a major stumbling block, because Apple had already decided to remain joined at the hip with Motorola. Sculley and McNealy also discussed, again, the possibility of Apple buying Sun and combining Sun's workstation line with Apple's Macintosh line. McNealy, though, insisted he be named president and chief operating officer of the combined companies, according to an individual involved with those talks. That would necessitate the demotion of Del Yocam, however, and Sculley wasn't yet in a mood to emasculate the guy who had stood up with him in the confrontation with Jobs.

Things were getting nowhere, so Berger and Homer packed their bags and headed east in the spring of 1986 to Boston, where they found a friendlier audience in the executive suite of Apollo Computer, a rival workstation maker based in suburban Chelmsford that, in fact, was still leading the market for workstations at the time. One reason Apollo was receptive was that its business was heading south, even as Sun's was on the way up. Apollo's executives were also concerned about the potential threat of personal computers becoming powerful enough to encroach into the low end of the workstation market, and Sun was also moving into that market fast. Apollo had just introduced the low-end DN 3000 for $10,000 in a bid to shore up that part of the market from attack when Apple came calling.

"We wanted to go down to five thousand dollars, but to get that far down we needed a cheaper architecture and a cheaper operating system," recalls Cheryl Vedoe, who was then manager of Apollo's low-end marketing. And that was where Apple could help. The two companies actually had a lot in common. Apollo, with its Domain operating system, was considered technically superior to its competition, just as Apple was with its Mac. And both Apollo and Apple maintained proprietary, or closed, systems that were not licensed to others.

Sun, however, had an open system that it licensed to others. In fact, Sun's very openness was the secret of its success. Founded in 1982 by Stanford graduate students McNealy, Andy Bechtolscheim, and Vinod Khosa, along with a man named Bill Joy from the University of California at Berkeley, Sun virtually gave away its software while letting other companies help manufacture its SPARC chip. This unleashed a licensing fury that overwhelmed workstation competitors such as Apollo, with their closed systems. McNealy, who ran Sun, gained a reputation as a butt kicker, because he kicked the butt not only of the competition but of employees to keep costs down while producing more.

So, in April 1986, Apple and Apollo sat down to talk. There were Chuck Berger and Mike Homer on one side of the table, representing Apple. On the other side were Cheryl Vedoe and Ed Zander, Apollo's vice president of marketing. At the time, Apple was developing the Macintosh II, which was going to be twice as fast as and far more powerful than the Macintosh Plus. At a retail price of up to $5,500, it was a high-end machine for Apple but ideal, Vedoe and Zander thought, to put onto the low end of the Apollo line. The idea, according to Vedoe and Homer, was for Apollo to buy the Macintosh II and repackage it as an Apollo workstation, with minor adjustments such as melding it with Apollo's Domain operating system. Over the next 10 months, Apple and Apollo worked together to make prototypes of Apollo Macs. In all, remembers Homer, Apollo planned to buy 40,000 Macintosh IIs over a two-year period for the initial stage of the cloning project.

All was going well, or so thought nearly everyone associated with the hush-hush project. Near the end, the Apple bigwigs got in on the act as Chief Operating Officer Del Yocam accompanied Sculley on a trip to Apollo. "When we were in Boston, we toured the facility and went into the lab to see the Mac OS running on Apollo boxes," Yocam recalls. "I remember leaving there thinking, 'Everything is positive.'"

In January 1987, Apollo's chief financial officer, Roland Pampel, signed an agreement to license the Mac IIs from Apple. Back in Cupertino, Berger and Homer put together a presentation for the executive staff at which they fully expected Sculley to give the deal the thumbs-up. "Apollo was there. It had licensed the Mac," recalls Homer, who is now senior vice president of marketing for Netscape Communications. "That would have been the first of many such deals."

The contract was in Sculley's hands. Excited about the new opportunity that lay ahead, Vedoe and an Apollo attorney went to Boston's Logan International Airport to catch a flight out west to finalize the deal. It would be a glorious day, or so they thought. That same morning, in Cupertino, Sculley was presiding over his regular executive staff meeting. With the contract in front of him and Berger and Homer waiting anxiously for a signature, Sculley pushed it to one side and shook his head. "He said he had decided not to do the deal because Apollo was a declining star and Sun was a rising star," says an executive intimately familiar with the meeting. When they got over their shock, Berger and Homer looked at each other and dashed out of the conference room to the phones. They had Vedoe paged at the Boston airport and stopped her and the attorney from boarding just as their plane was about to take off.

"We were in the boarding area when Mike Homer paged us," Vedoe recalls. "He said, 'Don't get on the airplane. We need to talk about it.' It did come as a shock." Actually, there was nothing to talk about, except that the whole deal was finis.

The ironic thing was Gassée, the Great Satan of licensing, had nothing to do with killing this deal. This was just Sculley, acting on a whim. He ended up like the dog that lost its bone by trying to snatch the reflection it saw in the water. After that, Sun expressed no more interest than before, quashing any hope of a deal there. Apple was again alone, just as it had always been. For poor Apollo, meanwhile, this was the beginning of the end. Having devoted much of its time and resources to getting out the new Apollo Mac, the company was left so far behind in the development of its own Apollo systems that it could not catch up when Sun moved in with its new low-end workstations. Apollo hemorrhaged more market share and was soon acquired by Hewlett-Packard. Both Vedoe and Zander packed up and joined Sun.

OK, forget licensing, Berger thought. With IBM-compatible computers proliferating like rabbits throughout the corporate world, could Apple at least do something to make the Mac fit in with all the others? As early as 1985, in the Macintosh office, the Macs had been rigged so that they could communicate with one another over the AppleTalk phone lines. But if you were a Mac person in an IBM office, you were on your own. "Everyplace we went, people were telling us, 'We love the Mac, but we can't even consider them unless you can hook into the networks of IBM computers,'" Berger later confided to a colleague. So Berger came up with a plan to help remedy that: put some extra software into the Mac so it could act like MS-DOS software and tap into the whole IBM network.

It was now almost 1988, and there was a clear trend in corporations to decentralize away from the huge, $1 million mainframe computers that had dominated the workplace in the past. In their place, corporations were deploying fleets of cheaper, sub-$5,000 desktop computers that could all be interconnected via smaller in-house networks. These new networks gave much more flexibility to employees and their managers, allowing files to be swapped and changed at will. The network was the way of the future, and it was time for Apple to get on board.

With Sculley's blessing, Berger got outside contractors to design and make these so-called IBM emulators. When an operating system emulates something, it looks and acts like another system. It doesn't run as fast as the system it is emulating, though, because it has to take extra time to translate that system's code. Companies such as Du Pont and Aetna expressed great interest in the Apple equipment and even started putting in large orders. The corporate appeal: office employees would be able to work on an easier-to-use Mac while still being hooked in to the rest of the IBM network. Gassée, however, didn't like this plan one bit. Just as he had dismissed the importance of licensing, Gassée never saw the need for Apple's computers to communicate with anything except other Apple computers. This philosophy had been demonstrated early on, in 1985, when he had addressed an Apple sales meeting in Hawaii and someone had had the audacity to ask what Apple's strategy would be in communicating with IBM-compatibles.

"Gassée stood up in front of the audience and held up a cutoff piece of a telephone wire and said, 'This is our communications strategy,'" recalls John Ziel, a sales manager in the Portland district who was in the auditorium. In other words, if you need to communicate with your IBM coworker, dial him or her through your computer modem. This was a terrible strategy, of course, because it did not seamlessly link Macintoshes with IBM-compatibles. "We looked at Gassée and said, 'Who is this guy?'"

Gassée elaborated on this novel concept in an encounter that same year with Peter Hirshberg, who was in charge of a fledgling program to develop networking and communications products. Hirshberg's group had already committed to a bunch of so-called connectivity products for corporate customers when Gassée sent word down that he was canceling them. Stunned, Hirshberg requested an explanation, and Gassée invited him to meet for lunch at Vivi's, a popular falafel eatery in Cupertino.

Over their falafels, Gassée repeated his telephone strategy. "He said, 'You must convince your customers that all they need is a simple telephone line,'" Hirshberg recalls. When asked, pray tell, how do you do that? Gassée smiled slyly and leaned forward to say, "Public relations. You must use public relations, not advertising." Whenever Gassée wanted to drive home a point, he would slip into a French accent even more pronounced than usual. This is when he explained to Hirshberg the difference between public relations and advertising.

"With advertising, I, Jean-Louis, say, 'I am the most wonderful lover in the world.' Of course, this would not work in attracting the woman." By now, Hirshberg had forgotten the networking argument and listened in fascination. "But if two of the most beautiful women in the world say they spent the evening with Jean-Louis, this would work. This is the difference between advertising and public relations."

Three years later, though, Gassée was in no mood to use sexual analogies in attacking Berger's latest scheme. Headquarters had by now moved to De Anza 7, and he was still telling Sculley and everyone else who would listen that putting Mac/IBM machines into corporations would be tantamount to "another Vietnam War." "He said Apple could never win in this battle because IBM would keep changing the rules," says an executive who heard the discussions. Berger would counter, "We can never sell a computer in big business unless we can hook into a business computer."

The tensions boiled over at a third-floor conference room in De Anza 7 during early 1988. Before approximately 100 people in the room, including Sculley and the entire executive staff, Gassée and Berger almost came to blows. "Jean-Louis was trying to make everyone believe we should return the engineers to engineering," recalls an executive in that meeting. "Chuck began screaming at Gassée. At one point, he jumped up out of his chair and slammed his notebook on the table so hard it broke the back of it." As usual, Gassée prevailed. The IBM emulators were tossed out. By this point, Berger had had his fill of Gassée and Apple. He quit soon after to sign up with Sun and the McNealy butt-kickers.

Jean-Louis Gassée had won nearly every fight. He was the undisputed master of engineering, the person who had almost always gotten his way. Now he would put another indelible stamp on Apple, one that would have repercussions as grave as the decision not to license. He wanted to keep those profit margins up - way up - and if that meant sacrificing market share, so be it. Sculley and the board could hardly have cared less at this point, since profits, revenues, and the stock price were all soaring to new heights. The gravy train was still chugging happily along, and John Sculley was still the apple of everyone's eye.