The untold story of the Microsoft antitrust case and what it means for the future of Bill Gates and his company.

I. THE HUMBLING

The judge in Chicago wanted his signature. Just two little words on the bottom line: Bill. Gates.

This article has been reproduced in a new format and may be missing content or contain faulty links. Contact wiredlabs@wired.com to report an issue.

It was early last March, three full months after the formal mediation between Microsoft and the Justice Department had started, and Gates knew he didn't have much time. Any day now, Judge Thomas Penfield Jackson would unveil his verdict in United States v. Microsoft, one of the largest antitrust actions in American history. Nobody doubted what the outcome would be: In November, Jackson had disgorged a 207-page "findings of fact" that was searing in its tone and staggering in the totality of its rejection of Microsoft's version of events. If the verdict fit the findings, it was going to be ugly – maybe ugly enough to bring about the company's dismemberment.

Microsoft's last hope for averting disaster lay in the hands of a different judge – the judge in Chicago, Richard Posner. Posner, the chief justice of the United States Court of Appeals for the Seventh Circuit, was a conservative jurist with a towering reputation as an antitrust scholar. Soon after Judge Jackson issued his findings of fact, he asked Judge Posner to step in as a mediator. For anyone else, trying to forge a peace between these combatants would have been a fool's errand. But given Posner's stature, Jackson hoped – prayed – he might just pull it off.

Every week since the end of November, Posner had summoned a team of lawyers from the DOJ and from Microsoft to his chambers in Chicago. Meeting with each side separately, he had, in effect, "retried the case," said one participant – rehearsing the arguments, reviewing the evidence. Gates himself had flown out to meet Posner and spent hours on the phone with him afterward, delving into the details of Microsoft's business. "The guy's super-smart," Gates told me later, bestowing on Posner his highest plaudit. By February, the judge had begun producing drafts of a proposed consent decree, which would put certain limits on Microsoft's conduct. After presenting each draft to the opposing parties, Posner solicited their comments and criticisms, then cranked out another draft to push the ball forward. For a month or so, it went on like this, back and forth, to and fro – until they arrived at Draft 14. With Draft 14, Posner seemed to think he'd come close to crafting a settlement Microsoft would accept, but which restricted its behavior in significant ways. So in order to show the DOJ that Microsoft was serious, Posner asked Gates to put his name to the proposal.

There were those at Microsoft who thought Posner naive. The government would never be satisfied, even if the company were to sacrifice its firstborn – or, something rather more precious, its source code. Others simply thought Draft 14 too draconian. But although Gates could see the skeptics' points, he was anxious to put this whole nightmare behind him. He swallowed hard and scrawled his signature.

The skeptics were right: It wasn't enough. Yet Posner still believed that a settlement could happen. For another month, he kept churning out drafts – Draft 15, Draft 16, Draft 17. In the last week in March, Posner asked Jackson for 10 days more; he was near, very near, to securing a deal. (So certain was Judge Jackson that the case would be settled, he took off on vacation to San Francisco.) By March 29, Draft 18 was complete. It reflected the DOJ's final offer.

In Gates' office in Redmond, the chairman's inner circle convened for one of the company's most fateful debates. Throughout the mediation, Gates had relied on this handful of people: Microsoft's newly elevated CEO, Steve Ballmer; its general counsel, Bill Neukom; and the senior executives Paul Maritz, Jim Allchin, and Bob Muglia. The document before them required that Microsoft set a uniform price list for Windows; prohibited it from striking exclusive contracts with Internet service and content providers; forced it to open its application programming interfaces. And although Draft 18 would let Microsoft add new features to Windows – features like Web browsing, which had provoked this lawsuit in the first place – PC makers would have the right to demand versions of the operating system without those features; and they would also be able to license the Windows source code, so they could modify the desktop, integrate rival software, or add features of their choosing.

There were critics who would say this was all trivial tinkering, modest stuff of marginal utility. But the Microsoft high command didn't see it that way. Even for those among them who'd been asked by Gates to be devil's advocates in favor of settlement, Draft 18 was a bridge too far. It was not a proposal that Gates could sign on to.

In Silicon Valley and in Washington, DC, Gates' decision to reject Draft 18 was seen as the latest blunder in his three-year battle with the federal government. It was an act of bloody-mindedness, of myopia, of hubris. But when I visited Gates recently in Redmond and asked him about his refusal to settle, he displayed not the slightest hint of doubt. The son of a lawyer, steeped in the language of contracts, Gates knew a bad deal when he saw one; and this was a deal that would have wrecked his business. Gates was aware that the courts were imperfect, and he considered Judge Jackson's more imperfect than most. But Gates had "faith," he told me, "that in the final analysis, the judicial system will come up with absolutely the right answer."

Whatever the logic of Gates' gamble, its immediate effect was swift and irrevocable. On March 31, Microsoft sent material to Posner that would form the basis of Draft 19, which he then read by phone to the DOJ. The very next afternoon, April 1, still four days shy of his self-imposed deadline, the mediator declared his mediation a failure.

In public, and even more pointedly in private, Microsoft blamed the breakdown on the coalition of state attorneys general who were the DOJ's partners in prosecution. In the crazy final days of the negotiations, the states had sent Posner a set of their own demands – demands considerably in excess of the DOJ's. In his only public statement on the talks, Posner was ambiguous about the precise cause of their collapse. Citing only "differences among the parties," he praised the professionalism of Microsoft and the DOJ, but made no mention of the attorneys general. However, in an early working version of Posner's statement that few people ever saw, the judge went out of his way to rap the states on the knuckles for their left-field intercession – while at the same time making clear that the truly insurmountable gap was "between" Microsoft and the DOJ.

With the mediation kaput, Jackson hustled back from the coast and delivered his verdict on April 3. It was nearly as gruesome as everyone expected. A month later, the DOJ and the states asked that the court split Microsoft in two. A month after that, Jackson agreed, ordering exactly the breakup the government requested.

It was the spring in Redmond when illusions were shattered, when old verities crumbled and the stock price tumbled, when everything that was solid melted into the air. By the time Jackson handed down his breakup order, Microsoft's Nasdaq value had been chopped nearly in half since March – wiping out more than $200 billion in wealth. Competitors crowed. The press piled on. Private class-action antitrust lawyers began to swarm. In the middle of June, Microsoft announced with great fanfare its grand new Internet strategy, and an industry that had so long hung on its every hiccup, that had trembled at the sound of its virtual footsteps, dismissed the initiative as half-cooked vaporware – or, more charitably, yawned. Three months later, in mid-September, a yearlong exodus of top executives reached its peak when Paul Maritz announced he was leaving the company. Even for the truest of true believers, faith had become a scarce commodity.

Gary Reback was known as a guy who got paid to complain about Bill Gates – the rough Silicon Valley equivalent of drawing a salary for breathing.

The humbling of Microsoft is the last great business story of the 20th century and the first great riddle of the 21st. There are fancier ways of putting it, but the riddle is this: How did it happen?

Perhaps no corporation in history has ever risen so far so fast. Having celebrated its 25th anniversary this summer, Microsoft is no longer a babe in the woods. Yet among the totemic firms of the past century, from Standard Oil and US Steel to General Motors and General Electric, none attained such stature, power, or profitability in such a breathtakingly short span of time. Even in the computer industry, where the awareness of Microsoft's ascent is acute, people often forget just how quickly it happened. As recently as 1992 or 1993, the company, though plenty influential, was hardly seen as an omnipotent leviathan. Five years later, that had changed. In the autumn of 1997, when the Justice Department first took after Microsoft in a serious way, Gates' many rivals in Silicon Valley applauded. But their pleasure was tempered by a perception that Microsoft was so indomitable, and the government so "dynamically anticlueful," as one digital quip-merchant put it, that nothing much would come of the DOJ's pursuit.

Theories abound as to why things turned out so spectacularly otherwise. Some now maintain that it was more or less inevitable; that Microsoft's business practices, once brought to light, would be enough to convict it in any court in the land. As a DOJ lawyer once said to me, "It was the stuff they did before the case was even filed that sealed their fate." Others suggest that Microsoft's history was bound to catch up with it in other respects; that its enemies in the Valley were lying in wait, ready to strike at the first opportunity. (Not surprisingly, this is a theory that Gates seems to favor.) Still others dwell on the tactical errors – on the ineptitude of Microsoft's lawyers and its halting incompetence in the realm of high politics. And still others focus on Gates himself; on his arrogance, and the insularity and isolation of the culture he'd wrought.

There are kernels of truth in each of these theories, but even taken together they fall short of eureka. What they fail to capture is the sometimes random confluence of forces: the way people with disparate agendas and mixed motivations came crashing together to produce an outcome that now seems obvious.

All through its conduct, the Microsoft trial was compared to a war. "The War of the Roses," said Judge Jackson, or "the fall of the House of Tudor. Something medieval." But war is hell not just because it's so bloody. War is hell because it's so unpredictable, so chaotic, so hot and dusty and shot through with confusion. The Microsoft trial was a war that neither side actually wanted to fight, in which unexpected alliances arose and old enmities surfaced at the most inopportune moments. It was war in which one hand rarely knew what the other was doing and carefully planned offensives ganged aft agley. Coincidence, timing, and blind shithouse luck all played their parts. And so did large acts of cowardice and small acts of courage, often committed by unknown soldiers.

This is the story of the generals in that war, of Bill Gates and Bill Neukom, Joel Klein and David Boies. But it is also the story of the unknown soldiers – people you've never heard of, whose stories have never been told. It's the story of Susan Creighton, the sweet-tempered antitrust lawyer who was Netscape's secret weapon. It's the story of Mark Tobey, the Texas crusader who took up the case when the Feds were still sleeping. It's the story of Mike Hirshland, the Republican Senate aide who found in Microsoft an unlikely passion, and it's the story of Dan Rubinfeld, the economist whose theories pushed the DOJ where it didn't want to go. It's the story of Steve McGeady, the Intel apostate who took the stand against Gates. And it's the story of Mike Morris, the lawyer from Sun Microsystems who mounted a lobbying campaign that brought together some of Microsoft's most powerful opponents, and that was one of the Valley's most closely guarded secrets – until now.

Acting sometimes in concert and sometimes alone, these anonymous characters and countless others like them accomplished things that had once seemed impossible. They thrust Silicon Valley neck-deep into the swamps of Washington, DC. They put the high tech industry's dirty laundry on display for all to see. They made antitrust law into national news. And they felled a giant that had once seemed invincible.

This is the story of the end of an era – and also of more than one kind of innocence.

II. THE CASE THAT ALMOST WASN'T

Though no one at the company knew it at the time, Microsoft's troubles with the Department of Justice began in earnest in the spring of 1996, with the literary aspirations of two amateur authors in Silicon Valley. Since 1990, when the Federal Trade Commission opened the first government probe into its practices, Microsoft had been under the antitrust microscope more or less constantly; not a year had passed without it receiving at least one civil investigative demand (CID) for documents. As one federal inquiry morphed into the next, Gates and Ballmer gradually came to see the investigations not merely as legal scrutiny but as a kind of proxy warfare (and, later, as nothing less than a vast high tech conspiracy) instigated by their rivals in the Valley and elsewhere. Yet as suspicious as they were about the source of their regulatory entanglements, Microsoft's leaders could not have dreamed that so much damage would be unleashed by a quiet woman who called herself a "law-and-order Republican," a shrill man who was regarded by some as mildly unhinged, and the book they wrote together – a book that was never published in any form, and whose contents, but for this story, would still be shrouded in secrecy.

Susan Creighton and Gary Reback were not, however, your typical wannabe wordsmiths. They were lawyers and antitrust specialists with the Valley's preeminent law firm, Wilson Sonsini Goodrich & Rosati. They were passionate, smart, articulate, and angry. They had been retained by Netscape to tell the world, not to mention the DOJ, about the myriad ways in which Microsoft was endeavoring to drive the pioneering startup six feet under. And they were rapidly approaching the end of their rope.

It was Reback who served as the duo's frontman. Throughout the computer business and the government, he was known as a guy who got paid to complain about Gates – the rough Silicon Valley equivalent of drawing a salary for breathing. Over the years, he had amassed a client roster that included some of the industry's most prominent firms – from Apple and Sun to Borland and Novell, though not all of them admitted it – and had earned a reputation as Redmond's most relentless and strident critic. (The cover of Wired 5.08 declared him "Bill Gates' Worst Nightmare.")

In Reback, Microsoft faced an adversary with a rare combination of technical savvy and antitrust expertise. As an undergrad at Yale, he had worked his way through school programming computers for the economics department; as a law student at Stanford, he had studied antitrust under the late William Baxter, who, as head of the DOJ's antitrust division under Ronald Reagan, would oversee the breakup of AT&T. Now in his late forties, Reback wore sharp suits, wire-rimmed glasses, and a perpetually pained expression. When he talked about Microsoft – which was pretty much constantly – his demeanor was fretfulness punctuated with blind outrage. His voice teetered on the edge of whine. "The only thing J. D. Rockefeller did that Bill Gates hasn't done," Reback would wail, "is use dynamite against his competitors!" Crusader and showboat, egotist and quote machine, he had a taste for avant-garde economic theories and a tendency to level extravagant accusations without much hard proof to back them up. He was, in the strictest sense, a zealot: a man both fanatical and fanatically earnest in his beliefs. Later, when the DOJ decided to go after Microsoft, a government lawyer was assigned to "deal" with Reback. "His heart's in the right place," this lawyer told me. "But he's twisted. He leaves me these voicemails in the middle of the night, raving about all kinds of stuff. He really needs some help." History might well have judged Reback a marginal figure, just another Gates-hating ranter, were it not for one inconvenient fact: Almost everything he claimed turned out to be true.

Reback's history with Microsoft was long, tangled, and not without its ironies. In the early 1980s, he secured for Apple the copyright registration for the Macintosh graphical user interface, a copyright that would eventually be at the center of a protracted lawsuit with Microsoft. Not long afterward, a bearded, elfin entrepreneur from Berkeley appeared on Reback's doorstep and asked for help in selling his fledgling software company. The company was called Dynamical Systems Research; the entrepreneur, Nathan Myhrvold. After Apple passed up the deal, Microsoft stepped in, buying Myhrvold's firm and Myhrvold along with it. Forever after, Reback was convinced that this transaction was pivotal to the rise of Windows, a conclusion that filled him with no end of guilt.

From then on, Reback became an anti-Microsoft missionary. As first the FTC and then the DOJ looked into the company, he peppered the Feds with briefs alleging a litany of predatory sins. In July 1994, the DOJ sued Microsoft for violating the Sherman Antitrust Act – but then dropped the suit after entering into a consent decree with the company. The consent decree contained only a few mild curbs; Gates himself summarized its effect bluntly: "nothing." At the behest of a clutch of Microsoft's rivals in the Valley, who saw the decree as a Potemkin remedy, Reback spearheaded a spirited, but ultimately futile, campaign in federal court to scuttle it.

Indeed, all of Reback's warnings went unheeded, with one exception. That fall, after Microsoft announced its plan to take over the financial-software firm Intuit for $1.5 billion, Reback, working primarily on behalf of an anonymous client (it was, in fact, the database company Sybase), prepared a white paper on the deal for the DOJ. Replete with novel economic concepts such as "network effects" and "increasing returns," the paper argued that if the merger weren't stopped, Microsoft would come to rule online financial services as it had the PC desktop. Reback was warned by the DOJ's chief economist that his analysis might be rejected as "totally preposterous." But it wasn't. In April 1995, the government moved to block the deal, and, rather than wage a costly battle, Microsoft bailed.

It was two months later, on June 21, that Reback received a call from Jim Clark, the chair of one of his firm's newest clients, Netscape. Earlier that day, Clark said, a team of Microsoft executives had visited Netscape's headquarters, met with its CEO, Jim Barksdale, its technical wunderkind, Marc Andreessen, and its marketing chief, Mike Homer, and offered them a "special relationship." If Netscape would abandon much of the browser market to Microsoft; if it would agree not to compete with Microsoft in other areas; if it would let Microsoft invest in Netscape and have a seat on its board, everything between the two companies would be wine and roses. If not …

"They basically said, OK, we have this nice shit sandwich for you," Mike Homer told me later. "You can put a little mustard on it if you want. You can put a little ketchup on it. But you're going to eat the fucking thing or we're going to put you out of business."

The next day, Reback phoned Joel Klein, the former deputy White House counsel who had recently been named the second-ranking lawyer in the antitrust division, and persuaded him to send Netscape a CID for some detailed notes Andreessen had taken during the meeting. A few weeks later, Reback flew to Washington with Clark, Andreessen, and Homer to state their case in person. The DOJ lawyers listened politely, jotted a few things down, said thanks – and then promptly forgot about it.

Thus began a pattern that would repeat itself again and again over the next two years. By the following spring, Barksdale & Co. were hearing a stream of reports about Microsoft's efforts to "cut off Netscape's air supply" – a phrase that would later acquire talismanic status – not least that Microsoft had threatened to cancel Compaq Computer's Windows license when Compaq tried to replace Internet Explorer with Netscape Navigator on some of its machines. With the browser war turning vicious and Netscape's complaints to the government getting nowhere, Reback and the company's general counsel, Roberta Katz, decided that desperate measures were in order. They would put Netscape's story down on paper, find a publisher, and present their plight in the bookstores of America.

Reback and Creighton's white paper read less like a legal treatise than a true-crime potboiler, a high tech Executioner's Song.

Penning this opus would fall to Susan Creighton. Cerebral and literary where Reback was blustery and verbal, Creighton was a Harvard- and Stanford-educated attorney who had clerked on the Supreme Court for Justice Sandra Day O'Connor. On May 1, Creighton sat down at her desk at home, surrounded by reams of documents, her infant child perched on her lap, and began tapping away.

Three months later, Creighton emerged with a 222-page piece of anti-Microsoft agitprop (with charts and tables courtesy of her husband, a local professor and desktop-publishing enthusiast). The tome would eventually bear the dust-dry title "White Paper Regarding Recent Anticompetitive Conduct of Microsoft Corporation," but it read less like a legal treatise than a true-crime potboiler, a high tech Executioner's Song. Creighton spun the tale of Microsoft's 20-year rise to power; of how it had employed a blend of strategic brilliance and nefarious tactics to destroy its competitors and hence "to acquire virtually complete control over what is arguably the most important tool in the American workplace"; and of how, faced with a potent new challenger, it had "engaged in a variety of anticompetitive acts that surpass its previous illegal conduct." The white paper accused Gates and his lieutenants of first attempting to divide the browser market with Netscape, and then, when that failed, of using their muscle with Internet service providers and OEMs – original equipment manufacturers, as PC makers are known – to shut down Netscape's distribution channels. Creighton accused them of illegally tying their browser to Windows. And of predatory pricing. And of exclusive dealing. And even of offering "secret side payments potentially amounting to hundreds of millions of dollars" to distributors to keep Netscape software off their customers' desktops.

Even more incendiary was Creighton's hypothesis as to Microsoft's motives. With help from Reback and Garth Saloner, a leading-edge Stanford economist who had assisted in drafting the Intuit white paper, Creighton put forward a nuanced theory of "monopoly maintenance": that Microsoft's primary objective was not to dominate the browser market for its own sake, but rather to protect its dominance over operating systems. What Gates realized, Creighton argued, was that the browser was more than just another software application – it was potentially a rival platform that held out the possibility of turning Windows into a commodity, and, as Gates himself put it, an "all but irrelevant" commodity at that.

"This is, at bottom, a very simple case," the white paper concluded. "It is about a monopolist (Microsoft) that has maintained its monopoly (desktop operating systems) for more than ten years. That monopoly is threatened by the introduction of a new technology (Web software) that is a partial substitute – and, in time, could become a complete substitute – for the monopoly product. Before that can happen, the monopolist decides to eliminate its principal rival (Netscape), and thereby protect its continued ability to receive monopoly rents. The monopolist is aided by the fact that circumstances are ideal for its predatory strategy: The monopolist has vast resources, while its rival has very modest ones; barriers to entry are high; and, once the rival is out of the way, the monopolist's road ahead looks clear."

When Creighton and Reback delivered the white paper to Netscape, the reaction was curiously schizophrenic. On one hand, Creighton remembers, "Barksdale and the others said to us, 'Thank you! Someone has finally put into words what we've been trying to say; it's like we've found our voice.'" Yet the white paper made chillingly clear just how dire Netscape's situation was. "As people saw what their position looked like in black and white, there was increasing concern about making it public," Creighton says. "They said, 'Jesus, there's no way we can let this get out.'" In particular, Barksdale was worried about the reaction of Wall Street. "My fear was that people would read it as the whinings of some sad-sack loser," he told me. "What would the markets think if we said, 'Well, if the government doesn't help us, we're doomed'?"

And so it was determined that the Netscape white paper would have an audience of one: the DOJ. Creighton was crestfallen; Reback, enraged. For not only had the DOJ already demonstrated its lack of interest in Netscape's ongoing evisceration, but now Joel Klein had been named the acting head of the antitrust division. Reback had no love for Klein, whose first major victory at the DOJ had come in 1995 when he defended the government's consent decree with Microsoft against Reback's challenge in federal court. Reback's suspicions, along with those of many in the Valley, only deepened when soon thereafter Klein took the lead in deciding that the DOJ would do nothing to halt Microsoft's plan to put an icon for its fledgling online service, the Microsoft Network, on the Windows 95 desktop.

For a brief moment, Reback's pessimism seemed mistaken. In September 1996, not long after the white paper was shipped off to Washington, the DOJ announced it was opening an investigation into Microsoft's Internet activities. Years later, after their triumph in court, Klein and his allies would point to this as evidence that, as soon as Netscape came forward with credible allegations, the DOJ jumped on the case like a dog on a bone. But this was revisionist history on a massive scale. The DOJ's investigative team consisted of a couple of lawyers working part-time on the matter in the San Francisco field office. In the course of the next year – a year in which, for all practical purposes, Netscape was reduced to rubble – those DOJ lawyers sent a single CID to Microsoft, limited in scope to the company's dealings with Internet access providers, and a single CID to Netscape. The San Francisco team, whose leader was a bookish fellow named Phil Malone, drove Reback to distraction. "One of them actually said to me, 'browser, schmowser,'" Reback recalls.

If Klein wouldn't act of his own volition, Reback and Creighton decided, they would simply have to goad him, or bait him, or shame him into doing it. The Netscape lawyers began lobbying anyone willing to lend them an ear. The FTC. The Senate Judiciary Committee. The European Commission. They drafted new white papers, these less secret. And they trawled for allies among firms outside Silicon Valley – American Airlines, Walt Disney, publishers, banks – which might one day find themselves reliant on, or beholden to, Microsoft.

The most promising nibble came from an unlikely pond: the office of the Texas Attorney General. Reback, of course, knew that Texas was home to a thriving high tech economy, and to two of the world's biggest PC manufacturers, Compaq and Dell. What he wasn't aware of was that it was also home to a populist, reformist assistant attorney general named Mark Tobey, who'd become suspicious of Gates' power after reading a story in Time magazine about the browser wars. Within weeks of examining the white paper, Tobey issued a set of CIDs to Microsoft and Netscape. When the documents arrived, he was fast persuaded the case was worth pursuing. From then on, Tobey became Reback's staunchest ally in lobbying states' attorneys general to look into Microsoft's behavior.

At first, the AGs were more than reluctant, but as the summer of 1997 unfolded, Microsoft seemed intent on giving them reasons to change their minds. First there was an article in The Wall Street Journal in which Reback's old friend Nathan Myhrvold was quoted as saying that Microsoft's strategy for Internet commerce was to get a "vig" (short for "vigorish," bookmakers' slang for a cut of the action) from every transaction on the Net that used Microsoft technology – every transaction on the Net, that is. Then came stories that Microsoft was negotiating a similar arrangement with cable-television firms when it came to digital TV. Then there was Microsoft's investment in Apple, a deal which marked the official end of what was once computing's fiercest rivalry and demonstrated that Steve Jobs' company was dependent on Bill Gates' for its very survival, and which was seen in the Valley as a mortal blow to Netscape, whose browser was being ousted from its last refuge, the Mac desktop. Suddenly, Reback's alarums were being met with the most welcome three words an agitator can hear: "Tell us more."

His campaign finally starting to catch sparks, Reback submitted to the DOJ a second Netscape white paper – in which he and Creighton contended that Microsoft's goal was to gain a chokehold over all of online commerce – and then quickly orchestrated a series of secret meetings with many of the allies he'd managed to round up, arranging for the DOJ's Phil Malone to witness the proceedings.

For two solid days in the last week of August, Reback turned Wilson Sonsini's Palo Alto offices into a kind of anti-Microsoft three-ring circus. In one conference room, lawyers on the staff of the Senate Judiciary Committee's chair, Orrin Hatch, huddled with an assortment of Silicon Valley executives, collecting leads and gathering evidence of Microsoft's alleged malfeasance. In another conference room down the hall, the general counsels to a number of Microsoft's competitors, including Netscape, Sun, and Sabre – the airline industry's computerized reservation system, which Microsoft planned to take on with its travel site Expedia.com – held brainstorming sessions to map out a wide-ranging political campaign against Redmond on the Hill, in the statehouses, and in the press. The meeting would prove to be the birth of ProComp, an anti-Microsoft lobbying outfit in Washington, DC.

But neither of these was the center ring. That was in the law firm's main conference room, where Mark Tobey, seated beside Malone, Reback, Creighton, Katz, and representatives from the AG's offices of several other states, conducted the first-ever depositions in what would become US v. Microsoft. There, Andreessen, Homer, and other Netscape executives laid out detailed accounts of many of the incidents in the white papers, including, most important, the June 1995 meeting at which Microsoft had allegedly made its market-division proposal. Asked by Tobey why he'd taken notes on the meeting, Andreessen replied, "I thought that it might be a topic of discussion at some point with the US government on antitrust issues." (During the trial, Microsoft would cite the comment as evidence that the meeting was a setup, and Netscape and the DOJ would retort that Andreessen was just being sarcastic. "Bullshit, on both counts," Andreessen told me. "I'd read all the books. I knew their MO. We were a little startup. They were Microsoft, coming to town. I thought, Uh-oh. I know what happens now.")

Malone sat silently and took it all in. For the past year, he had been in charge of the DOJ's desultory inquiry; now he was watching as a state-level law-enforcement official – from Texas, no less – seized the initiative in an investigation of the world's second most valuable corporation. Although Reback was taunting him mercilessly – "Phil, whaddaya think? That didn't sound like a market-division proposal, now did it?" – Malone somehow managed not to lose his composure. Until the very end, that is.

"When the depositions were over," Reback recalls, "Tobey goes up to Malone and says, 'This looks like the endgame. The only remedy I can see is to break Microsoft up.' And Malone turned purple. Purple! Here the DOJ isn't doing anything, and Tobey is saying, Hey guys, it's over. I really thought Phil was about to have a coronary."

For Reback and Creighton, the August meetings at Wilson Sonsini marked a turning point. The lawyers from the Senate Judiciary Committee were leaning their way, and had started talking about the possibility of holding hearings on competition (or the lack thereof) in the software industry – and even, perhaps, of summoning Gates himself to Capitol Hill. Tobey and the states, a contingent that had grown to include Massachusetts and New York, were in hot pursuit. With the founding of ProComp, Microsoft's congenitally disorganized competitors seemed for once to be getting their act together. And, through the good offices of a shaken Phil Malone, the Netscape attorneys had fired a loud, bracing shot across the DOJ's bow.

The message was clear: The Microsoft matter wasn't going away. The real question, however, remained: Was Joel Klein finally ready to listen?

"Microsoft is a company run by engineers. Engineers like simplicity. They don't like nuance. That's how Bill sees the world. And if it's how Bill sees the world, it's how Microsoft sees the world."

III. THE ACCIDENTAL TRUSTBUSTER

Mike Hirshland thought not. Hirshland was Orrin Hatch's number two staffer on the Senate Judiciary Committee. He was barely 30, garrulous, and wicked smart, a former clerk for Supreme Court Justice Anthony Kennedy. He was also a diehard Republican, a free-marketeer, and therefore a man instinctively averse to government meddling in the affairs of commerce. But what Hirshland had learned about Microsoft's behavior troubled him deeply. Returning to Washington from the Valley in the fall of 1997, he began calling up computer manufacturers such as Compaq and Internet service providers such as EarthLink to see if the allegations in the white papers about Microsoft's exclusionary practices held water. After a few weeks of poking around, he was convinced that "this was pretty damn serious."

On a gorgeous autumn day, Hirshland and the Judiciary Committee's chief counsel went down to the DOJ to meet Klein and his deputies. "They told us, 'If what you're basing this on are the Netscape white papers, forget about it,'" Hirshland recalls. "They said, 'A lot of those leads just didn't pan out. Reback? You can't trust that guy; he makes stuff up. And besides, we're not really sure that tying the browser to the operating system is illegal anyway.'"

"What about all the exclusionary contracts?" Hirshland countered. "What about the OEMs? The ISPs? EarthLink? AOL? Gateway? Compaq?" Klein and his team fell silent. "Next thing you know," Hirshland told me, "they had their notebooks out and were writing everything down."

Afterwards, Hirshland and his boss walked back to the Hill. "Jesus Christ, that was all news to them!" Hirshland exclaimed. "Those guys aren't going to do jack."

This was not a unique assessment in the fall of 1997. Joel Klein had been around Washington a long time, and a fairly clear consensus had emerged as to what kind of an antitrust chief he was likely to be. Klein was brilliant, scholarly, and sophisticated; also careful, cautious, and pathologically pragmatic. Politically astute and avowedly pro-business, he was nobody's idea of a tough-talking trustbuster in the tradition of Teddy Roosevelt or William Howard Taft. He would only take cases he knew he could win. And so, therefore, he'd leave Microsoft alone.

In his early fifties, Klein is short and slight, with a perpetual tan and a shiny bald pate. He walks and talks softly, and seems on first inspection to carry no stick at all. The son of a postman, he grew up in Queens, hoping to be a professional athlete. Robbed of that dream by the cruelties of genetics, he focused on academics, graduating magna cum laude from both Columbia University, where he majored in economics, and Harvard Law School. After stints as a clerk to Justice Lewis Powell and an advocate for the mentally ill, he went on to be a founding partner in a boutique Washington law firm specializing in complex trial and appellate work. In the 1980s, he earned a reputation as one of the most accomplished Supreme Court advocates of his generation, arguing eleven cases before the Court and winning eight – a record he may yet have the chance to improve in the Microsoft case.

For Klein, who had wanted badly to be Solicitor General, the antitrust post was a consolation prize – and a prize, in the end, that was almost denied him. Having sailed through his confirmation hearings in the spring of 1997, he hit unexpectedly turbulent waters in the Senate when his name came up for final approval, in large part because of his approval of the controversial merger of the phone giants Bell Atlantic and Nynex. "We've got an antitrust fellow here who rolls over and plays dead," said Senator Ernest Hollings of South Carolina, one of several who put a formal hold on his nomination. With The New York Times calling Klein "a weak nominee" and editorializing that the administration should withdraw him, and with his opponents obstinate and apparently committed, he seemed for a moment to be in serious trouble.

What few people knew was that one of those opponents was Gary Reback, who lobbied Senator Conrad Burns of Montana to put a hold on Klein as well. On Capitol Hill, where the only thing that moves faster than a senator sprinting toward a TV camera is confirmation scuttlebutt, word spread quickly about Reback's maneuvers, and found its way, inevitably, to the ears of Joel Klein. "Of course I heard," Klein told me later. "It did make me smile when Microsoft said I was carrying Netscape's water."

Yet even if Reback's finaglings didn't hurt Netscape's cause, they certainly didn't help. "The situation wasn't good," says Christine Varney, who became Netscape's chief Washington counsel that fall and was an old friend of Klein's. "Netscape found itself in a position where its principal antitrust lawyer had fought tooth and nail to defeat Joel's nomination, and now, lo and behold, Joel was the antitrust AG. As I said: not good."

In the DOJ's antitrust division, Klein was surrounded by lawyers so sober they made him look impetuous. But there was one dissenter to the hypercautious consensus: Dan Rubinfeld, a joint professor of law and economics at UC Berkeley, who had just taken over, at Klein's invitation, as the division's chief economist. Another small bald man with a low-key demeanor and a high-pitched metabolism, Rubinfeld seemed at first glance no more likely than his boss be eager to take a whack at Bill Gates. As a private-sector consultant, Rubinfeld had a lengthy record of appearing as an expert witness in corporate lawsuits, almost always on the side of the defense. In fact, years before, Rubinfeld had served as Microsoft's main expert in its prolonged, and successful, copyright litigation with Apple. "I had no anti-Microsoft bias when I got here," he told me. "I knew those people well. I respected them. I had spent a lot of time up there." Rubinfeld paused. "Though I don't expect I'll get another invitation to Redmond anytime soon."

When Rubinfeld took a look at the white papers, what struck him wasn't so much the catalogue of abuses they accused Microsoft of perpetrating as the clarity of Reback and Creighton's analysis. Since the 1970s, antitrust economics had been dominated by the free-market orthodoxies brought into vogue by a group of University of Chicago scholars such as Milton Friedman and Ronald Coase, who argued that the market functioned so well that government intervention was unnecessary and even harmful. As an academic, Rubinfeld was part of a growing vanguard of "post-Chicago School" economists who rejected those orthodoxies; Garth Saloner, the Stanford professor who worked with Creighton and Reback, was another. Like Saloner, Rubinfeld had spent the past few years thinking about dynamic high tech industries and had embraced the new economic ideas, from network effects to technological "lock-in," being advanced to explain how such industries work – ideas at the heart of the Netscape briefs.

The more Rubinfeld studied the situation, the more he worried about the impending launch of the new version of Microsoft's browser, IE4, which was designed to be more tightly bound to Windows than any previous browser had been. "Nobody would fight over which browser is on the desktop," Saloner told Rubinfeld at a meeting engineered by Reback. "This is about control of the gateway to electronic commerce. This is about somebody" – Microsoft – "potentially owning commerce. We're talking airlines, cars, banks, you name it."

At Rubinfeld's urging, Klein called Phil Malone in San Francisco and told him to send another CID to Microsoft. Broader than the CID a year earlier, it focused particularly on the company's OEM licensing agreements with respect to IE4. As the Microsoft papers began to pour in, the DOJ was startled not only by what they said but by the sheer baldness of how they said it. Two emails sent in late 1996 and early 1997 from Jim Allchin, Microsoft's top Windows executive, to Gates' third-in-command, Paul Maritz, stood out. In one, Allchin began, "I don't understand how IE is going to win. The current path is simply to copy everything that Netscape does packaging and product-wise. Let's suppose IE is as good as Navigator/Communicator. Who wins? The one with 80% market share … My conclusion is that we must leverage Windows more." In the other, he wrote, "You see browser share as job 1. The real issue deals with not losing control of the APIs on the client and not losing control of the end-user experience … We have to be competitive with [browser] features, but we need something more – Windows integration."

Soon the investigators also had in hand evidence supporting several key allegations by Hirshland about exclusionary contracts with OEMs and ISPs, and especially about Microsoft having threatened to revoke Compaq's Windows license if it removed IE in favor of Navigator.

Yet even then, a debate raged in the DOJ about what to do. There were many voices who urged Klein to hold his fire; to investigate further and bring a broad lawsuit, if it was justified, later. Rubinfeld disagreed. Under the 1995 consent decree, Microsoft was prohibited from requiring OEMs to license any other product as a condition of their Windows licenses. But according to Microsoft's marketing plans, that was exactly what it intended to do with IE4. Indeed, the DOJ now had proof that Microsoft had been doing the same for some time with IE3. Why not simply sue the company for violating the consent decree, Rubinfeld asked, and put off any decision about a broader case until later? "The browser market hasn't tipped yet, but it's really close," he said. By filing a narrow case now, perhaps the DOJ could keep that from happening.

Klein had entered the DOJ with little background in antitrust. But in the past two years, he'd learned enough to know that the issues of "tying" and bundling were some of the murkiest areas in antitrust law – areas made even murkier by the subtle and abstract nature of the product in question: code. Still, it was hard for Klein to imagine a clearer case of illegal tying than the one Microsoft was planning with IE4, nor one more manifestly at odds with the letter and the spirit of the consent decree. Moreover, he was aware that in the few months since his confirmation the political winds around Microsoft had shifted appreciably. He knew that the contingent of states looking into the company, which seemed to swell by the week, was charging ahead, and was likely to take action whether he did or not. After a few more conversations between his staff and Mike Hirshland, he knew that the Senate Judiciary Committee was planning to hold hearings. He knew that Democrats on the Hill still had doubts about whether he had the stomach to joust with big business. And, while he still was far from sure about bringing a broader suit, he sensed in his gut this was one he could win.

And so, on October 20, 1997, Klein stood next to Attorney General Janet Reno, with flashbulbs popping and cameras whirring, as she announced that the DOJ was not only seeking an injunction against Microsoft for violating the consent decree, but was asking a federal court to impose a fine of $1 million a day – the largest civil fine in Justice Department history – until the company stopped tying its browser to Windows. "Even as we go forward with this action today," Klein added, "we also want to make clear that we have an ongoing and wide-ranging investigation to determine whether Microsoft's actions are stifling innovation and consumer choice."

Out in Silicon Valley, Gary Reback heard that, laughed, and wondered if Klein was just blowing smoke. "This filing is a fine first step, but it's only a first step," Reback muttered to me over the phone. "All we can do is hope that it's just the first shoe to drop."

What no one could have guessed – not Reback, not Klein, and certainly not Gates – was that, for Microsoft, it was the start of a hailstorm of footwear that would continue, amazingly and unabatedly, for the next three years.

"Bill said, 'Of course, I have as much power as the president has.' Melinda's eyes got wide, and she kicked him under the table, so then he tried to play it off as a joke. But it was too late."

IV. THE SHADOW OF THE MAN

The morning the news from Washington broke, Gates was in the high desert outside Phoenix, attending a high-end high tech conference called Agenda at the famously opulent Phoenician Hotel. That night, rather than mingle with the rest of the industry's A-list – Andy Grove, John Chambers, Steve Case, Scott McNealy – at the official dinner, Microsoft's CEO retreated to a private supper with a handful of friends. When the conversation turned to the DOJ, he explained in a tone at once dismissive and defiant why the government was wrong, why Microsoft was right, and why, in the end, he had nothing to worry about. Gates held forth at some length on these subjects, but it was a single sentence from his former girlfriend, the Silicon Valley venture capitalist Ann Winblad, who had spent the better part of the day holed up with him in his room as he absorbed the details of the case, that most perfectly captured his reaction to the lawsuit:

"These people have no idea who they're dealing with."

The next day, the man the government was dealing with took his turn on the Agenda stage. Dressed in a madras-plaid shirt and a pair of khakis, Gates laid out his company's arguments in unequivocal terms: that the consent decree specifically allowed Microsoft to develop "integrated products," and that IE was just such a product – fundamentally melded into Windows. "There's no magic line between an application and an operating system that some bureaucrat in Washington should draw. It's like saying that as of 1932, cars didn't have radios in them, so they should never have radios in them." The central question, Gates contended, was this: "Is one company excluded from innovation, or not?"

From the audience, Gates was asked about public opinion, about the growing sense, not only in Washington but in the industry at large, that Microsoft was wielding its power too wantonly. "You're sort of asking us if we're going to change, to start telling engineers, 'Slow down, slow down. Go home,'" Gates replied. "No, we're not."

Through most of the session, Gates was calm and collected, if occasionally curt. Then Rob Glaser, a former protégé of his at Microsoft and now the CEO of the Web media-streaming firm RealNetworks, stepped up to the microphone. "Bill, do you really think there is no limit to what should or should not be included in the operating system?" Glaser asked. "If there is a limit, who should set it? Microsoft? The Justice Department?"

"Look, look, this is called capitalism!" Gates snapped. "We create a product called Windows. Who decides what's in Windows? The customers who buy it."

For Gates, the Q&A at Agenda was a gentle preview of what lay ahead. Although the DOJ was the primary provocateur, it wasn't the only one. The European Commission had opened an investigation of its own. Soon, the Japanese government would do the same. Ralph Nader, the old economy's hoariest rabble-rouser, was organizing an anti-Microsoft summit in Washington, featuring some of Redmond's most vocal foes. One of them was Sun CEO Scott McNealy, who had just filed a separate lawsuit over Microsoft's use of the trendy software technology Java, in which Sun accused Microsoft of breach of contract, trademark infringement, false advertising, and unfair competition.

Thus in the fall of 1997 did Microsoft find itself subject to a degree of public scrutiny unlike any it had received in its 20-year history. Its reaction was telling.

First there was Steve Ballmer, standing on a stage in San Jose a few days after the DOJ's filing, bellowing, "To heck with Janet Reno!" Then there was Microsoft's first formal response to the case, a legal memorandum which labeled the DOJ's arguments "perverse," "uninformed," "misguided," "misleading," "wrong," "just wrong," "simply wrong," and "without merit," and which suggested that the government was acting not on behalf of consumers but the company's competitors.

Then there was the business with the ham sandwich. When the DOJ issued its reply brief to Microsoft's memorandum, one passage stood out. "Microsoft asserts that 'integrated' means whatever Microsoft says it means," the brief said. "Indeed, in its discussions with the government before the Petition was filed, Microsoft flatly stated that its interpretation of the [consent decree] would enable it to require OEMs to put 'orange juice' or 'a ham sandwich' in the box with a PC preinstalled with Windows 95."

This was true. At a meeting with the DOJ before Klein pulled the trigger, Richard Urowsky of the New York firm Sullivan & Cromwell – Microsoft's primary outside legal counsel – had let his flair for the dramatic flourish run away with him. Even today, three years later, Microsoft's legal team still fumes over what it calls the government's "ham-sandwich leak." "It was taken totally out of context," a Microsoft lawyer tells me. "What he said was, 'We could put in a ham sandwich, but nobody would buy it.' It was a perfectly legitimate thing to say. People wouldn't buy it if we put a ham sandwich in the OS. It was a metaphor for consumer choice." Unfortunately for Microsoft, Urowsky's declamation, repeated endlessly in the press, was taken as another sort of metaphor entirely: a metaphor for its arrogance, for its unwillingness to acknowledge any limits to its power.

As autumn began to fade into winter, Microsoft was being roughed up in the media, and the company's reaction seemed to grow only more clumsy and paranoid each day. The trend reach new heights at its annual shareholders meeting, when Gates lashed out at the "witch-hunt atmosphere" being ginned up by his enemies in the Valley and in DC. All through its history, Microsoft had been adroit, even masterful, at presenting its image to the public; now it seemed to be melting down. The sight was so strange, so unexpected, I was sure the press accounts were conjuring an exaggerated impression. There was no way the company could really be as rattled as it seemed.

Then I went to see Steve Ballmer.

Ballmer is Gates' best friend, a classmate of his at Harvard who worked briefly at Procter & Gamble and spent a year at Stanford's business school before joining Microsoft in 1980. He has worn a number of official hats at the company, but, unofficially, he has always been Gates' number two. If Gates is Microsoft's ego, Ballmer – beefy, boisterous, a natural-born cheerleader – is its rampaging id.

Even so, I was unready for what occurred when we met on a chill December afternoon in San Francisco, where Ballmer had come to deliver a speech to some customers. Sitting in a windowless conference room at the Westin St. Francis hotel, I asked Ballmer about an internal Microsoft document concerning Microsoft's licensing of Java, which had come to light in the DOJ's investigation. In it, PaulMaritz stated that the company's goal was to "get control of" and "neutralize" Java, whose cross-platform raison d'être was seen as posing a threat to Windows. Scott McNealy had told me he considered the document prima facie evidence that Microsoft had signed its contract in bad faith. I asked Ballmer if McNealy was right.

"Sun is just a very dumb company," Ballmer began.

"We always honored our license. We always intended to. We always have." His voice quickly rising, Ballmer continued, "Sun wasn't confused. We weren't coming in there saying, Hallelujah, brother! We love you, Sun! We said, We don't like you as a company – nice people; I like Scott – and you don't like us! We said, Hey Sun, you want to get on the back of us and ride, baby, ride You want on? OK, here's the terms!"

Ballmer's face was beet-red now, and he was screaming so loudly that, had there been any windowshades, they would have been rattling. Up on his feet, leaning across the table so that his face was no more than 6 inches from mine, pounding his meaty fists on the tabletop so hard that my tape recorder leapt and skittered, he roared, "Nobody was ever one little teeny tiny bit confused that we and Sun had this wonderful dovetailing of strategic interests! Those sub-50-IQ people who work at Sun who believe that are either uninformed, crazy, or sleeping!"

I took this as a Yes.

Extending a long middle finger to the government and your competitors is not conventional behavior among the top executives of most blue-chip companies. But, of course, Microsoft was different – self-consciously so. Populated by an army of young men (mainly), most of them unusually bright, many of them abnormally wealthy, working endless hours and pulling frequent all-nighters, Microsoft has always retained the air of a fraternity – a fraternity of rich eggheads, but a fraternity nonetheless. For years, Softies were wont to sport buttons that read FYIFV: Fuck You, I'm Fully Vested. Another favorite acronym, meant to suggest how far the company would go, in Ballmer's words, to "get the business, get the business, get the business," was BOGU: Bend Over, Grease Up.

Machismo, callowness, and profanity are not exactly unique to Microsoft. What is unique, however, is the intense insularity of the Redmond culture. Situated hundreds if not thousands of miles from its competitors and partners, staffed mostly by folks who have never worked anywhere else, Microsoft is the frat house from another planet. Time and again, its engineers express apparently genuine surprise and a lack of comprehension that other high tech companies harbor deep and abiding suspicions of their employer. Even Ballmer, a sharp guy despite all the hollering, was quoted this past June in Newsweek saying, "People say a lot of things about us, but never has anyone said we're untrustworthy." Hello?

At the very heart of the Microsoft culture is technology – an assertion that will sound either axiomatic or ludicrous depending on your prejudices. To most Americans, Microsoft is more than a technology culture; it is the technology culture. In the Valley, though, the view is different. There, even among some Microsoft allies, it is an article of faith that the company is incapable of innovating; that it is a copycat, a "fast follower," an assimilator of breakthroughs achieved elsewhere; that its products, despite their awesome popularity, are crashingly mediocre.

No matter what outsiders may think, Microsoft executives fervently believe that their company does in fact innovate, a belief they support by pointing to the extraordinary $3 billion the company spends each year on research and development, in areas ranging from voice recognition to artificial intelligence. Yet starting in the early 1990s, the company also devoted vast resources to buffing its image, waging multimillion-dollar advertising campaigns and carefully orchestrating press coverage to turn Microsoft, Windows, and Gates himself into household names. One of the clearest indications that Microsoft was becoming as much a marketing culture as an engineering culture came in 1994, with the hiring of Robert Herbold as chief operating officer. A mild fellow of middling age, middling stature, and a certain bland charm, Herbold was a computer-science PhD who had risen to become the top marketing executive at Procter & Gamble. He spoke the lingo of branding, of corporate identity, of making "deposits" in the "key mental bank accounts" of customers. On arriving at Microsoft, he quickly implemented the full complement of consumer-research techniques that he'd used at P&G, from extensive polling to focus groups.

As Microsoft began its recalcitrant flailing in the fall of 1997, I couldn't help wondering what Herbold was thinking. Here was his company, violating every conceivable rule in the big-brand handbook of crisis management. Consider: What would McDonald's do if it found itself in similar straits? What would Coca-Cola do? Or Disney? Answer: Their CEOs would appear on the DOJ's doorstep and ask, in voices sugary with solicitude, What can we do to make the problem go away? Yet this approach never seemed to have occurred to anyone at Microsoft. A few months later I visited Herbold in Redmond and asked if it made sense to interpret the company's belligerence as a sign that Microsoft had failed to internalize the idea that its success rested on its image as well as its technology.

"Yeah, it does," Herbold said. "But there comes a point in any company's life where, if a fundamental principle as to how you operate is being threatened, you have no choice but to stand tall." Like Gates and everyone else I spoke to at Microsoft, Herbold was adamant that the consent-decree case threatened to undermine the company's ability to innovate. If thwarting that meant extreme and even potentially self-destructive measures, so be it.

"Always keep in mind, Microsoft is a company run by engineers," a departed Microsoft executive, himself an engineer, said to me later. "Engineers like simplicity. They like clarity. They like rules. They don't like nuance. They don't like shades of gray. They're totally binary. Ones or zeros. Black or white. Right or wrong. Innovate or not innovate. That's how Bill sees the world. And if it's how Bill sees the world, it's how Microsoft sees the world.

"Remember, no one has ever accused Microsoft of being a democracy."

Gates "was going through a period where he kept saying, 'I hate my job. I hate my life. I hate this situation. I don't know what to do.'"

Nowhere in the annals of modern business has Emerson's aperçu that "an institution is the lengthened shadow of one man" held more abundantly true than at Microsoft. From the moment the company was founded, everything about it – good and bad, strong and weak – has been a pure crystalline reflection of Gates' mind, his personality, his character. In the computer industry, few founders have been able or willing to stick with their firms as they've grown, guiding them from birth to maturity. Scott McNealy is a notable exception; so is Larry Ellison, of Oracle. But although McNealy and Ellison are both forceful and dynamic CEOs, neither has ever come close to exerting the type of hold over his company that Gates has always maintained over Microsoft.

Gates inspires this intense following without being, in any conventional sense, a charismatic or especially winning figure. What he is, is very smart, and in the Microsoft culture that he himself has engendered, smartness is valued above all. "There are probably more smart people per square foot right here than anywhere else in the world," the former Microsoft executive Mike Maples has said. "But Bill is just smarter."

The slavish fealty accorded Gates at Microsoft draws gales of derision from critics and competitors. Netscape's former counsel, Roberta Katz, says it was the "blind obedience, the willingness to suspend all judgment and follow the party line, all this zombielike devotion to the Maximum Leader" that led Microsoft inexorably to its fate in the courts. "It's the whole voice-of-God thing," says Bill Joy, Sun's chief scientist. "They're always asking, What would Bill think? As if Bill's the oracle. As if Bill knows best. It's hard to be creative in that kind of environment, and it's very hard to do clean-sheet work, because all the old stuff is the oracle's stuff, and who's going to tear that up to start fresh? It's why they can't innovate no matter how many smart people they hire." Gates, says Joy, is "the low priest of a low cult."

While the notion that Gates is a technological genius is a central part of his public legend, the depiction elicits eye-rolling (and less charitable responses) in computing circles, where his technical gifts are regarded almost universally as solid but unexceptional. "Neither Bill nor Paul &##91;Allen] was tremendously technically sophisticated when they started Microsoft, and they're not now," says David Liddle, the former director of Allen's now-defunct think tank, Interval Research, and a friend of both men. In 25 years working in software, Gates personally has made no significant contributions to computer science. He holds but one patent. Yet at Microsoft, top-flight scientists speak of his technical fluency in tones of awe. Gates, they say, is a fox and not a hedgehog; a technologist whose strength is breadth, not depth. Craig Mundie, the Microsoft executive who has spent more time recently with Gates discussing the future of technology than anyone, told me, "Bill's great gift is synthesis: his ability to accumulate a huge amount of information and then synthesize it on a grand scale."

In a way, the myth of Gates as a mighty technologist has overshadowed his rightful claim to genius as a businessman. Of course, Gates is often credited, and justly so, with being among the first to discern that software could be the basis of an enterprise; with having appreciated that software, not hardware, was where the serious money would be made in personal computing; and with having shrewdly persuaded IBM, when it asked Microsoft to provide an operating system for its first PC, in 1980, to allow his firm to retain the rights to that software, MS-DOS. But Gates' insights were far more sweeping than that. Before he arrived on the scene, the computer industry had always been organized vertically. That is, it consisted of companies like IBM and DEC that built their own machines, designed and manufactured their own chips, and developed their own operating systems and applications, all of them proprietary. Side by side with Intel's CEO, Andy Grove, Gates envisioned a different structure, a horizontal structure, in which specialized competition would take place in each layer of the industry: chip company versus chip company, software company versus software company, computer company versus computer company. He figured out, again with Grove, that the position of maximum power and profit in this new structure came from owning one of two critical industry standards: the OS or the microprocessor. And, finally, he understood that Microsoft's control of the OS standard could be leveraged in ways that would give the company enormous advantages in competing for other software markets.

Gates' strategic foresight was twinned to a tactical discipline and a single-mindedness that were unusually fierce. For a long time, he seemed oblivious to the marginalia of corporate life, the perks and the status symbols, that distract so many executives. His office was modest. He disdained titles. He flew coach. And while he never suffered from a deficit of ego, he was relatively immune to intellectual vanity, keeping close tabs on ideas and trends gaining currency beyond Microsoft's borders. "He carefully reads the wind and weather," Liddle says, "and he does not have false pride about admitting he's wrong" – as he did most famously in turning Microsoft around in the mid-1990s after initially missing the emergence of the Internet.

Nor was he prone to technical vanity. Where other high tech CEOs wasted time and money pursuing perfect, elegant solutions, Gates refused to let the great be the enemy of the good, or even to let the good be the enemy of the minimally serviceable. Over and over, he attacked new markets with the same pragmatic sequence of moves: Dive in fast with a half-assed product to establish an early foothold, improve it steadily (even Microsofties joke that the company never gets anything right until version 3.0), then use clout, low prices, and any other means necessary to gobble up the market. About the extent of Microsoft's appetites, Gates and his lieutenants were unabashed. "My job is to get a fair share of the software applications market," Mike Maples said in 1991, on the eve of the launch of Office. "And, to me, that's 100 percent."

Gates' hunger for new conquests left a trail of bloody bodies strewn in Microsoft's wake. Digital Research. WordPerfect. Novell. Lotus. Borland. Apple. "Bill [had an] incredible desire to win, and to beat other people," ex-Microsoft executive Jean Richardson recalled in the PBS documentary Triumph of the Nerds. "At Microsoft, the whole idea was that we would put people under."

But while Gates' style of competition was at once relentless and remorseless, it seemed to be fueled as much by anxiety as by cruelty. Long before Andy Grove made "Only the paranoid survive" the watchphrase of Silicon Valley, Gates was living the mantra at Microsoft. "Bill runs scared much more than people think," says William Randolph Hearst III, a Valley venture capitalist and one of Gates' closer friends. "He does what he does out of fear, not sadism. The history of business is full of guys looking out the 50th-story window of their corporate headquarters, seeing some little pipsqueaks down below, and going, 'Oh, forget it; how could they ever threaten us?' And then getting their clocks cleaned. Bill just knows he doesn't want to be one of those guys."

Or, as Gates himself said to me one day in his office, "The fact that you can't name the place you're going to die doesn't mean you shouldn't pay attention to your health."

The mortality of skyscraper-dwelling overlords was a phenomenon with which Gates was intimately familiar. When his company's partnership with IBM began, Big Blue was arguably the exemplary corporation of the modern age. It was 3,000 times the size of Microsoft and had defined commercial computing for three decades. "It's easy to forget how pervasive IBM's influence over this industry was," Gates recalled. "When you talk to people who've come into the industry recently, there's no way you can get it into their heads: IBM was the environment." Then the men from Armonk met Gates, and everything changed. By the early 1990s, not only had IBM's hegemony been shattered, but the company was on the ropes – losing billions of dollars a year, laying off employees by the thousands, struggling for its very survival. Meanwhile, Microsoft was ascendant. In January 1993 it surpassed IBM in market value and never looked back; a few weeks later, IBM's board tried in vain to recruit Gates to become the company's chair. The role reversal was complete: Microsoft was now the environment.

The fall of IBM was a seminal experience for Gates and Ballmer, shaping their perspectives in countless ways, both obvious and subtle. "If you asked me where I learned more about business than anyplace else, I wouldn't point to school, I wouldn't point to my two years at Procter & Gamble, I wouldn't point to Microsoft," Ballmer told me. "I would point to my 10 years working with IBM." With the passage of time, he and Gates would come to extol and emulate IBM's strengths – its devotion to research, its attentiveness to customers. But during Microsoft's formative years, their opinions were somewhat less favorable.

"We hated IBM," says Peter Neupert, a former Microsoft executive who worked with Big Blue on the joint development of the operating system OS/2 and is now the CEO of drugstore.com. "We hated their decisionmaking process, which was incredibly bureaucratic and stilted. We hated their silly rules and requirements; the red tape was unbelievable. And we had zero respect for their engineering talent. The core of Microsoft is: Great talent matters. We had a great team; theirs was big, slow, and sloppy." (Among the OS/2 coders, IBM stood for Incredible Bunch of Morons.) "We fought bigness at every stage. We had no processes. We had no planning department. Anything that would slow decisions down was rejected by design. Bill wanted to preserve a freewheeling style, where you made decisions fast and didn't get bogged down. It all comes from his programmer orientation. The people who were rewarded most at Microsoft were cowboys and misfits – the guys IBM would never hire. That was a point of pride."

If IBM provided Gates with an object lesson in the perils of gigantism, it also offered him a case study in how debilitating a constant fear of government intrusion could be. From the early 1950s until the early 1980s, IBM had been continually under investigation by, or in litigation with, federal antitrust authorities. In 1956, the company had signed a consent decree that forced it to license its patents at a "reasonable" price to all comers; and in 1969, the DOJ had launched its landmark 13-year lawsuit accusing IBM of illegal monopolization of the computer industry – a lawsuit which, despite being dropped in 1982, saddled the company with a legacy of competitive restraint and legalistic caution that played no small part in its vulnerability to the PC revolution that Microsoft spearheaded. "Every decision they made – on products, packaging, marketing – was based at least in part on legal constraints or perceived legal constraints," Neupert recalls. "It was screwy." And it made a large and lasting impression on the boys from Redmond. "Bill thought a lot about it. The question was: How important are we going to let the lawyers be at Microsoft? In dealing with IBM, they'd have lawyers in technical meetings. Ludicrous."

Gates' answer to the question was: Not very. It would prove fateful. In 1985, the year before Microsoft went public, its legal department consisted of Bill Neukom and two other employees. In the next 15 years the department would steadily expand to more than 400 employees, 150 of them attorneys. Yet despite all those warm bodies, through the 1980s and most of the 1990s Microsoft failed to adopt an official antitrust-compliance policy or a comprehensive antitrust-training regime for its employees.

Today, Microsoft's lawyers are at pains to deny that this is so. They produce documents listing an array of programs (Executive Competition Counseling, Consent Decree Training, Legal Road Shows) intended "to ensure that Microsoft employees understand and comply with legal obligations under US and other antitrust laws." Antitrust training has even been incorporated into the "Microsoft 101 training vehicle" for all new employees – although that incorporation took place in 1999, well after the company's imbroglio with the government began.

Ballmer recently insisted to me that Microsoft has had "antitrust audits, antitrust reviews, antitrust training" since the mid-1980s. "Now, do we train every Tom, Dick, and Harry in the company?" he said. "No. But it's not every Tom, Dick, and Harry that's making the decisions." Yet in dozens of interviews with current and former Microsoft executives, I found few who could recall having received antitrust training, and of those who could, even fewer who remembered anything they'd been taught, beyond the vague instruction to "obey the law." (On the stand in the trial, Paul Maritz would testify that he knew of no antitrust-compliance policy at Microsoft.)

To trustbusters such as Joel Klein, Gates' unwillingness to implement a thorough antitrust program was a plain sign of his immaturity as a CEO. "Major corporations in America have these things – they just do," Klein told me. "It's just sensible; it's just prudent." Even in high tech, the absence of such a program at Microsoft has long raised eyebrows, including those of Gates' ally, Andy Grove. Grove, who would no more concede that his company has a monopoly on PC microchips than he would admit a fondness for New Age management techniques, instituted a far-reaching antitrust regimen at Intel as long ago as 1986. For years thereafter, he periodically raised the issue with Gates, and then complained to other Intel executives about Gates' "pigheaded" refusal to follow suit. Yet something more complex and calculating than mere pigheadedness was at work. To Gates' way of thinking, being without an antitrust program may have carried with it certain legal risks, but the risks of enacting one were even greater. "Bill's thought was that once we accept even self-imposed regulation, the culture of the company will change in bad ways," a former Microsoft executive told me. "It would crush our competitive spirit."

Morris' top-secret "Project Sherman," which comprised a superstar group of antitrust authorities, would span three months and consume $3 million of Sun's money.

Or, as Gates himself put it to another of the industry's leading CEOs, "The minute we start worrying too much about antitrust, we become IBM."

Years later, when bemused analysts and commentators tried to explain the behavior that got Microsoft into such hot water with the government, one theory in particular came into vogue: After years of seeing itself as David, the feisty underdog doing battle with the industry's behemoths, Microsoft had failed to realize that somewhere along the way it had become Goliath – and that Goliaths were subject to a stricter set of rules than Davids were. The truth, however, was slightly different. Gates hadn't failed to recognize anything. After witnessing the collapse of IBM up close and personal, he was determined not to let Microsoft fall prey to a similar syndrome, and had repeatedly taken explicit steps to preserve the company's Davidian attitudes and attributes in spite of its mass and muscle. The result was a culture built on a willing suspension of disbelief; a culture whose public posture was, in 1997, neatly – and ridiculously – summed up by COO Bob Herbold thus: "Think about the technology business in its broadest sense. Microsoft is a small but important player in that very large industry."

In private, though, when the man who ran Microsoft let down his guard, he betrayed no confusion about what he and his company had become. A close friend of Gates' recalls a dinner with him and his then-fiancée (now wife) Melinda French back in 1993. "We were talking about Clinton, who'd just been elected, and Bill was saying blah, blah, blah about whatever the issue was," this friend remembers. "Then Bill stopped and said, 'Of course, I have as much power as the president has.' And Melinda's eyes got wide, and she kicked him under the table, so then he tried to play it off as a joke. But it was too late; the truth was there. If Bill ever thought of himself as a scrappy little guy, he didn't anymore."

By the middle of the 1990s, Gates may have been as powerful as the president in some ways, yet he remained as paranoid as a speed freak at the end of a very long binge. The proximate cause of his paranoia was Netscape. In May 1995, in a now-famous memo titled "The Internet Tidal Wave," Gates argued that the startup's browser held the potential to "commoditize the underlying operating system" – Windows. What worried him, Gates told me, wasn't merely the threat posed by the browser or other forms of middleware but the sudden momentum Netscape had gained in the industry. "Lightning struck," Gates said. "There was a belief that they were the exciting thing, they were the coming company. You'd go to their developer conferences, go to Marc Andreessen's press conferences, read the article about what flavor of pizza he ordered. That phenomenon was getting developers to pay a lot of attention to the Netscape browser." He added, "Expectations are a form of first-class truth: If people believe it, it's true." And people were believing in Netscape.

As was Microsoft, in a sense. When Andreessen and his colleagues first started talking about turning their lean little browser into a full-blown platform, the idea struck Gates and Ballmer as perfectly plausible – not surprisingly, since Microsoft had pulled off the same trick in the course of ten years with Windows, which was originally nothing more than an application running on top of DOS.

The only thing that surprised Microsoft about Netscape's strategy was the brazenness with which the upstarts shouted it to the world. Nathan Myhrvold told me, "There's a good analogy to bicycle racing. In bicycle racing, you don't want to be first until the end. What you want to do is draft the guy in front of you. And then, in the last minute, you dart out. The middleware gambit is about drafting the leader." Yet here was Andreessen publicly proclaiming in the summer of 1995 that Netscape's plan was to reduce Windows to "a poorly debugged set of device drivers." "They didn't save it up," Myhrvold said. "They fucking pulled up alongside us and said, 'Hey, sorry, that guy's already history.'"

The tactic drove Redmond into a rage. The day after Andreessen's quote appeared in the press, John Doerr, the prominent venture capitalist and Netscape board member, received a chilling email from Jon Lazarus, one of Gates' key advisers. In its entirety, it read: "Boy waves large red flag in front of herd of charging bulls and is then surprised to wake up gored."

The consent-decree case resulted from Microsoft's very first thrust: the decision to bundle and then integrate IE into Windows. Even apart from its effect on Netscape, Gates firmly believed that Web browsing was a natural addition to any OS, one that would serve consumers and make computing easier. Adding IE to Windows for free, he told me, was "the most defensible thing we've ever done." It was also indisputably legal, he said. When Microsoft had bargained with the DOJ (and with the European Commission, which was simultaneously pursuing its own investigation) over the consent decree in 1994, Gates had taken great care to be sure that the provision on tying was worded broadly enough to give Microsoft unfettered freedom to put new features into Windows. Indeed, when Neukom presented Gates with a proposed draft of the decree which stated that Microsoft would not be prohibited from "developing integrated products which offer technological advantages," Gates barked, "Remove those last four words!"

Gates, Neukom, and the rest of Microsoft's legal team were therefore stunned when the DOJ filed the consent-decree case. It seemed to them that the Feds were either woefully unaware of the negotiating history of the decree (given that the deal was cut under Klein's predecessor) or had willfully chosen to ignore it. Equally maddening was the premise of the DOJ's claim: that because Explorer was distributed to PC makers on a different disk from Windows, and because it was also marketed as a standalone product, it was by definition not "integrated." At a meeting with the DOJ that fall, Klein held aloft the two disks and said, "See? Two separate products." To Microsoft, the gesture was glaring evidence of Klein's technological cluelessness. Once IE and Windows were installed together, they fused into one seamless whole; the fact that they were distributed on separate disks, as software products often are, was irrelevant. "It's all just bits," Neukom said to me later. "Antitrust law isn't about how you distribute the bits; it's about how the bits relate to each other."

Klein may have been clueless about the commingling of code, but the DOJ's argument found a friendly pair of ears on the large round head of Thomas Penfield Jackson. Jackson was the gruff, grandfatherly federal judge who had somehow lucked into hearing the consent-decree case. After nearly two months of legal volleys, on December 11 he issued a stopgap split decision that cut sharply against Microsoft. On one hand, the company had offered a "plausible interpretation" of the term "integrated" and a "reasonable explanation" as to why its behavior was kosher under the consent decree; so Jackson rejected the government's motion to fine Microsoft $1 million a day. On the other, though the judge remained undecided on the merits of the case and needed more time to sort out the issues, he found that the DOJ "appears to have a substantial likelihood of success" and that "the probability that Microsoft might also acquire yet another monopoly in the Internet browser market is simply too great to tolerate indefinitely until the issue is finally resolved." And so Jackson handed down a preliminary injunction ordering Microsoft to "cease and desist" from requiring PC makers to install IE as a condition of their Windows licenses. Until the case was decided, Microsoft was to offer them a browser-free version of the OS.

Microsoft's response was flagrant, provocative, and ill-considered. Having maintained all along that removing the browser code from Windows would break the OS, the company decided to comply with Jackson's order in a remarkable fashion: by offering OEMs a choice of either a two-year-old version of Windows without IE or a current version that simply didn't function. Joel Klein was livid. "Usually the phrase 'contempt of court' is metaphoric," he sputtered to me. "In this case, it was literal."

Microsoft's maneuver led to the consent-decree case's most famous – and famously comical – incident. Before a packed courtroom, Judge Jackson announced that he and his clerks had been doing some hacking, and had found that IE could be uninstalled with no noticeable harm to Windows in "less than 90 seconds."

A few weeks later, in mid-January, after another hearing in which Jackson heaped scorn on Microsoft and its witnesses, the company backed down. In consultation with the DOJ, it agreed to offer computer makers a version of Windows that still contained some IE code, but in which the browser was disabled and hidden from view. Today, Gates and his lawyers still refuse to admit that this was what they should have done in the first place, not least because most PC manufacturers would have continued (and in fact did continue) to take the version of Windows that included IE. "Do I wish we'd found a more politically, personally, atmospherically palatable response?" one of Microsoft's senior lawyers muses. "Sure. But we couldn't then and we still can't."

"Maybe we should have gone to the DOJ and said, Hey, this won't work. Why don't we go to the judge and try to figure it out?" another Microsoft attorney tells me. "But we were in an adversarial situation, remember. And we were trying to make a point that was lost on the court."

The price of making that point would prove to be greater than Microsoft could ever have imagined. Two and a half years later, when Jackson issued his order that the company be split up, he cited its "illusory" and "disingenuous" compliance with his injunction in the consent-decree case as evidence that Microsoft was "untrustworthy" and that conduct remedies alone weren't sufficient to rein in its power. And even in the short term, the damage was severe. In America and abroad, in the news columns and in editorial cartoons, criticism, sarcasm, and even mockery suddenly appeared where once there'd been little besides adulation. For the first time ever, Ballmer acknowledged that the company's polling and focus groups had begun to show that the negative publicity was taking a toll on Microsoft's image. "It's not cataclysmic, but it's clear," he said.

At the same time, Microsoft's insolence seemed only to have emboldened the DOJ and the states as they turned their attention to the question of whether to launch a full-scale antitrust action against the company. If anyone had a doubt that they were serious, one piece of news should have instantly dispelled it: that Klein had retained David Boies, the famed New York litigator who had successfully defended IBM against the government's antitrust charges in the 1970s and 1980s, as a consultant.

The gathering storm was unlike anything Gates had ever weathered. Competitors had been assailing him and his company in every fashion imaginable for more than a decade. But what was happening now … this was different. This wasn't business. This was the government, an adversary not unknown to Gates, but one against whose slings and arrows his defenses weren't nearly so robust.

In the months ahead, it would often be said that, for a company of its importance Microsoft had paid dangerously little attention to politics over the years. As recently as 1995, the company had no government-affairs office in Washington, DC. Yet Gates didn't think of himself as a political innocent. He had never been partisan, but who was these days? He had issues he cared about – trade, immigration, encryption, taxes – and had lobbied on behalf of. He had even dabbled a bit in the art of the schmooze. He had golfed with Bill Clinton on more than one occasion. He had dined with Newt Gingrich back when that meant something. He had hosted Al Gore on a visit to Microsoft. (For a time, Gore's daughter Karenna had worked at >Slate.) More to the point, Gates believed that he and Microsoft had delivered to this administration perhaps the greatest political gift of the postwar era: the new economy. Who had done more than he had to spark the PC revolution? What company had done more to provide the underpinnings of the information age?

Reback assembled a Murderer's Row of Valley executives, financiers, and technologists who would parade before the Project Shermanites during a single daylong session.

Directly and indirectly, Microsoft had generated untold wealth. In Windows, it had built a platform on which much of the high tech economy stood. It had created products on which millions of workers relied. It had propelled the Nasdaq to improbable heights. And now, after all this, after all he had done, the government that should have been showering him with praise and gratitude was casting him as a villain, a scoundrel, a grasping monopolist. It was crazy, infuriating. And it was starting to get under his skin.

As the consent-decree contretemps wound to a close, the blind outrage that had colored Gates' mood for months remained intact, but increasingly it was overshadowed by something darker. Among his small circle of close friends, word began to spread that Gates had fallen into a deep blue funk. "His own government suing him, that's not chocolate sundae," his father would later tell Newsweek. "He was concerned, he was angry, he was distracted from things he'd rather be doing." Actually, it was much worse than that. According to one old friend, "He was going through a period where he kept saying, 'I hate my job. I hate my life. I hate this situation. I don't know what to do.'"

Seeing Gates so demoralized disturbed his friends. It also worried the Microsoft board. On January 24, the directors (who included Steve Ballmer, Paul Allen, the ex-Microsoft president Jon Shirley, venture capitalist Dave Marquardt, Mattel CEO Jill Barad, and a Hewlett-Packard executive named Richard Hackborn) gathered for their monthly meeting. It was a gray Saturday just 72 hours after the company had come to terms with Jackson and the DOJ on the preliminary injunction, and the board expected that much of the meeting would be taken up with discussion of the consent-decree case and the broader suit that the government was contemplating. At least a few of Microsoft's directors were hoping to raise another issue as well: the possibility of promoting Ballmer to president from his current position as executive vice president of sales and support – in order, as one board member put it to me, "to take some of the burden off Bill's shoulders." Yet it was only when Gates began to speak that anyone fully realized how great the burden had become.

Looking haggard, as though he hadn't slept in days, Gates plunged into an extended and emotional tirade, railing at the DOJ, castigating the judge, bemoaning the sheer irrationality of what had befallen his company. Everyone in the room was familiar with Gates' outbursts, which were, after all, a signature of his leadership style. But this was a different brand of diatribe – more stream-of-consciousness than usual, and far more personal. His voice quavered; his body quaked. And where Gates in full lather was normally condescending and sometimes cruel, now he was seized by unbridled self-pity. The DOJ was demonizing him. The press hated him. His rivals were conspiring to take him down. The political establishment was ganging up on him. His enemies were legion; his defenders, mute.

How had this happened? What could he do?

Gates' eyes reddened. "The whole thing is crashing in on me," he said. "It's all crashing in."

And with that, the richest man in the world fell silent, and began to cry.

V. THINGS FALL APART

It was said by admirers and antagonists alike that Gates was endowed with a greater ability than perhaps any CEO in history not only to see several chess moves ahead but to do so on several chessboards simultaneously. Yet no matter how many chess games are being played, the same rules apply from board to board. Knowable rules. Fixed rules. The trouble for Gates and Microsoft was that the ordeal they now confronted was less like a chess match than a piece of improvisational theater, where the stage is full of actors armed with different scripts, motivations, and objectives. Careering around the proscenium, this motley cast – Microsoft, the DOJ, the states, Silicon Valley, Judge Jackson, and the rest – would at times stay in character; at times not. At times they would read out well-rehearsed lines; at times they would extemporize wildly.

For Microsoft, the most baffling of subplots was the one playing out in the realm of politics. Starting in 1997, a number of Valley figures had begun building bridges to Washington, DC, in a manner unprecedented in the high tech industry. The institutional form this outreach took was a bipartisan organization called TechNet, whose cochairs were Netscape's CEO, Jim Barksdale, and John Doerr, the venture capitalist who had funded not only Netscape but Sun, Intuit, @Home, and an array of other Microsoft rivals, and who was famously tight with Al Gore. In Redmond, suspicions ran rampant that Barksdale, Doerr, and other TechNetters were using their newly minted access in the capital to lobby the administration and Congress to take on Microsoft.

These suspicions weren't entirely unfounded. In the fall of 1997, TechNet had arranged a trip to the Valley for the White House's then-deputy chief of staff, John Podesta, during which executives repeatedly raised the Microsoft issue with him. And according to someone close to the group, on at least one occasion, a Valley figure spoke about it with President Clinton. How did Clinton respond? "He expressed sympathy with our point of view," this person said. "But then, this was Clinton, so it could have been meaningless."

The effects of such lobbying was probably nil. Though Silicon Valley is a rich vein of campaign cash, the politics of pursuing Microsoft were highly fraught. "It's a no-winner," says Greg Simon, a Gore campaign official who once served as the vice president's cyberpolicy guru. "People say, 'Why are you going after them? Do you want to kill the goose that laid the golden egg?'" Still, Microsoft was right to worry that their foes were playing the influence game more adroitly than they were. For if mining the muddy Clintonite middle yielded few tangible (or at least public) results for the Valley, it hit paydirt among those with more concrete ideologies.

On the left there was of course Ralph Nader, but more unexpected, and more influential, was the support the Valley stirred up on the right. Most notably, Netscape and ProComp together reeled in Robert Bork, the conservative jurist whose 1978 book, The Antitrust Paradox, was a sacred text for Chicago School economists and the generation of conservative judges named to the bench by Nixon and Reagan, because of its potent arguments that antitrust enforcement was only justified in the rarest of cases. Bork was initially skeptical of Netscape's complaint – until he took a gander at the first white paper. There he found that Susan Creighton had cited a case in support of her arguments that was also cited conspicuously in Bork's own book. "You're right, I wrote this," Bork said. "And it applies, perfectly."

Among the relatively few executives in Redmond with a background in politics, the Valley's success in nailing down support at both ends of the political spectrum was troubling. As one executive said to me, "If Ralph Nader and Bob Bork agree about Microsoft, my God, there really is no political risk in going after us."

Enter Orrin Hatch. In February, Hatch announced that he was planning to hold a hearing on Microsoft – and to invite Bill Gates to attend. The idea belonged to Mike Hirshland. Assuming Gates showed, the hearing guaranteed adulation from Silicon Valley and a copious quantity of TV time – and thus promised to feed Hatch's twin joneses, for campaign cash and national publicity. To assure that Gates felt he was being treated fairly, Hatch set aside a full hour for a briefing with Gates the day before the hearing, despite his normal practice of never – ever – allotting more than 20 minutes for any meeting.

On the drizzly Monday afternoon of March 2, Gates arrived with an entourage of nearly a dozen at Hatch's first-floor office in the Russell Senate Office Building. The decor of Hatch's digs was classic early-modern senatorial drab – blue carpet, dark wood, flag in the corner. Hatch was on the Senate floor, casting a vote, but soon he strode in and apologized for being late. Gates stared at the clock on the wall, turned to the chair of the Senate Judiciary Committee, and said coolly, "Well, given that we're starting 15 minutes late and I'm only going to have 45 minutes now, we should get right to it."

Hatch, thunderstruck, said nothing.

It went downhill from there. When Gates told Hatch that the DOJ was trying to force Microsoft to remove IE from Windows, Hirshland piped up and said he was mischaracterizing the government's position. Whipping around, Gates snapped, "You don't know what you're talking about." When Gates demanded to see a summary of the questions he might be asked, Hirshland handed him a list of broad categories. Pointing to one topic, Gates wailed, "If you ask about that, this will be a kangaroo court!" Then Gates inquired about the seating arrangements for the hearing. When he was told he'd be seated between Barksdale and McNealy, Gates leapt to his feet and exploded, "No! No! No! If you put me between them, I will not appear at this hearing!"

Hatch, by now more amused than annoyed, leaned back and said, "OK, OK, we'll put you on one end of the table and we'll let you speak first. Happy?"

Compared with the prelude, the hearing itself was a bit of a letdown. Hundreds of gawkers lined up outside to catch a glimpse of Gates decked out like a kid in a wedding – in a suit and tie and decent leather shoes, his hair freshly cut and plastered down. Gates' handlers had studiously prepared him, putting him through mock hearings in which a Microsoft lawyer posed as Hatch and two Microsoft executives played McNealy and Barksdale. Even so, Gates' performance ranged from passable to poor. He was often evasive. He repeatedly contended that Microsoft was not a monopoly, a statement met with pervasive skepticism. And, in the hearing's final minutes, an actual moment of drama arose, in which a dogged (and well-briefed) Hatch was able to extract an admission from him that Microsoft's contracts with Internet content companies barred them from promoting Netscape's browser.

To many observers, and especially those ill versed in the kabuki that passes for communication inside the Beltway, Hatch's convocation seemed to have accomplished little or nothing. But the senator's message wasn't lost on Klein. Two weeks after the hearing, Klein told me, "I knew there was political support for taking on Microsoft. That was not a shock to me. But the Senate hearing provided a real sense of comfort. The politics of this thing were becoming clearer. Microsoft goes up to the Hill and says they don't have a monopoly, and people just say, That's silly."

To other politicos, silly was an understatement. Jeff Eisenach, the head of the Progress & Freedom Foundation, the think tank once known as Gingrich's braintrust, said to me at the time, "When Gates walked out of that hearing, he was a lot closer to a broad Sherman Act case than when he walked in. When you're the richest man in the world and not a single senator speaks up on your behalf, you know you've got problems."

"In all my years practicing antitrust law, I have never seen such powerful people so scared," Michael Sohn told Klein and his team. "It utterly amazed me."

For two of Gates' most outspoken rivals, the Hatch hearing was a day at the circus: the media circus. Barksdale had a ball. Silver-haired and Southern-fried, with courtly manners and a hint of hambone, Netscape's CEO seemed vaguely senatorial himself. He began his opening remarks by turning to the gallery and asking, in his best Mississippi drawl, how many people in the room had a PC. Maybe three-quarters of them raised their hands.

Barksdale asked, "How many of you use a PC without Microsoft's operating system?"

The hands all fell.

"Gentlemen, that's a monopoly."

McNealy, by contrast, seemed a touch nervous. He also committed a thumping faux pas by getting up abruptly in the middle of the hearing and heading for a business meeting in New York. Before he left, however, McNealy snapped off a winner of his own, quipping that "the only thing I'd rather own than Windows is English … because then I could charge you $249 [for the] right to speak English, and I could charge you an upgrade fee when I add new letters like N and T."

Even before the Senate shindig, Barksdale and McNealy had emerged as the public faces of the anti-Microsoft movement. Their companies were loosely but indisputably aligned, despite the feuds that flared sporadically between their employees. Sun was a hardware firm that dabbled in software, and it was far larger and more established than its ally, with $8.6 billion in sales in 1997 compared with Netscape's $533 million. But when it came to the legal campaign against Microsoft, Netscape was the senior partner, both in front of the camera and backstage. It was Netscape's bid to topple the ogre – bold, romantic, inspiring, doomed – that had captured the public's imagination in a way that Sun, even with Java, never had. It was Netscape that was Microsoft's main victim. And it was Netscape, with its white papers and the tireless lobbying by Reback and Creighton, that had finally surmounted the DOJ's inertia and got things cooking in the courts.

Then, on the first working day of January 1998, Netscape announced it had badly missed its fourth-quarter earnings estimates; ultimately, it would report an $88 million loss and fire 400 of its 3,200 employees. At that moment, things changed. While Netscape would remain forever the poster child of the Microsoft case – imagine Marc Andreessen's picture on the side of a milk carton – the pioneering startup was no longer the brains or the heart of the anti-Microsoft coalition. Sun was.

Although McNealy had a reputation as Gates' most caustic and extravagant critic, he assumed the mantle of leadership skittishly. Despite having cultivated a public image as a brash, high-sticking, trash-talking corporate rebel, McNealy's private persona was cautious and conflict-averse to the point of phobia. He was known to be incapable of firing anyone (for that deed he used surrogates) and rarely made decisions without consensus among his senior staff. "His demeanor is radical," Mike Morris, Sun's general counsel, remarked to me. "But his instincts are conservative."

McNealy's instincts were forever at war with his antipathy for Microsoft, which was real and deep and unforgiving. As Sun had transformed itself from an obscure workstation maker into a leading manufacturer of high-end servers, competing with giants such as IBM and HP, some of McNealy's lieutenants, and especially his number two, Ed Zander, had encouraged him to mute his Microsoft attacks. We need détente with Redmond, they said; our customers are begging for it. McNealy was also hesitant about lobbying the government, even on Microsoft, because he didn't believe in it – the government, that is. "Washington, DC, is my least favorite town in the world," he told me at one point. "I see all these unbelievable monuments to government, agencies that have no reason for being on the planet – the Department of Agriculture, Transportation, FEMA, Health, Education, Commerce – all these huge erections of brick and mortar and masses of people running around redistributing wealth. The whole thing drives me absolutely into a freaking funk."

I noted that McNealy hadn't included the DOJ in his list of huge erections. He smiled. I asked what he thought the Feds should do to deal with Microsoft. "Shut down some of the bullshit the government is spending money on and use it to buy all the Microsoft stock. Then put all their intellectual property in the public domain. Free Windows for everyone! Then we could just bronze Gates, turn him into a statue, and stick him in front of the Commerce Department."

Had McNealy's legal theories been all that Sun brought to the anti-Microsoft movement, Redmond could have rested easy. But the company brought Mike Morris, too. A small man with a round belly, a brown beard, and pudding-bowl bangs across his forehead, Morris had been Sun's chief lawyer since 1987. Like McNealy, he was a Michigan native, but they had grown up on decidedly different sides of the tracks – McNealy in posh Bloomfield Hills, as the son of a senior auto-industry executive, Morris in the sticks, as the son of tool-and-die maker. And that was the least of the differences between them. Where McNealy was a blunderbuss when it came to politics, and a libertarian whose tastes in presidential candidates ran to Steve Forbes, Morris was a capital-L liberal with the cagey instincts of a natural-born political consultant. Where McNealy was screamingly heterosexual, Morris was openly gay, and a bit of an activist. And where McNealy shied away from conflict and confrontation, Morris reveled in it, especially when his adversary happened to be Microsoft. It was Morris who had pushed McNealy to file the Java lawsuit in October 1997. After claiming a victory there, he persuaded his boss to file another Java suit, this one more radical, in that it asked the court to order Microsoft to make changes in Windows. In the midst of a furious internal debate over filing the second suit, Ed Zander accused Morris of being a "fanatic."

"I'm not a fanatic, I'm just realistic," Morris said angrily. "We've got our boots on their throats. The right thing to do is to press until they stop breathing. If you're going to strike at the king, you better cut his head off."

Decapitating Microsoft was on Morris' mind again when, a few days into 1998, he picked up the phone and called Joel Klein. For the past nine months, Morris had been in contact with Klein as part of a three-way effort to nudge the government toward a case against Microsoft. His partners in the triad were Netscape's Roberta Katz and Sabre's counsel, Andy Steinberg. Together they had founded ProComp; lobbied the DOJ; assisted Mike Hirshland in his inquiries; told their tale in concert – from the multiple, harmonious viewpoints of a systems company, a software company, and a content company – to anyone who'd listen; and urged wary Silicon Valley bosses to talk in confidence to the DOJ. Now Morris was plotting a solo mission: to put together a sort of private blue-ribbon commission of nationally renowned antitrust lawyers and economists, have them draw up an outline of the kind of Sherman Act case that would make sense for the DOJ to file, including a discussion of possible remedies, and then present the whole thing to Klein and his people.

Might the DOJ find that helpful? Morris asked him.

Sure we would, Klein replied.

So began a project that would span three months and consume $3 million of Sun's money: "Project Sherman." As Morris intended, Project Sherman comprised a superstar group of antitrust authorities, including the famed Houston litigator Harry Reasoner; University of Chicago economist Dennis Carlton and several of his colleagues from the economic consulting firm Lexecon; Arnold & Porter chair and prominent Washington attorney Michael Sohn; Stanford economist Garth Saloner; and former FTC general counsel Kevin Arquit, who handled Sun's antitrust work in Washington. In choosing his experts, Morris took care to select people with impeccable credentials – mainstream credentials, Establishment credentials; the kind of people who spoke Joel Klein's language; the kind who might come across as reasonably objective despite the fact that Sun was paying them $600 to $700 an hour. The political sensitivity of the project was, needless to say, extremely high, for here was one of Microsoft's most ardent competitors bankrolling a costly endeavor to influence the DOJ – an endeavor undertaken with the department's encouragement. And so it was done in utmost secret. Apart from McNealy, Morris informed almost no one at Sun, and the other participants were sworn to strict confidentiality. One of them said to me, "I haven't even told my wife about this."

From mid-January to mid-April, the Project Sherman crew met every two weeks, usually at the O'Hare Hilton in Chicago. At first, the meetings were contentious. For one thing, "There was an awful lot of ego in that room," one person recalls. "An awful lot of grandstanding." For another, the group quickly divided into factions: lawyers and economists; tech-savvy and tech-challenged; Washington insiders and Washington outsiders. "We had these people who claimed to know Joel well," recalls one participant. "They'd say all the time, 'Let me tell you, I know Joel, and Joel will never go for that.'" The FOJ problem was especially nettlesome when it came to the question of remedies. One economist told me recently, "The Washington people kept arguing for conduct remedies because they were so sure Joel would never agree to a structural remedy." He laughed. "Boy, I guess they must be feeling pretty dumb now."

There was another debilitating split within the group. Among those from the Valley, the idea that Microsoft's monopoly and its predatory practices had chilled innovation and distorted investment was taken for granted; it was a given. But for people like Reasoner, Carlton, and Sohn – the big guns, whom Morris intended to wheel out in front of Klein – it was speculation garnished with hearsay. Reasoner kept asking, "Where the hell is the evidence?"

Morris' plan was to bring the Project Sherman gang to the Valley and expose them firsthand to Microsoft's influence. He turned to Gary Reback, asking him to arrange a series of hush-hush meetings with industry figures who could address the question with authority. Nothing gave Reback more kicks than a covert operation where he was pulling the strings. Within days, he had assembled a Murderer's Row of Valley executives, financiers, and technologists who would parade before Morris' group during a single daylong session. Reback told his witnesses that the meeting was important and that it might help influence the DOJ, but he told them little else; not the names of the economists and lawyers they'd be addressing, or who their fellow witnesses would be, or the identity of the meeting's sponsor. To keep the bigwigs from running into one another at Wilson Sonsini's offices, he instructed them to enter and exit through different lobbies.

The tutorial the Project Shermanites received on the appointed late-March day was wide-ranging, and, according to one person who attended, they reacted to certain parts of it with shock and amazement. They heard from Eric Schmidt, the CEO of Novell, about the vulnerability of being a firm that both competes with, and is reliant on, Microsoft software. They heard from the Apple software wizard Avie Tevanian about why conduct remedies like opening up Microsoft's APIs wouldn't accomplish anything. They heard from Sun's Bill Joy (who had no idea that his company was paying for this show) about why Tevanian was right, but why splitting Microsoft into three identical firms, the so-called Baby Bills solution, might be worse: "I keep thinking of 'The Sorcerer's Apprentice.'" They heard from John Doerr about Microsoft's recent habit of gathering together the Valley's VCs and offering helpful suggestions about which technologies it might be advisable to invest in and which might be best left to Redmond. "My firm's policy is never to back a venture that competes directly with Microsoft," Doerr said. "Only damned fools stand in the way of oncoming trains."

And they heard from Jim Clark. "When I left Silicon Graphics I had a net worth of $16 million and I invested $5 million to start Netscape," Clark said. "Microsoft has practically killed Netscape. I'll never invest in another thing to compete with them. I'll never touch another market that has anything remotely to do with Microsoft's path. And if I'd known four years ago what I know now – that Microsoft would destroy us and that the government wouldn't do anything about it for three fucking years – I never would've started Netscape in the first place."

A few weeks later, Morris and a select subset of his experts (big guns plus Saloner; no Reback) flew out to Washington for their audience with the DOJ. It was now the middle of April. Four months had passed since the consent-decree case had climaxed, and Morris knew little more about where the DOJ's investigation stood than what he read in the papers. Certainly the trustbusters seemed eager to see him: Klein had called twice to try and move up the date of the presentation, and, arriving at the DOJ, Morris found himself playing to a packed house. Klein, his number two Doug Melamed, Rubinfeld, Malone, and Boies were there, along with a swarm of junior antitrust-division staffers, all crowded into the conference room next to Klein's office.

Taking seats across the table from Klein and his deputies, Morris' team proceeded to outline the case they believed the DOJ should file. Just as the Netscape white papers had argued, the core of that case was illegal monopoly maintenance and monopoly extension – a violation of Section 2 of the Sherman Act. For years, Microsoft had leveraged its power over the desktop to invade adjacent markets, from productivity applications to server operating systems. Sometimes those markets were tremendously valuable in their own right; Office alone raked in billions each year for Gates' company, and Microsoft's next target – the server space in which Sun was a leader – was even richer. Other times, the market itself was worth next to nothing in terms of dollars and cents, but controlling it was essential to preserving Microsoft's dominance on the desktop. Browsers were one example of this. But Java was an equally compelling one. By letting programmers write software that would run on any OS, Java threatened to render Windows irrelevant, if not obsolete. Microsoft's response had been to license Java from Sun and then violate that license by creating a Windows-only variant of the technology in an attempt to subvert its cross-platform purpose. With both Java and the browser, as Saloner put it later, Microsoft's MO was the same: "We will embrace it, we will make it ours, we will apply it to our operating system, and we will kill it. We will do what we must to protect the mothership – the OS."

A Microsoft executive had made a suggestion about what Apple should do with its multi- media technology QuickTime: "Knife the baby."

The Sun presentation ran for nearly four hours. Deploying his experts to make most of the arguments – Reasoner and Sohn on the law, Carlton on the economics – Morris tried to anticipate and shoot down Microsoft's defenses. In particular, the team addressed the question of harm, of who'd been hurt by Microsoft's actions. After all, the company would say, consumers are happy; prices are falling; high tech is thriving; so is Sun, by the way. What that picture left out, however, was the damage to innovation – the products left undeveloped, the areas of technology left unexplored. For example, there was almost no R&D on operating systems anymore. What did that imply for the future of technology? And how long could innovation continue to flourish in an industry suffused with fear?

"I went out to Silicon Valley," Mike Sohn told Klein and his team. "In all my years practicing antitrust law, I have never seen such powerful people so scared. It utterly amazed me."

At the end of the afternoon, the talk turned to remedies, and Dennis Carlton took the floor. In a way, Carlton was the least likely, and thus the most impressive, member of the Sun team. One of the best-regarded economists in the country, he was also a classic conservative straight out of the Chicago School: suspicious of plaintiffs, friendly to business, inherently skeptical of government intervention in general and antitrust enforcement in particular. All of which was why Morris had worked doggedly to recruit him in the first place.

Throughout the day, Carlton had spoken with calm conviction about the economics of the case; about monopoly maintenance, market power, and Microsoft's predation. Now, with the DOJ officials hanging on every word, Carlton did what had once been nearly unthinkable. First, he laid out a range of conduct remedies (contract restrictions, technical requirements) and methodically described the pros and cons of each, in every case listing more cons than pros. Then, without the slightest hesitation, he presented the case for a structural remedy – not a full-blown breakup of Microsoft, but a scheme that would force the firm to license all its intellectual property to some number of third parties, giving birth to a set of clone companies that would create competition in the markets for operating systems and applications.

Garth Saloner knew it was coming, but even he found it a powerful moment. "This is not one of us Silicon Valley loonies saying this," Saloner later observed. "This isn't Gary Reback. This isn't Roberta Katz. This isn't Garth Saloner. This is Dennis Carlton. Things have moved. The world has changed. If you're Joel Klein or Dan Rubinfeld, I would think you'd take comfort in that."

Mike Morris had no illusions that Klein and his colleagues would swallow the case his team put forward – let alone the remedy – whole. Instead he was trying, as he put it, simply "to give them a sense that this wasn't a wild goose chase; that this was a good case, a real case."

As the meeting drew to a close it was impossible to know if the effort had succeeded. For four hours, the DOJ officials had maintained what one participant described as a "highway-patrol demeanor": professional, poker-faced, pristinely neutral. They had asked countless questions but given nothing away.

But many months later, when I asked Rubinfeld about the Sun presentation, he answered in a way that would have made Morris' day: "It was memorable. It was impressive. It told us some things that we did not know. But mostly, and this can't be underestimated, it reinforced in our minds that what we were doing wasn't crazy."

What the DOJ was doing was girding for war. By mid-April, Klein had persuaded David Boies to sign on as the antitrust division's "special trial counsel" for about one-fifteenth of his customary $600-an-hour fee. ("It didn't take a lot of persuading," Klein recalls. "About a half a second after I asked, he said, 'When do I start?'") Klein also brought another pivotal player into his inner circle: Jeffrey Blattner, a former chief counsel to Ted Kennedy's staff on the Senate Judiciary Committee who'd made a reputation in Washington as a sharp operator during the bloody 1987 battle to keep Robert Bork off the Supreme Court. Blattner's title was Special Counsel for Information Technology, but his de facto role was chief of staff for the Microsoft case, with duties that would include stroking the Hill, spinning the press, and plugging any (unwanted) leaks from within the division.

In short, all the smoke signals wafting out of the DOJ's Pennsylvania Avenue headquarters indicated that Klein was on the verge of filing a broad Sherman Act suit. The only questions were: How broad? And to what end?

To find out, I arranged to meet Klein on the Saturday morning after the Sun presentation. It was a brilliant spring day, with Washington aswirl in cherry blossoms and dogwood. Over the next two years, Klein and I would have nearly a dozen of these discussions, all on the understanding that nothing he said would appear in print until after the trial was done. The setting was always the same: Klein's fourth-floor corner office, where he would sit in a high-backed leather chair, dressed usually in a dark suit and tie, and talk for an hour or two about the strategy, tactics, and legal principles at stake in a case he believed would help set the rules of competition for the digital age. He spoke quickly, quietly, candidly, and not without humor, in a voice still tinged with the accents of Astoria and Bensonhurst.

"I think we're at decisionmaking time," Klein began, pointing out that the introduction of Windows 98 was only a few weeks away. After months of concentrated investigation, Klein was satisfied that he had sufficient evidence to level a number of charges against Microsoft: that its exclusive contracts with ISPs and content providers were anticompetitive; that its contracts with OEMs placing "first-screen" restrictions on how they could modify the Windows desktop and boot-up sequence were illegal; and that its integration of IE with Windows constituted an unlawful tying of two separate products together. In all this, Klein said, the company's motives were clear, and clearly predatory. "When you see document after document, from Gates on down, saying that Netscape could basically commoditize the operating system, that's important stuff," he said. "That's what was going on in the minds of these people when they say, Well, what we ought to do in response is go right at 'em and cut off their oxygen."

Klein felt confident that each of these tactics was a violation of Section 1 of the Sherman Act, which states: "Every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations, is declared to be illegal." The question was whether to go further and accuse Microsoft of monopoly maintenance under Section 2. Section 2 says: "Every person who shall monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States … shall be deemed guilty of a felony."

For all the exhortations of Netscape and Sun, monopoly maintenance was not a mainstream case to bring. And it wouldn't be a simple one to prove, especially given the products in question. To start with, explaining how the combination of Java and Netscape's browser, neither of them a direct rival to Windows, nonetheless posed a threat to the OS would require the DOJ not only to elucidate the intricacies of software APIs but to do so in a vocabulary Judge Jackson (who, having presided over the consent-decree case, would handle any related Sherman Act suit) could readily comprehend. No mean feat, that.

Inside the DOJ, a pitched debate was still being waged among those who preferred to keep it simple, to stick with a more traditional Section 1 case, and those angling hard for Section 2. Just as he had during the run-up to the consent-decree suit, Dan Rubinfeld led the hawks, although now he had backing from the hardline Boies. "Contrary to what a lot of other economists and lawyers in the division believed, I thought it would be perhaps easier to win a bigger case than a narrower one," Rubinfeld recalled. "What we had with Microsoft was a pattern of practices where the whole was greater than the sum of the parts." And although the majority of misdeeds the DOJ had nailed down so far – including the June 1995 meeting between Microsoft and Netscape – revolved around the browser, the investigation had begun to unearth evidence of Microsoft malfeasance involving other competitors. "We hadn't had time to flesh out the pattern of bad acts completely," Rubinfeld said. By filing a Section 2 claim, "we could put a legal placeholder in our complaint and try to fill it in later. If we could stand it up, the case would be broad. If we couldn't, it would be a browser case."

For nine months, Klein had heard from every putative Microsoft victim in the known world. He'd heard tales of treachery, duplicity, and outright thuggery. He had watched as the mood in Washington had turned decisively against Gates and his company. And yet, far from spoiling for a fight, he still seemed wary, slightly gun-shy. On the question of what sort of remedy he might seek, Klein expressed a preference for something "surgical." Did that mean he wasn't considering a breakup? "I think that's accurate – at least for now," he replied. "There are real costs that one has to be very cognizant of to breaking up a company like Microsoft."

I asked Klein if he'd ever met Gates, and he said that he hadn't. Was he looking forward to that day?

"I don't know. People ask me this a lot. Maybe it reflects a blind spot. I mean, obviously there's something about meeting Bill Gates – though, as my kids would tell me, it's not as exciting as meeting some rock star. I feel weird, because I have the sense that everybody expects there's gonna be this great day. But I don't personalize this stuff. I really don't."

"I am one happy camper," Klein whispered to me in the back of the hall at the Agenda conference. "We really kicked their butts."

The great day arrived just two weeks later, when Gates and Bill Neukom traveled from Seattle to Washington for a summit with Klein and his lieutenants. The Justice Department had put Microsoft on notice that it intended to file suit sometime before May 15, the Windows 98 ship date; at least a dozen state attorneys general were prepared to do the same. Now it was time to offer the soon-to-be defendant a final opportunity to reach an accord outside of court – a meeting known inside the antitrust division as "last rites."

And so on the evening of May 5 the two camps convened in the offices of Microsoft's main outside law firm, Sullivan & Cromwell, in an eighth-floor conference room with windows looking out on the Old Executive Office Building. On Microsoft's side of the table sat Gates, Neukom, and a pair of S&C attorneys; on the DOJ's side were Klein, Boies, Blattner, and Melamed. Typically, when a company and the government get together in an effort to avert a massive lawsuit, the tenor of the discussions is all about give and take, with each side laboring, however misguided it thinks the other is, to find common ground. But Gates' approach "was more in the nature of a lecture – the world according to Gates – than a constructive dialogue," Klein told me later. For the next two hours he held forth – forcefully, passionately, and often patronizingly – about the nature of the software business and the needs of his company. He asked no questions of the DOJ, and his answers to theirs took the form of prolonged soliloquies.

In the world according to Gates, the notion that Microsoft had a monopoly was ludicrous. "Give me any seat at the table – Java, OS/2, Linux – and I'd end up where I am," he proclaimed. "I could blow Microsoft away! I'd have programmers in India clone our APIs. If you were smart enough, you could do it." Asked if Netscape's browser was designed to compete with Windows, Gates shot back, "Not compete. Eliminate."

When the DOJ team tried to get Gates to address the full range of their concerns – the exclusive contracts, the first-screen restrictions – he repeatedly brushed them off, returning again and again to a single issue: integration. Klein recalled, "He made the argument in myriad different ways that the future of technology was through product integration; that he'd put billions of things into the operating system and he needed to be able to keep putting whatever he wanted into Windows. And if the government blocked that, it would basically kill his business. That was the clear top line, bottom line, and every line in between." What startled Klein was the personal terms in which Gates made these points. "It wasn't just, You're going to kill my business; it was, You're going to kill me. And clearly we, the government, were the instrument of this great personal affliction."

As David Boies watched Gates give no quarter, he couldn't help thinking that the CEO was dangerously underestimating his adversary. From his decade-long stint in the trenches of the IBM case, Boies knew well that the DOJ was not just another adversary; that it had "the same resources, the same imperatives, the same commitment" as any corporation, no matter how resolute. It was a point worth making, he told himself. So as the meeting drew to an end, Boies looked across the table at Gates and Neukom and asked if he could offer a word of advice.

"You know," Boies said, "once the United States government files suit against you, everything changes. People you thought you could trust turn against you. People you thought were your allies turn out to be enemies. Everyone is more willing to question you, to resist you. The whole world changes."

Gates and Neukom stared back blankly. "The government kept making these melodramatic statements," one of Microsoft's senior lawyers said. "They just didn't understand the fundamentals of our business. It was a bit like two ships passing in the night."

Initially, Klein felt the same way. But as he turned the meeting over in his mind afterward, he began to discern in the contours of Gates' intransigence what he thought were the faint outlines of a settlement. Microsoft seemed to be signaling that the first-screen limitations and restrictive contracts meant little to it. Maybe, if the company were willing to give significant ground there, and if the DOJ showed flexibility on product integration, a deal could be done that would satisfy both parties.

For the next nine days, Klein and Neukom burned up the phone lines with proposals and counteroffers. From the Microsoft side came a series of concessions to loosen the firm's grip on the first screen and give OEMs greater freedom over the Windows desktop. The company also offered a variety of ideas – a "browser folder," perhaps, or a "ballot screen" whereby users could choose between IE and Navigator – to create a more level playing field for Netscape. Indeed, at 1:30 am on the Thursday morning that the DOJ was set to file suit, Gates himself phoned Klein at home to discuss whether Microsoft might agree to a "must-carry" provision wherein it would ship Netscape's browser with every copy of Windows. A few hours later, after another conversation with Neukom, Klein decided to delay launching the suit until the following Monday, so that Microsoft and the DOJ could devote the weekend to face-to-face negotiations.

In Silicon Valley the sound that greeted Klein's announcement was the gnashing of high tech teeth; in Washington, it was the low murmur of cynical assumptions being confirmed. What the Valley had long feared and the political class had long expected finally seemed to be coming to pass: At the eleventh hour, Joel Klein was caving. And although that judgment was too harsh, at its core was a kernel of truth: Klein wanted a settlement and he wanted it badly.

The reasons were almost too numerous to count. In suing Microsoft, Klein would be taking on a company with unlimited resources and the best legal talent that money could buy, not to mention a PR operation populated by literally hundreds of footsoldiers, strategists, and high-priced ad gurus. For all the tarnishing Gates' image had recently suffered, Microsoft's CEO remained an icon of the new economy. Even for a man more daring by nature than Klein, the political and legal risks of challenging Gates would have been daunting, the rewards uncertain. If he settled the case, Klein could declare victory and go home. The victory would be limited, but it would also be immediate – no small thing in an industry running on Internet time. And it would avert a protracted lawsuit in which the government's prospects were decidedly murky. A month earlier, the DOJ and Microsoft had argued the appeal of the consent-decree case before a three-judge panel on the US Court of Appeals for the District of Columbia, and the judges had seemed markedly hostile to the government's position. As for the broader case the DOJ was about to unleash, the antitrust Establishment (Mike Morris' experts notwithstanding) regarded it as a shot in the dark.

Even David Boies had his doubts. "At that point, we didn't have all the evidence that we would subsequently get," he told me later. "We had some evidence of Microsoft's broader conduct, but they denied that conduct. We had a lot of things we believed, but whether we could ultimately prove them or not was very uncertain. We had a judge who we thought was a good judge, but he was a careful judge, a very conservative judge. We knew he was going to make us prove every element of the offense. So we were in a situation where, if we could have achieved anything like a reasonable settlement, I think we would have jumped at it."

Dan Rubinfeld remembers thinking that Microsoft could have capitalized on the DOJ's ardor. "If I had been free to give them advice, that was the moment I would have said, 'Look, this is the time. Do a deal with us. You know me. You trust me. Really. Do it.'"

Instead, Neukom flew back to Washington, sat down with the DOJ and the states on a Friday afternoon, and played a brand of hardball that quickly brought the negotiations to a halt. Not long into the first session, it became clear to the government that certain compromises Microsoft had already offered – in particular, ceding power over the desktop to OEMs – were now being yanked off the table. If that was so, there was not much to talk about. For Microsoft's part, one of its top attorneys said that the government's "basic attitude throughout was arms-folded, we-need-more, we-need-more. They made no counteroffers. We were not difficult or nonchalant. We tried our darnedest."

Late on Saturday morning, Neukom drafted a memo laying out Microsoft's stance (which included dropping its restrictive contracts, adopting the browser "ballot page," and not much else) and handed it to Jeff Blattner. Blattner, who was leading the DOJ's negotiating posse, could see the talks were about to fall apart and suspected Microsoft might leak the memo to the press. Pushing it back across the table, he said flatly, "I don't negotiate from a list." Roughly translated, that meant sayonara.

In retrospect, Microsoft's failure to settle seems a colossal and inexplicable blunder. Even putting aside its mysterious retreat from its first-screen concessions (Had Neukom gotten ahead of Gates? Had Gates himself had a change of heart? Had the DOJ misunderstood the company's previous offers?), there were any number of other solutions at hand. In the consent-decree case, for example, Microsoft had agreed to offer OEMs two versions of Windows 95, one with IE visible, the other with it hidden; already it was clear that most of the OEMs were taking the version that the company preferred. Had Gates proposed that the same arrangement be applied to Windows 98, the company would have sacrificed little in business terms and conceded nothing about its future right to integrate features into the operating system. Meanwhile, the government would have been hard-pressed to spurn the offer, as its officials acknowledge today. Yet, when I raised the point with Gates, Neukom, and the rest of the Microsoft legal team, they said uniformly that this entirely obvious idea had never been entertained by the company; and that, even if it had, the DOJ would never have accepted anything less than forcing Microsoft to carry Netscape's browser.

There is, however, an alternative explanation: that despite the Sturm und Drang of those 10 days in May, Microsoft's aim in the settlement negotiations was something other than settlement. "It was a fishing expedition," Christine Varney, Netscape's Washington counsel, opines. "They wanted to find out what was in the case. When you're a litigant, you want to know as much as possible about what you're facing – if there's some smoking gun that you don't know about. So you find out, then you recalibrate and decide whether to settle or not."

What Microsoft found out – or thought it found out – was that the suit the DOJ intended to bring wasn't nearly as sweeping as the company had feared. To Microsoft's lawyers, it sounded like a browser case, a tying case, and tying was the legal ground on which they believed that their standing was firmest. "They thought, 'This is going to be a narrow case, so let's fight it,'" Boies told me. "'If we lose, we lose a narrow issue. We can afford to fight this case and lose.'" He went on, "Also, remember that Microsoft had been fighting with the government in one way or another for almost 10 years. And every time, they'd managed to come out really well. I think they thought they were smarter than we were. I think they thought they knew more than we did. And both of those things may very well have been true. But I think they underestimated our ability and willingness to learn."

Microsoft wasn't alone in its view that the government's case was a narrow one. When the DOJ and 20 state attorneys general filed suit on May 18, the Monday after the settlement talks collapsed, the complaint charged Microsoft with four counts of violating the Sherman Act: exclusive dealing and unlawful tying under Section 1; monopoly maintenance in the OS market and attempted monopolization of the browser market under Section 2. Yet the narrative that Klein spun around the case painted Netscape as its hero and victim, and the short-term remedy the DOJ was seeking was inescapably Netscape-centric: a preliminary injunction forcing Microsoft either to offer a version of Windows 98 without IE or to bundle Navigator with the OS as well. Netscape was thrilled: It certainly looked like a browser case to Jim Barksdale. Sun was disconsolate: It looked like a browser case to Mike Morris too. And the rest of the Valley rolled its eyes: Didn't the government realize that the browser war was over? "If they'd done two years ago what they did today, it might have been useful," Gary Reback groaned from a New York phone booth. "It's been a long haul to get this far. It's going to be a long march to get where we need to be. And some of us are getting awfully tired."

Some of them were worse off than that. Since the early days of Reback's efforts, no one in government had been a more steadfast ally than Mark Tobey. The assistant attorney general from Texas had got the ball rolling with his Netscape depositions and then lobbied furiously to create a groundswell among the states. But a few days before the Sherman Act case was filed, Texas had been forced to withdrawal its support, under pressure from the state's computermaking kingpins, Compaq and Dell. Because both companies were reliant on Microsoft, the widespread assumption was that they were acting on orders from Redmond. Tobey told Reback, "I never dreamed they'd be able to shut me down entirely."

Then came another blow to the anti-Microsoft movement, a development that also plunged the DOJ into sudden despair. On June 23, the federal appeals court issued its ruling in the consent decree case. Striking down Judge Jackson's preliminary injunction, the Appeals Court found that he had "erred procedurally," by not giving Microsoft a chance to contest the injunction, and "substantively," by misreading the law on tying. "Antitrust scholars have long recognized the undesirability of having courts oversee product design, and any dampening of technological innovation would be at cross-purposes with antitrust law," the court's opinion read. "We suggest here only that the limited competence of courts to evaluate high tech product designs and the high cost of error should make them wary of second-guessing the claimed benefits of a particular design decision."

In Microsoft's eyes, it was an overwhelming victory. The next morning, Gates picked up The New York Times and read that even Joel Klein's allies shared Microsoft's assessment of the Appeals Court's ruling. "This cuts the legs out from under the Justice Department on their new case," the former DOJ antitrust official Robert Litan was quoted as saying. "It's potentially devastating."

McGeady was "quarantined" from the rest of Intel. "Nobody even told me I was on the witness list. I read about it in my underwear in The New York Times."

For the first time in months, reading the Times made Bill Gates smile.

David Boies was smiling too, though it made his associates think he'd gone crazy. By common consensus, Boies was the most brilliant litigator of his generation. Before setting up his own shop in 1997, he had spent 30 years with the York firm Cravath, Swaine & Moore after graduating second in his class from Yale Law School. Over the years, Boies had represented a vast assortment of splashy clients against an array of even splashier opponents. In addition to his antitrust work for IBM, he had defended CBS against a takeover bid by Ted Turner and a libel suit by General William Westmoreland. He had helped Texaco fight off the corporate raider Carl Icahn and helped Westinghouse take on Philippine president Corazon Aquino. On behalf of George Steinbrenner, he'd sued Major League Baseball; on behalf of the government, he'd sued Michael Milken. He rarely lost at trial and had never had a victory overturned on appeal.

In his mid-fifties, Boies had thinning brown hair, a flat Midwestern twang, and a downmarket demeanor (Lands' End suits worn with blue knit ties he bought by the bagful) that belied an annual paycheck of more than $2 million. His courtroom manner was casual and conversational, which tended to lull his adversaries into a fatal haze of complacency. His memory was borderline photographic; his competitiveness, modestly terrifying. To a colleague at Cravath he once uttered the words that will surely be his epitaph: "Would you rather sleep or win?"

Boies got hold of a copy of the Appeals Court decision just before boarding a flight from New York to San Francisco. By the time the plane landed he was certain that, far from being a death knell, the opinion actually worked to the DOJ's advantage. "It helped in three ways," he told me later. First, though the court was plainly on Microsoft's side, it made no bones about the fact that the company had a monopoly. Second, Boies said, when it came to tying, "The court said that if you can prove that they don't need the tie to achieve the benefits, then that's just bolting two products together, and that violates the tying laws." Third, he went on, "The court said that if you can prove that they did it not for efficiency purposes but for anticompetitive purposes, that trumps everything. In other words, the court was adopting an intent standard, and, given the Microsoft documents we had in hand, that was a standard I thought we could meet."

The Appeals Court had effectively provided Boies with a kind of road map, a guide for framing his arguments on product integration. At the same time, what the court said didn't touch the DOJ's Section 2 claims – claims that Boies now believed it was essential for the government to beef up before the trial began. But more than all that, the ruling gave Boies an overarching sense of confidence about the case in general. "Even in a decision that was really quite pro-Microsoft, there was no hint that the court was saying 'The antitrust laws don't apply here; we're going to give the software industry or Microsoft a free pass,'" he noted. "And once I knew they didn't get a free pass, I knew we could prove an antitrust violation."

To do that, however, and especially to flesh out a powerful case of monopoly maintenance, Boies would need witnesses – strong and credible ones. And, unfortunately, the DOJ would have precious little time to round them up. When the government filed suit, both sides had assumed that Judge Jackson would hold a quick hearing on the DOJ's request for a preliminary injunction and then schedule a full trial to start perhaps a year later. But apparently Jackson had other ideas. In a surprise maneuver, he decided to put aside any preliminary hearing and move directly to trial – and soon, setting a date to commence in early September. If Jackson had his way, US v. Microsoft would be short and sweet. To ensure that it was, he adopted an unusual procedure limiting each side to 12 witnesses, all of whom would deliver their direct testimony not on the stand but in writing, with courtroom hours being reserved strictly for cross-examination.

The accelerated schedule set the DOJ a formidable task: Klein and his team had the summer months to chase down all the allegations of misconduct they had heard, establish their veracity, and then persuade a reasonable number of the aggrieved parties to step forward, under oath and in the glare of a high-profile trial, and testify. Shortly after the Appeals Court decision came down, Gary Reback had breakfast with Klein in Washington and found the assistant attorney general in a nerve-addled state. "We've filed this thing," Klein told him, "but we have no witnesses."

"If I were Joel, I'd have been pissing my pants right then," Reback recalls. "The judge said 12 witnesses. I kept looking at my fingers and thinking, how are we ever going to get there?"

In the tech world, the memory of the 1995 consent decree, seen universally as a dismal failure, remained fresh. And even with the Sherman Act case, the current crop of trustbusters hadn't gone very far toward restoring the industry's confidence in the DOJ. "There was a lot of trepidation in the Valley about whether the government was capable of getting any of this right," Reback remembers. "Nobody wanted to get anywhere near this thing. Nobody wanted to be subpoenaed. Nobody was sure they could pull it off."

Reback, naturally, joined in the hunt for witnesses; so did Orrin Hatch and Mike Hirshland. Klein's team of 20-odd lawyers talked with dozens of companies in Microsoft's crosshairs. Software companies and hardware companies. Internet tyros and Fortune 500 stalwarts. They talked with Yahoo!, Excite, RealNetworks, Palm. And with most of the OEMs – Compaq, Acer, Gateway, Packard Bell, HP, Sony. Yet by the middle of July, the DOJ's witness list was so barren that Klein told me he was considering filling a quarter of his slots with Netscape executives, another slot with someone from Sun, and many of the rest with economists and technical experts. He didn't have much choice. After several weeks of tree-shaking, the DOJ's efforts had produced a meager harvest.

And then, quite suddenly, some fruit began to fall.

It began with Intuit, whose CEO, Bill Campbell, had spent much of his estimable career at the sharp end of the Microsoft stick. In the 1980s, Campbell worked at Apple and helped launch the Macintosh, and then became CEO of the doomstruck pen-computing company GO, whose executives claimed Gates first stole their ideas and then muscled OEMs to keep them from allying with the startup – nixing a key deal with Compaq, in particular, when GO was on the brink of bankruptcy. At Intuit, Campbell and board member John Doerr (who backed both Intuit and GO) were, in Campbell's words, "the last holdouts" against the plan by the firm's chair, Scott Cook, to sell Intuit to Microsoft back in 1995. After the DOJ scotched the deal, Microsoft waged a fierce campaign to topple Intuit in the financial-software market. Against the odds, Campbell prevailed, doing whatever was necessary – including abandoning an alliance with Netscape – to keep Intuit's place on the Windows desktop.

The DOJ had long believed that Intuit had a story to tell. In its court papers, the government cited a Microsoft email in which Gates wrote, "I was quite frank with [Cook] that if he had a favor we could do for him that would cost us something like $1M … in return for switching browsers in the next few months I would be open to doing that." But Campbell wanted no part of the DOJ. He regarded its lawyers as woefully overmatched ("I told them, the Bill Neukoms of the world are going to cream you government pantywaists") and the suit's short-term remedy as worse than meaningless ("They've got to put both browsers in the OS? Great. Now I have to pay double ransom"). Then that summer he got a call from Mike Hirshland, who told him the DOJ had hard evidence Microsoft had indeed killed the Compaq deal that might have saved GO. He also got a call from Hirshland's boss. "You know damn well there is some unethical behavior out there that's possibly illegal," Hatch said. "The only way we can expand the case is if people like you are willing to talk."

Campbell thought about it, and by late July, he was ready to raise the subject with Intuit's board and its top executives. In favor of putting someone forward to testify was John Doerr, who argued, "If we feel we're getting screwed, we ought to say so." Against it was Cook, who said that helping the government would be an admission of defeat; it would put Intuit on a par with the Valley's congenital complainers. At the end of the three-hour meeting, a vote was taken: Everyone except Cook agreed that Intuit should testify. For Campbell, a former college football coach, it all came down to a question of cojones: "I thought, goddammit, we ought to be strong enough to stand up and be counted."

Just around the time Campbell was climbing on board, the DOJ caught another big break. The government investigators had been trying for months, without much luck, to nail down rumors that several years earlier Microsoft had strong-armed its ally Intel over Intel's plans regarding the Internet. Now, as the DOJ was taking depositions from various Netscape officials, Jim Clark recalled that an Intel executive named Steve McGeady had once told him about a meeting in which Gates had declared his intent to "take Netscape's air." Clark shot off an email to McGeady asking if he'd be willing to talk to the DOJ. McGeady wrote back almost instantly, correcting Clark's memory (it was Paul Maritz, not Gates, who'd alluded to Netscape's impending lack of oxygen) but adding, "If the DOJ asks me to testify to that effect, I will, without hesitation." In short order, the government arranged to depose McGeady.

The DOJ should already have been aware of Steve McGeady. Three years earlier, on a tip from Reback, the antitrust division had sent him a CID for documents concerning a clash between Intel and Microsoft over an Intel software technology called Native Signal Processing. But like the ark of the covenant at the end of the first Indiana Jones movie, the NSP documents had apparently been buried somewhere deep in the bowels of the DOJ, and the entire issue had faded from its collective memory – and from Intel's as well. "Four days before my deposition, I say to my Intel lawyer, I assume you've reviewed these documents from 1995," McGeady told me. "He says, 'What documents?' He doesn't know. So he calls the Justice Department. They don't know either!" McGeady rolled his eyes. "It was like the Keystone Kops do antitrust."

McGeady's deposition was dynamite stuff. At the same time, the DOJ's dealings with Intel were wary and delicate. For nearly 20 years, Intel and Microsoft had collaborated so closely that they were often regarded as a unitary being: "Wintel." The moniker was misleading, for the relationship was riven with fractures and fissures; Andy Grove liked to refer to the companies not as strategic partners ("I really hate that phrase," he snarled) but as "fellow travelers" – not soulmates, but seatmates on the same train. But because Intel was hugely dependent on Microsoft, and vice versa, keeping peace with Gates was one of Grove's prime priorities. So when Intel eventually confirmed that McGeady would be testifying in the trial, the company took pains to assume a posture of perfect neutrality. McGeady was not being "sent" to testify; he was merely being "allowed" to testify. What choice do we have? Intel said, in effect. The government wants him; we can hardly refuse.

Behind the scenes, Intel's neutrality was far from perfect. With the stealth and finesse of an accomplished Byzantine courtier, the company's general counsel, Peter Detkin, was helping drive the stiletto into Microsoft's back. Detkin, a former partner at Wilson Sonsini, had been a longtime colleague of Gary Reback. There was no love lost between the two men, but over the years Reback had conducted what he called "deep-throat meetings" with Detkin and other Intel lawyers in the bar at Hyatt Rickeys in Palo Alto. When the government started asking about McGeady, Detkin turned to Reback and Susan Creighton as a covert back-channel to the DOJ. "Peter used Wilson Sonsini as a safe conduit to pass information to the government," a lawyer close to the situation told me. "The nature of the information was: If you look here, or here, or here, you'll find something interesting."

News that the DOJ had deposed McGeady hit the Valley like a thunderbolt from a clear blue sky. If Intel was cooperating with the government (as everyone assumed it was, no matter what the company was saying), then the DOJ's case was undeniably gathering steam. With Intel and Intuit on board, Boies was able to lock down witnesses from two companies at which he had close connections: IBM, where the ancient Microsoft hatreds still burned, and AOL, whose head of government affairs, George Vradenburg, had years earlier hired Boies to handle the Westmoreland libel case when Vradenburg was in-house counsel to CBS.

McGeady's testimony took on the flavor of a software-world Scenes from a Marriage : Microsoft clearly wore the pants in the family, while Intel played the part of the long-suffering spouse.

The DOJ got another boost with Judge Jackson's decision in mid-September to delay the trial's start until mid-October. The extra month would buy the DOJ some breathing room. It would also provide a chance to go after the most glittering prize of all: Steve Jobs and Apple.

The DOJ's interest in Apple was twofold. First, there was the headline-grabbing deal between Cupertino and Redmond in August 1997, in which, the government believed, Microsoft had threatened to cancel Office for the Macintosh unless Apple replaced Navigator with IE as the Mac's default browser. Then there was multimedia. The DOJ had recently received from Reback another of his patented white papers, this one focused on the Apple multimedia technology QuickTime. The white paper alleged that over the previous two years Microsoft had engaged in a passel of predatory tactics to stifle QuickTime – tactics that loudly echoed its approach to Netscape's browser. According to the Reback document, Microsoft had proposed to carve up the multimedia market with Apple; it had then pressured OEMs to drop QuickTime; it had inserted technical incompatibilities that disabled QuickTime in Windows; and it had struck exclusionary deals with content providers to develop only for Microsoft's competing NetShow technology. At one point, a Microsoft biz-dev manager had made a suggestion about what Apple should do to QuickTime which was so irresistibly colorful that Reback made it the white paper's title: "Knife the Baby."

In the autumn of 1998, Apple's recovery under Jobs was still fragile, its relationship with Microsoft forever precarious. If the DOJ had any prayer of persuading the company to throw caution aside and sign up for the trial, Reback was clearly the man to see. In the mad scramble for fresh evidence and plausible witnesses, whatever lingering resentments Klein harbored toward the monomaniacal lawyer had receded. Reback was simply too useful, too plugged-in and switched-on, to be ignored. In a series of phone calls that September, Klein told Reback that he desperately wanted the Apple story to be part of the trial – and he wanted Jobs to be the one to tell it. Though the DOJ's witness list was shaping up nicely, Klein was concerned that it lacked star power, featuring as it did only one big-name CEO – Jim Barksdale. Klein told Reback, "We have an uuml;bermenschen problem."

Jobs was certainly uuml;ber, but no one had ever accused him of being a mensch. Visionary, volatile, volcanic, and vain, Apple's CEO had made no secret of his skepticism about the DOJ's capacity to prosecute Microsoft. "The government is bullshit! The government is bullshit!" he'd barked when a government lawyer visited him that spring to ask for his help in building the case. "You guys have done nothing, you haven't figured it out, you've been too slow, you'll never change anything. This is an incredibly sensitive time for Apple. Why should I jeopardize the future of my company when I have no faith that the government is going to do anything real?"

To Jobs, "real" meant one thing: breaking Microsoft up. For all his doubts about the DOJ's competence, he was now grudgingly impressed by the government's progress. In late September, after several lengthy talks with Reback, his friend Bill Campbell, and a number of DOJ intermediaries in the Valley, Jobs agreed to have a conversation with Klein about the possibility of testifying. When the two men connected by phone, with Jobs on vacation in Hawaii, he wasted no time in getting to the point. He wanted to hear Klein's thinking on remedies.

Are you going to do something serious? Jobs demanded. Or, he asked, "Is it going to be dickless?"

At the other end of the line, Joel Klein squirmed. Even if he had settled on a remedy, which he manifestly hadn't, it would be grossly inappropriate to discuss it with Jobs – or with any other Microsoft competitor. Klein told Jobs this. He told him he could offer him no commitments, no promises. Klein said, "It's a chicken-and-egg problem; the power of the remedy will be determined by the quality of the case."

Jobs was singularly unimpressed, and he let Reback know it. Frustrated, irritated, Reback called Mike Hirshland to commiserate. "Joel blew it," Reback sighed. Jobs hadn't needed a firm commitment. He needed to be sold on the notion that the DOJ was, à la Microsoft, hardcore about the case. But Klein hadn't been selling; he'd been legalistic, stilted, excessively circumspect. He'd been … Joel.

As he listened to Reback moan, Hirshland had a brainstorm. Why not have Boies give Jobs a call? Not being a DOJ official, the litigator might have more freedom, more latitude, to deliver a proper pitch. After hanging up with Reback, Hirshland called Boies and ran the idea past him. Sure, Boies said, but I'll need Joel's blessing. "It might be delicate," Boies went on. "Can you get Senator Hatch to call Joel and tell him this needs to happen?" Which Hirshland promptly did.

Meanwhile, Reback had had a bright idea of his own. Realizing that part of Jobs' reluctance to testify revolved around the fear (a rare one for him) of standing alone at center stage, of being by far the most significant person in computing to be opposing Gates in so public a forum, Reback suggested to Apple's CEO that perhaps there was a way to give him some cover. What if another industry figure of Jobs' stature were to testify alongside him? Jobs liked the idea – though to his mind there was only one person who belonged in that category: Andy Grove.

Thus began a brief but frantic spell during which the DOJ and much of the anti-Microsoft movement was seized by the most feverish of fantasies: the Grove-Jobs twofer – get one, get both.

And a fantasy is precisely what it was. Not only was Grove the archetypal practitioner of corporate realpolitik, but at that moment Intel was engulfed in a substantial antitrust inquiry of its own, one being conducted by the FTC.

Nevertheless, Grove was deluged at home with plaintive calls from the DOJ's Silicon Valley surrogates. He heard from Hatch, and even from Steve Jobs. What none of Grove's suitors knew was that he was also receiving entreaties from Gates and Neukom, begging him to testify on Microsoft's behalf. Grove's reply to both sides was the same: Intel is neutral in this case and so am I. Besides, he told them, any testimony he gave was certain to be a double-edged sword. "I've been in the middle of all this shit for years," Grove told me. "I don't lie. I particularly don't lie under oath. And I particularly don't lie under oath when there's no reason to. I would have said things that neither side would have been happy to hear."

With Grove's irrevocable refusal, the DOJ lost its chance at Jobs. By the time Boies called Apple's CEO, "He'd made up his mind," the lawyer recalls. "He just didn't want to testify." Yet in failing to land the Valley's two reigning kingfish, the DOJ came away with two less spectacular but important victories. All along in its dealings with Intel, the government had feared a double cross; that, under pressure from Gates, the company would provide a witness, and perhaps even Grove, to testify for the defense. Now Grove had given his word that that wouldn't happen. And while Jobs rebuffed Boies when it came to testifying himself, he pledged to send Avie Tevanian in his stead.

By early October, with the inclusion of Tevanian and another software expert, James Gosling of Sun, the DOJ's witness list was complete. In the end, it had but one gaping hole: No OEM official would testify about how Microsoft leveraged its Windows monopoly to exercise coercive power over computer manufacturers. (IBM's witness, John Soyring, would talk only about the development of OS/2.) The search for an OEM whistleblower had consumed more man-hours at the DOJ than securing any other witness, but no amount of suasion was enough to convince PC makers that they had more to gain than lose by airing their grievances. "Most of the major OEMs are simply afraid," Klein told me that October. "A lot of them said to us, 'What you're doing is terrific, but we just can't afford to stick out our necks.' The power that Microsoft has over these people with the Windows license and the Office license is simply extraordinary."

The failure to land an OEM was frustrating for Klein, but it did nothing to diminish his sense of how far his team had come. After months of anxiety and hand-wringing, Boies and Klein were happy warriors – happier than anyone realized. For the DOJ's lawyers knew something few others did: They had a surprise witness up their sleeves. A witness of unimpeachable authority. A witness with power beyond reckoning and cash beyond counting. A witness guaranteed to overshadow even the brightest lights on the list they'd announced. A witness – need it be said? – who would soon have Microsoft's defenders paraphrasing Pogo: We have seen the enemy, and he is Gates.

Stylistically, Bill Neukom was an odd man out at Microsoft. In his fifties, he had a wavy pompadour of silver hair, a handsome face, and a vaguely patrician air. He was tall and trim and impeccably dressed, his suits well-pressed and invariably accented with suspenders and florid bow ties. Polite and formal, Neukom spoke in precise sentences that he strung together to form perfect paragraphs. He was occasionally turgid and always verbose. Once, after I'd finished a long interview with him, another Microsoft executive remarked, "I'm sure he crammed 20 minutes of substance into those two hours."

In 1988, three years after becoming Microsoft's in-house counsel, Neukom directed the defense against the Apple copyright suit, which threatened, Gates told me, to "absolutely put us out of business." The case dragged on for five years, and the received wisdom in the press was that Microsoft was in the wrong; that it had plainly ripped off Apple's graphical user interface to create Windows. But Neukom advised Gates to ignore the headlines and focus on the law, which the attorney was certain supported Microsoft's position. The court's vindication of that view, in 1993, was Neukom's greatest triumph, and a source of Gates' trust in his judgment.

Just as he had in the Apple dispute, Neukom believed unequivocally that the law was on Microsoft's side against the DOJ. To prove it, he and his lawyers set out over the summer of 1998 to pull together evidence to show that, far from being a monopolist, Microsoft faced competition from all sides; that the company's contracts with OEMs and ISPs were commonplace in the industry; that the infamous June 1995 meeting with Netscape was nothing more than a routine powwow between an operating-system vendor and an applications provider; that integrating IE into Windows wasn't part of a nefarious plot to wipe out Netscape but a natural extension of the OS, just as Microsoft's past inclusion of features such as printer drivers and memory management had been; that, in fact, the company's plans to incorporate browsing into Windows had begun before Netscape had even been born. In support of these claims they came up with hundreds of internal documents and email. They took dozens of depositions. And they assembled a witness list composed almost entirely of Microsoft executives, who would tell the company's story in court.

As the Microsoft lawyers readied their case, the most potent of all their potential witnesses dropped out of sight. In late July, Gates, as his board of directors had been urging him to do for months, named Steve Ballmer Microsoft's president. In an email to employees, Gates said that, from then on, Ballmer and Bob Herbold, the COO, would be responsible for running the company day-to-day, while he would spend his time on product development and new technology. "In no way am I pulling back," Gates wrote. "The hours I put in and my enjoyment of the work I do will be absolutely the same." And with that, he took off on a weeks-long vacation.

Nathan Myhrvold attributed the government's crusade to the impulses of "very successful people whose deepest regret is that they're not as rich as Bill."

Yet even when Gates was at play, business and the trial were never far from his mind. "He seemed totally in the loop," one person who saw him during that time told me. "He was aware of the issues, he'd read all the evidence and read up on the law, the procedures, the timing – everything."

Ten days before Gates was scheduled to be deposed by the DOJ, he jetted down to Silicon Valley for a dinner hosted by his friend Heidi Roizen, a software entrepreneur and former Apple executive, who had recently signed on as Microsoft's informal ambassador to the Valley. It was August 17, the day Bill Clinton came clean with Ken Starr – and with the nation – admitting for the first time his dalliance with Monica Lewinsky, and when Roizen's guests arrived at cocktail hour they eagerly scurried upstairs to watch Clinton's speech to the nation on the big screen in their hosts' bedroom. Perched at the edge of the bed, Gates heckled the president mercilessly, with a degree of venom that took many of the others aback. Clinton was a loser, he said; his speech was "hot air," a "pile of crap." To more than one person, it seemed obvious that Gates blamed Clinton for his antitrust woes. "If I did what he did in my office," Gates squawked, "the shareholders would throw me out!"

On August 27, in a windowless conference room in Microsoft's Building 8, Gates sat down across from his own Ken Starr for an extended spell of exquisite torture. "I expected that the Bill Gates I'd be facing would be the same Bill Gates I'd been in a room with that spring," David Boies told me. "The Bill Gates I'd met was smart and tough and articulate, a very passionate and effective spokesman for his point of view." Boies grinned. "Needless to say, that was not the Bill Gates who showed up for the deposition."

The Bill Gates who showed up for the videotaped deposition was not only the polar opposite of his public persona, he was a caricature of the polar opposite. He was dour and edgy. He was petulant and passive-aggressive, obfuscatory and obscurantist. He was a quibbler, a pedant, an amnesiac, a baby. He was the sort of CEO who would profess not to recall countless emails he'd written and who would claim to be ignorant of his company's strategies. Who would quarrel stubbornly over the meanings of words like "concern," "compete," "definition," "ask," and "very." Who would take five minutes to concede that when another Microsoft executive talked about "pissing on" Java, it was not, as Boies put it, a "code word that means saying nice things." Who, when asked who had attended a meeting of Microsoft's executive staff, would reply, "Probably members of the executive staff."

For his part, David Boies stayed cool. He was patient and persistent, asking certain questions again and again and again, often using precisely the same phrasing, until Gates either coughed up a straight answer or provided Boies an equally valuable display of prevarication. Judge Jackson had decreed that Boies could take as long as he wanted for the deposition. Early on, the lawyer remarked placidly, "I've got as much time as I need to finish the examination, sir, and I'm prepared to spend as many days here as I have to." In the end, that would be three, yielding 20 hours of Bill Gates unplugged.

At the end of Day One, Boies phoned Klein. "They're never going to call him as a witness now," he said confidently.

Klein was incredulous. "Well, that's not what we've heard," he replied. "We've heard that they're telling everybody they're going to bring him."

"They're not going to bring him. He's already said too many things he could never explain on the stand."

At the end of Day Two, although Boies hadn't yet covered all the ground he intended to, he was so delighted with the material he'd already garnered that he seriously considered ending the deposition right there. Gates was headed off for a long weekend on an Alaskan cruise hosted by Paul Allen, and Boies, already perplexed that Gates' lawyers hadn't stepped in to curb his behavior, assumed his quarry would return in greater command of himself. But Boies decided to risk it. On Day Three, his reward – among others – was one of the deposition's genuinely priceless exchanges. Handing Gates an email he'd written, Boies offhandedly remarked that at the top of the message Gates had typed "Importance: High."

"No," Gates said curtly.

"No?"

"No, I didn't type that."

Then who did?

"A computer."

Gates' performance was an unmitigated disaster, and not only in terms of PR. As a piece of evidence, it handed Boies the largest, most gnarled club imaginable with which to bludgeon both Gates and Microsoft as a whole, for the deposition fairly screamed that the dissembling at the company started at the top. It was a point that would not be lost on Judge Jackson. "Here is the guy who is the head of the organization, and his testimony is inherently without credibility," he told The New York Times after the case was over. "At the start, it makes you skeptical about the rest of the trial. You are saying, if you can't believe this guy, who else can you believe?"

Many observers would blame Gates' lawyers for the deposition fiasco, but Boies believes it wasn't that simple. "I've said many times that if I had been his attorney I would have stopped the deposition," Boies said. "But the thing I don't know, and the thing nobody will know unless I get Bill Neukom more drunk than he should be, is how much of it was the lawyers' unwillingness to act and how much of it was the client rejecting their absolutely unambiguous instructions." Not that Boies doesn't have his suspicions. "You have in Gates someone who is very smart, very rich, very powerful, and very much in command. He's a very hard client to say no to."

Very hard – or perhaps impossible. Since Microsoft's birth, Gates has seen himself as its chief legal strategist, Bill Neukom's presence notwithstanding. Reared in a lawyerly household, schooled by his father in lawyerly thinking, Gates' lawyerly proclivities have shaped the company and the software business profoundly. It was Gates who, in 1976, published a kind of manifesto, "An Open Letter to Hobbyists," in an early computer hobbyist newsletter, which asserted for the first time that software, like hardware, was a valuable commodity – it was intellectual property, and as such its creators deserve to be compensated. It was Gates whose grasp of the fine points of contracts had allowed him to outmaneuver IBM in the MS-DOS deal that would be the foundation of Microsoft's empire. And for all the accolades heaped on Neukom for the outcome of the Apple suit, the greater credit actually belonged to his boss. "Neukom did yeoman's work, but make no mistake, it was Bill who won the Apple case," a former Microsoft executive asserts. "He was deeply engaged in the case, he knew the issues, both technical and legal, and he played a huge role in framing them for the court. Hell, he practically wrote our briefs himself."

As the Sherman Act trial drew near, Gates boned up on antitrust, studying the law, poring over precedents. "Bill knows the courts to an amazing degree," a senior Microsoft manager told me. "He knows all about the judges – who they are, how they've decided in the past, district by district, all over the country. This is not a normal client who just sits across from his lawyers and takes their advice. No way."

To this day, Gates insists that his deposition performance has been badly mischaracterized. He answered honestly and precisely, he says. He seems especially wounded by the portrayal of him as forgetful, insisting over and over, in Rain Man-like cadences, "I have an excellent memory, a most excellent memory." "Did I fence with Boies?" Gates asks rhetorically. "I plead guilty. Whatever that penalty is should be levied against me: rudeness to Boies in the first degree." His tone of voice was regrettable, and so were the camera angles, he says, yet all this was mere atmospherics and therefore irrelevant.

Microsoft's lawyers are somewhat less sanguine. Compelled by reality (and a concern for their own reputations) to acknowledge the damage done by the Gates tapes, they blame Judge Jackson, who had issued a pretrial order that led them to believe the tapes would never be shown in court. Had they thought otherwise, Neukom told me, they would have prepared Gates differently – but only in terms of style, not substance. (They would also have made sure that the lighting was more flattering.)

Boies scoffs at the idea that Microsoft didn't know the tapes would be aired: "What, they thought I was taking them for my memory book?" He offers his own theory, which revolves around Gates' assumption going into the deposition that he would be called as a witness by one side or the other. "He must have thought that if he came as a witness, we wouldn't be able to introduce the videotape," Boies said. "And he was probably right about that. If he had been a witness, I don't think the judge would have let us play it. As a result, he wasn't really focused on how he looked in the deposition. He was prepared to stonewall. He was prepared to do all kinds of things that you might do if you believed nobody was going to see it."

Instead, the degree of Gates' stonewalling was so great, and his evasions were so egregious, that the deposition set in motion a cascade of unintended consequences. Suddenly, Microsoft had to keep its most powerful witness off the stand, lest he be humiliated in the attempt to defend the indefensible and explain the inexplicable. The DOJ, meanwhile, now had no reason to call Gates, for whatever he said in the courtroom could hardly serve the government's purposes more effectively than the testimony it already had in the can. The world's richest man had no date to the dance. And the video was fair game.

"It was like the Russian Revolution," Boies concluded. "Everything had to fall into place just so for it to turn out as it did."

Like the czars in Petrograd in 1917, Microsoft in the late summer of 1998 could feel the ground shifting beneath its feet. Nearly a year had passed since the DOJ had filed the consent-decree case, and in that time, almost everything that could possibly go wrong had. Surrounded by Bolsheviks and Mensheviks, populists and nihilists, the old regime began, for the first time, to betray a hint of what Gates, on a carefree day, would have called "concern," but that others might properly have described as panic.

Some months after his testimony, Dick Schmalensee would tell a fellow economist privately, "The lawyers are not in charge. All the shots are being called by Gates."

Trying to roll back the tide, Microsoft flooded Judge Jackson's chambers with pretrial motions – nine of them in September and October. The motions' themes cried out from their titles: "Motion to Limit Issues for Trial"; "Arguments for Excluding Extraneous Last-Minute Issues from Trial"; "Motion for a Continuance Needed to Address Testimony of Plaintiffs' New Trial Witnesses"; and so on. The case the government had filed in May, Microsoft argued, was all about browsers and a tad about Java. Broadening it beyond those issues was illegitimate, unfair, and a sign that the DOJ realized the Appeals Court's decision had "eviscerated" the core of its original complaint. At the very least, Microsoft maintained, the company needed more time to build a thorough defense.

The DOJ's response was swift, emphatic, and gently mocking. In one of its reply briefs, it wrote, "To the limited extent that plaintiffs offer evidence adduced in discovery concerning events and transactions not strictly limited to browsers and Java, those events and transactions (a) directly evidence monopoly power and barriers to entry, which issues are (of course) part of plaintiffs' complaints, and of every Sherman Act Section 2 case; (b) demonstrate Microsoft's intent to monopolize, which issue is (of course) also part of plaintiffs' complaints, and of every Sherman Act Section 2 attempt case; and/or (c) demonstrate a pattern that is relevant to understanding and establishing Microsoft's conduct with respect to browsers and Java." The DOJ's Jeff Blattner put it more colorfully: "We haven't broadened the case – we've broadened the evidence. In a murder case, you refer to the body in the filing. But at trial you bring out the bloody glove, the bloody shoes, the murder weapon."

Right up to the eve of the trial's first day, the back-and-forth between the sides continued unabated. But with every volley, the referee remained consistent. Time and again, in written orders and pretrial hearings, Judge Jackson informed Microsoft that the trial would be broad, and that it would focus on one large question: whether the company had "maintained its operating-system monopoly through exclusionary and predatory conduct." As Jackson put it matter-of-factly to Bill Neukom and his team, "My view of the case is not as narrow as yours."

And so it was that, on the morning of October 19, the courtroom phase of the Microsoft case began. For three solid hours, David Boies, slightly stoned on antihistamines and armed with only a few scribbled notes on one side of a manila folder, held the room pretty much in the palm of his hand. There was nothing soaring about his oratory, nothing ornate or mellifluous. Instead, the power of his opening lay in the narrative he unfurled and the evidence he unveiled in support of it. The story he told was straightforward: Faced with the threat posed by the browser and Java, Microsoft had tried first to coerce Netscape into not competing with it, and then, after being rebuffed, it had put the screws to the entire industry in an effort to destroy the startup and keep its grip on the desktop. As he walked Judge Jackson through the government's claims, Boies displayed on the courtroom monitors a sequence of documents that painted Gates and Microsoft as the most rapacious (and unsubtle) of monopolists. And, the pièce de résistance, there was Gates unplugged.

Here was Microsoft's CEO onscreen, denying knowledge of the June 1995 meeting – saying, indeed, "I had no sense of what Netscape was doing" at the time. And here was an email from Gates to Maritz and other Microsoft brass a few weeks before the meeting: "I think there is a very powerful deal of some kind we can do with Netscape…. We could even pay them money as part of the deal, buying some piece of them or something. I would really like to see something like this happen!!"

When Boies was done and the court gaveled out of session, Bill Neukom appeared before a gaggle of reporters on the courthouse steps. Calmly but adamantly, he denounced Boies' tactics as hollow theatrics, accusing him of using "loose rhetoric and out-of-context snippets" to disguise the fact that he had no case and adding that "none of these snippets, none of this rhetoric, even approaches proof of anticompetitive conduct."

The next day, Joel Klein flew to Scottsdale, Arizona. On the one-year anniversary of the consent-decree case, he was scheduled to give a keynote address at Agenda, the conference at which Gates had first heard the news that his government was suing him. The speech Klein would deliver was a high-minded affair, a discussion of regulation, market failure, and "the case for government involvement in the computer industry." He would offer few comments on the trial, and those he would offer were as dry and arid as the high-desert air. Klein knew better than anyone that the government had a long row to hoe. He expected Microsoft to mount an awesome defense. And he knew that one good day in court was no cause for chest-thumping.

Still, that one good day had been a very good day. In the back of the hall, Klein whispered to me, "I am one happy camper. We really kicked their butts."

VI. IN THE DOCK

The E. Barrett Prettyman federal courthouse squats at the northwest edge of Capitol Hill and bears all the hallmarks of the neo-brutalist architectural style that came into vogue in Washington in the 1950s. The six-story facade is gray and granite and imposingly free of inspiration. Inside, the walls are of marble – light gray streaked with darker gray. Down in the basement, a barebones cafeteria serves food, also gray, to the several hundred maintenance people and clerks who work in the building. (The judges tend to take their meals elsewhere – in Judge Jackson's case, at his club, the Metropolitan.) And yet, however mundane its appearance, the courthouse has provided the setting for more historic legal confrontations than anyplace but the Supreme Court itself. The Watergate trials, the arguments over the Pentagon Papers, the Whitewater/Lewinsky grand jury hearings – all were conducted here, at the corner of Constitution Avenue and Third Street, NW.

The Microsoft trial took place on the second floor, in Courtroom No. 2, a small space with five rows of pews in the back providing seats for just 100 spectators. Given the level of interest among the press, Judge Jackson had been urged to hear the case in the large ceremonial courtroom upstairs. But Courtroom 2 was where Judge John J. Sirica had tried the Watergate defendants, and Jackson told his clerks, "This case ain't any bigger than that one." In addition to being relatively cramped, the courtroom was windowless, airless, and charmless, bathed in fluorescent light and perfumed with the scent of stale arguments and fresh acrimony. In the absence of a jury, the jury box was occupied by sketch artists, who often surveyed the scene through special eyewear that resembled the night-vision goggles worn by Navy SEALs and Green Berets.

The lawyers from each side huddled around tables at Jackson's feet. Based strictly on appearances, it wasn't hard to see why oddsmakers favored Microsoft's team, which was composed of men in slick suits with hard eyes and harder hair. The government's table, by contrast, had a slightly ragtag look to it, the clothes off-the-rack, the coiffures pure Supercuts. Boies, with his mail-order apparel and scuffed black sneakers, could easily have passed for a GS-11 from the Department of Agriculture.

As much as the government's case had widened, its sine qua non remained Netscape, so the first witness Boies called was Barksdale. The job of questioning him fell to Microsoft's lead litigator, John Warden, a Sullivan & Cromwell partner with great experience in antitrust. In 1979, Warden had won the Appeals Court decision in Berkey Photo v. Kodak, which held that "any firm, even a monopolist, may generally bring its products to market whenever and however it chooses." A rotund man with dark-framed glasses, Warden spoke in a deep Southern drawl that rose up from his throat like a foghorn booming from the bottom of a well. (In private, Barksdale and Netscape's lawyers took to calling him "Boomer.") Between the lawyer and the witness, a Mississippi native, there were times in the next few days when, if you closed your eyes, you could imagine you were in a county courthouse far down below the Mason-Dixon line. While Warden mangled the names of Netscape's multiethnic employees, Barksdale peppered his answers with downhomeisms such as "We put a little Kentucky windage on it" and "it still irritated the stew out of me."

Barksdale's written testimony ran to 126 pages, and Warden appeared intent on refuting everything in it. Yet the matter that drew his most sustained fire was Barksdale's account of the June 1995 meeting. Warden first posited that, far from being a feared aggressor, Microsoft had been invited – no, begged – to do a deal by Netscape. In support of this assertion Warden produced an email to Microsoft from Jim Clark, written at 3 am on December 29, 1994. "We have never planned to compete with you," Clark wrote. "We'd like to work with you. Working together could be in your self-interest as well as ours. Depending on your interest level, you might take an equity position in Netscape, with the ability to expand that position later."

Barksdale reeled. At the time the email was sent, he had been a few days away from becoming Netscape's CEO; he had not known Clark had written it. And although he'd recently been forewarned by the DOJ that there had been a Clark-Gates exchange, no one had mentioned that Clark had, in effect, offered to sell the company to Microsoft. Barksdale told Warden that Clark had written the email in a "moment of weakness." He said Clark had been freelancing, that his note never represented the company's true strategy. But as Barksdale stared at the email on the courtroom monitor, all he could think to himself was, "Well, goddamn."

Warden asked Barksdale if Clark enjoyed "a public reputation for veracity."

Long pause. "I couldn't comment on that," Barksdale said. "I don't know."

"Do you regard him as a truthful man?"

Even longer pause. "I regard him as a salesman."

Throughout the Microsoft trial there would be moments that revealed what the journalist Joe Nocera called "the secret history of the software industry." This was one of them. To a Silicon Valley outsider, Barksdale's disavowals of Clark could only seem incredible – a prime example of Kentucky windage. Clark was Netscape's chair, after all, the man above Barksdale on the company org chart. But the truth of it was, Barksdale had never taken orders from Clark, who had a reputation for being modestly insane; indeed, he had agreed to become CEO only after receiving assurances from VC John Doerr that he would have complete freedom to ignore Clark's advice – which he'd done with impunity. As for the suggestion that Netscape was begging for a deal, it ignored how much had changed in the six months between Clark's email and the meeting in question. In December 1994, Netscape's sales were zero and its capital evaporating; by June 1995, it was the fastest-growing software company in history, one whose board had just voted to launch the IPO that would ignite the Internet boom.

The next day, Warden hammered away at Andreessen's notes from the meeting – "These notes of his aren't verbatim, are they?" – and at a chronology of events supplied to the DOJ a month later by Reback, which failed to mention the "stunning proposal to divide markets" that Barksdale was now alleging. "If you look at the whole record of events up to the June 21, 1995, meeting," Warden bellowed, "the only fair conclusion that can be reached is that Marc Andreessen invented or imagined a proposal to divide markets and that you and your company signed on to that invention or imaginary concoction in order to assist in the prosecution of this lawsuit!"

"I absolutely disagree," Barksdale said sternly, his face turning crimson. "I was in the meeting. I know what I know. I was a witness to it, and you weren't."

Out in the Valley, Reback heard about Warden's argument and was stunned. Whatever the chronology said, Reback knew that he'd phoned Klein and requested a CID the day after the June meeting, that Klein had complied hours later, and that Reback had sent in Andreessen's notes the following day. Digging through his records, Reback found a copy of the CID and faxed it off to Klein. (Apparently, the DOJ's had been buried away with the paperwork around the fruitless Microsoft Network investigation that had been going on at the time.) Over the weekend, the DOJ turned the documents over to Microsoft. The next Monday morning, Warden resumed with a new line of attack: Given the immediacy of Reback's request and the rapidity of the DOJ's response, didn't it all smack of a conspiracy?

"Isn't it a fact, Mr. Barksdale," Boomer boomed, "that the June 21, 1995, meeting was held for the purpose of creating something that could be called a record and delivered to the Department of Justice to spur them on to action against Microsoft?"

Jackson's disdain for Microsoft's defense became more and more evident. "The code of tribal wisdom," he said, "says that when you discover you are riding a dead horse, the best strategy is to dismount."

Barksdale: "That's absurd."

Afterward, on the courthouse steps, Microsoft's foes gleefully mocked Warden's gambit. Squinting into a warm October sun, Netscape's outside counsel Christine Varney quipped, "We've gone from Alice in Wonderland to Oliver Stone's JFK."

"In my experience as a litigator," Boies chimed in, "there are few signs more encouraging than when the opposition starts saying, 'They set us up.'"

Barksdale had expected to testify for two days; he spent a week on the stand. When it was over, Microsoft had scored on a number of fronts. It had got him to admit that he hadn't actually heard anyone from Microsoft speak of cutting off Netscape's air supply; in fact, Barksdale allowed that he'd first come across the phrase in a biography of Larry Ellison – an admission that pointed up that much of the government's evidence was hearsay and that Microsoft wasn't the only software outfit given to rough talk or hyperbolic metaphor. More significant, Barksdale acknowledged that more than 26 million copies of Navigator had been downloaded over the Net in the first eight months of the year, and that the company planned to distribute another 159 million copies in the next twelve. If that were true, Warden asked, how could the DOJ claim that Microsoft had foreclosed Netscape's distribution channels? If people could still "freely choose at no cost Netscape's Web-browsing software," as Warden put it, how could consumers possibly have been harmed?

Yet the overarching impression conveyed by Microsoft's defense was one of indiscriminate flailing. In the space of a few days, Warden had argued that Microsoft couldn't be said to have destroyed Netscape because Netscape was alive and well – but if Netscape was on the ropes, it was the company's own fault. He had argued that Microsoft hadn't acted like a bully – but if it had, it was acceptable, because everyone else in the industry did it. He had argued that the June 1995 meeting was either an elaborate frame-up, or an elaborate fiction, or a cordial meeting between potential allies, or the wary circling of potential rivals. Lawyers call this "arguing in the alternative." Generally, it is not a compliment.

The government's next witness was David Colburn of AOL. A legendary hardass, Colburn was the guy sent into every big deal at nut-cutting time. In March 1996, he had engineered the browser war's most famous double cross, in which AOL agreed to license Navigator one day, only to announce the next that it had chosen IE as its default browser, under terms that rendered the Netscape deal worthless. At great length, Warden attempted to induce Colburn to admit that AOL had done this because Microsoft's browser was superior. At even greater length, Colburn insisted that it just wasn't so; that, technically speaking, the products were a wash; and that the decisive factor was Microsoft's ability to give AOL's icon prime placement on the Windows desktop.

When Warden tired of this colloquy, he turned his attention to a series of emails in late 1995 between AOL's CEO, Steve Case, and Barksdale. In one of them, Barksdale argued that the two companies should team up to take on Microsoft. Case agreed, proposing a "grand alliance" that might also include Sun; suggesting that members of the alliance not invade one another's primary markets; and endorsing an idea of Andreessen's that "we can use our unique respective strengths to go kick the shit out of the Beast from Redmond that wants to see us both dead."

Warden asked Colburn, "A market-division proposal, isn't that correct?"

"I wouldn't call it that," Colburn deadpanned. "What it seemed like to me was a strategic relationship."

Once again, Warden was saying: Everyone does it. To which Boies, on the courthouse steps, responded: "The antitrust rules make a big distinction between what a monopolist can do and what everyone else can do." The difference, Boies said, was that "neither Netscape nor AOL had monopoly power."

Apple's Avie Tevanian, by all accounts one of the best minds in software, proved a lethal witness. For three weeks, Judge Jackson had absorbed the testimony of a parade of executives and lawyers who, when it came down to it, knew next to nothing about the raw material at the heart of the case – code. Jackson was ready to hear Tevanian's allegations that Microsoft had tried to divide the multimedia market with Apple; had pressured OEMs (and Compaq in particular) to drop QuickTime, even when Apple was letting them bundle it for free; and had wielded the threat of canceling Mac Office to blackmail Apple into adopting IE as its default browser. But what the judge wanted most from the witness, it turned out, was a software tutorial. Tevanian was only too happy to oblige.

The lawyer cross-examining him was S&C's Ted Edelman, who, like a tag-team wrestler, had stepped into the ring to relieve a fatigued John Warden. Edelman, a clever young man with serrated edges, realized he was in trouble early in Tevanian's second day on the stand, when, without warning, Jackson started questioning the witness himself. "What is a codec?" the judge inquired tentatively. Soon the proceeding was spinning out of Edelman's control. Every time he asked a question, Tevanian would turn and address his answer to the judge. When Edelman attempted to pin Tevanian down on one point, Jackson slapped the lawyer around: "Mr. Edelman, you keep mischaracterizing what he's told you. It's misleading language, and it's not acceptable to me." Eventually, Edelman found himself cut out of the loop completely, as Jackson and Tevanian engaged in a lengthy – and, for Microsoft, damaging – dialogue on the question of tying.

"From a technological perspective," Jackson asked, uttering a phrase that must have felt like Swahili as it left his lips, "what benefit, if any, is there, do you believe in integrating a browser as distinguished from bundling it with an operating system?" Less than none, replied Tevanian. "What you're telling me is, you don't think there is any benefit and there may be a detriment to the ultimate consumer?" That's right, replied Tevanian. "My final question: Is it possible to extricate your browser from the operating system without otherwise impairing the operation of the system?" Certainly, replied Tevanian. At which point Jackson – memories of the consent-decree case surely galloping across his cerebrum – nodded gravely, jotted a note, and then shot a glare toward the defense table.

The Microsoft team wore masks of misery. By the time Tevanian exited stage left, the defense was showing its first signs of disarray, with Neukom calling courtroom huddles during breaks and improvising tactics on the fly. After the case was over, Microsoft's lawyers and its PR people would agree on one thing, at least: Tevanian had been the government's best witness, his turn on the stand the moment when it first occurred to them that Microsoft might actually lose the case.

The DOJ was well-pleased with Tevanian, and with Barksdale and Colburn too, but Boies had no time for self-congratulation. Next on the stand would be Steve McGeady. Intel had declined to let McGeady submit written testimony, and thus he would be the only government witness whom Boies would examine directly. He was, as Klein put it to me, "the one wild card in our deck." And while the spectacle of an Intel official airing the Wintel alliance's soiled laundry in public would have been wild enough on its own, the drama was heightened immeasurably by one stark reality: Nobody – literally, nobody – knew what McGeady was going to say.

There were two salient facts about Steve McGeady. One was that he was exceedingly intelligent. The other was that he detested Microsoft. Whether these facts were connected was a moot point, but they had unquestionably defined his career at Intel.

McGeady was a Reed College Unix hacker who studied physics and philosophy, never graduated, and joined Intel in 1985, at age 27. Though few people are aware of it, Intel employs several thousand software engineers, most of whom write code that's embedded in its chips. (As Andy Grove says, "Silicon is frozen software.") It was from this crowd that McGeady emerged as a rising star. In 1991, he became one of the founders of the Intel Architecture Labs, an operation in Hillsboro, Oregon, that Grove hoped to turn into an R&D facility for the entire PC industry. Yet because many of its projects involved software, IAL was in constant conflict with Microsoft; indeed, the lab was a hotbed of what McGeady calls "a whole subculture of Microsoft-haters," of whom he was loudest and most acerbic. Not long after IAL's inception, he was asked to address a high-level strategy meeting on "the software environment" at Intel's headquarters in Santa Clara. After listening to Grove describe Intel and Microsoft as fellow travelers and another executive talk about being "hungry for a new relationship" with Redmond, McGeady opened his speech by saying, "I'll tell you, when I think of hungry fellow travelers, I think of the Donner party."

In the early 1990s, McGeady was involved in a series of increasingly bitter run-ins with Microsoft. Matters came to a head in the spring and summer of 1995, when a two-pronged rift pushed the two companies to the brink of open war. One prong was NSP, which was a layer of multimedia software developed by IAL that Microsoft opposed; the other was Intel's support for Netscape and Java, of which McGeady, Intel's chief Internet evangelist, was a primary champion. On both fronts, McGeady believed not only that Grove caved to pressure from Gates, but that IAL was "gelded" in the process. At that point, McGeady retreated into self-imposed exile, heading off to spend a year at the MIT Media Lab. On returning, he was put in charge of Intel's Internet health care initiative – a pet project of Grove's, who had been diagnosed with prostate cancer. McGeady's prospects were fine, but the wounds of the past remained open and raw. He told me, "I really think Microsoft is a fucking evil corporation; they're way out of line in all this." So when the chance to testify presented itself, McGeady jumped first and asked questions later.

From the moment of his deposition in August, McGeady was "quarantined," as he puts it, from the rest of Intel. He spoke to no one about the case except Intel's lawyers. He had no idea what Grove was thinking, no idea what the company was telling the DOJ. ("Nobody even told me I was on the witness list; I read about it in my underwear in The New York Times.") McGeady assumed Intel was cooperating, at least tacitly, because it hadn't fought the CIDs or tried to block his deposition. At the same time, however, he had been informed by Intel's lawyers that he wouldn't be submitting his direct testimony in writing. Also, Intel's lawyers were being squirrely about whether they represented him personally or only in his capacity as an Intel executive. Then, in early October, McGeady learned that Microsoft wanted to depose him a second time; and that the men from Sullivan & Cromwell were demanding his personnel file, including his performance reviews and salary records. Things, it seemed, were about to get nasty. The time had come to get his own lawyer.

One of the first things McGeady learned from his new attorney was that the DOJ had asked repeatedly to interview him – requests that Intel's lawyers had failed to convey. In the absence of written testimony from McGeady, the government wanted a clearer sense than could be divined from his deposition of what he'd be willing to say on the stand. Sure, McGeady said. A DOJ lawyer was coming to Oregon for the second deposition, on October 7. McGeady would meet him the next morning, whether or not Intel approved.

Disapproval would be too anemic a word for Intel's reaction. With its own FTC investigation under way, and the health of its relationship with Microsoft hanging in the balance, Intel was on a very thin tightrope. Grove had assured both sides that the company was neutral. He had assured Gates, especially, that Intel was doing nothing willingly to assist the government. Keeping up appearances was essential here – and now McGeady was about to make a great big mess.

At 7 am on the day of his interview with the DOJ, McGeady's phone rings, waking him, and the voice of his Intel lawyer, Jim Murray, pipes through the receiver.

Don't talk to the government today, Murray tells McGeady; we want to maintain neutrality.

Gates' bottom line: They'd done nothing wrong. They'd made no mistakes. In the end, they'd be exonerated: "Every action that's been attacked in this case was Microsoft working on behalf of consumers." In a word, everything at Microsoft was A-OK.

"Nobody ever asked me about that," McGeady replies.

"We don't have to ask you. You're an employee."

"Fuck that. I'm going."

A half-hour later, while McGeady is in the shower, the phone rings again. This time it's Intel's chief counsel, Peter Detkin, in a state of barely controlled fury.

"You're violating Intel confidences!" Detkin yells. "If you do this, it's a fireable offense!"

Detkin doesn't know Steve McGeady very well; he doesn't know that the best way to ensure that he'll do something is to tell him not to; he doesn't know that McGeady has, in his own words, "a real severe authority problem." So McGeady's reaction is unexpected.

"Pound sand, Peter. This is the fucking US government, OK? Just because you think you want to be seen as neutral doesn't mean I do. This is my reputation and my morality. So fuck you."

As McGeady arrives at his lawyer's office for the meeting with the DOJ, the phone rings yet again. Apparently, the situation has escalated to DefCon 3: Grove's second-in-command, Craig Barrett, is on the line now. Barrett's message is the same, and emphatically stated: Don't do this.

"Sorry, Craig," McGeady says. "If the government doesn't want to talk to me, I won't talk. But if they do, I will."

Hanging up the phone, McGeady walks into the conference room, shakes hands with the DOJ attorney, sits down, and starts to chat. One more time, a phone call arrives – but this time it's not for McGeady. It's for the government lawyer – Joel Klein is holding. Not three minutes later, the lawyer returns, apologizes, gathers his things, and leaves.

There it was: Intel had phoned Klein and twisted the screws. Boies later told me, "They said very bluntly, If you insist on meeting with McGeady, then you're going to make us hostile, you're going to make us an enemy. We've been neutral up till now, but if you do this, we are not going to be neutral anymore."

If the turn of events was profoundly unsettling for McGeady, it was hardly more comfortable for the DOJ. "First we can't get a written statement from the guy," Boies recalls. "Then we can't meet with him before we name him as a witness. Then we can't meet him before or after his depositions. I put that son of a bitch on the stand without ever having talked to him!"

Steve McGeady testified for three days in the middle of November, dressed in a dark suit and a patterned tie, wearing eyeglasses, a thick gray-brown beard, and an implacable expression. He sat motionless in the witness box and proceeded to pull back the curtain on the most lucrative partnership in the history of modern business.

Before Boies began his questioning, he played some excerpts from the Gates deposition. On the courtroom monitors, the lawyer asked Microsoft's CEO, "Did you ever express any concern to anyone at Intel … concerning Intel's Internet software work?" After an interminable pause, Gates replied, "I don't think Intel ever did any Internet software work."

Boies: "And if they did, I take it that it's your testimony that no one ever told you about it?"

Gates: "That's right."

Boies: "Did you or others on behalf of Microsoft tell Intel that Microsoft would hold up support for Intel's microprocessors if Intel did not cooperate with Microsoft?"

Gates: "No."

Boies: "Did you, Mr. Gates, ever yourself try to get Intel to reduce its support of Netscape?"

Gates: "I'm not aware of any work that Intel did in supporting Netscape."

It would take McGeady roughly two hours to make Gates out to be a liar on all this and more. In response to Boies' questions, McGeady told the court that Gates had been briefed many times on Intel's Internet software development – once, at least, by McGeady himself. Gates "became quite enraged," McGeady said, about "the software engineers in IAL who were, in his view, competing with Microsoft." McGeady told the court that at one 1995 meeting, "Bill made it very clear that Microsoft would not support our next processor offerings if we did not get alignment" on platform software – a threat that McGeady called "both credible and fairly terrifying." He told the court how Intel's NSP had caused a "conniption" at Microsoft, which saw the software as an invasion of its turf. He told how Intel's support of Java had been, in the words of one email, a "show stopper." And he testified that "it was Microsoft's desire that we essentially clear and get approval for our software programs from them before proceeding."

McGeady also told a story about Paul Maritz – a story that gave credence to one of the trial's least consequential but most highly publicized claims. In the fall of 1995, McGeady said, he'd attended a meeting where Maritz laid out for a handful of Intel executives Microsoft's strategy for defeating their "common enemy," Netscape. The strategy had three elements: Microsoft would "embrace, extend, and extinguish" open Internet standards; it would fight Netscape "with both arms," meaning both its OS and its applications; and, Maritz fatefully declared, it would "cut off Netscape's air supply" by giving away IE for free.

McGeady's testimony was buttressed by an assortment of astonishing documents, the most explosive of which was a memo he'd written after an August 1995 meeting attended by both firms' CEOs. Bearing the title "Sympathy for the Devil," the memo said, "Bill Gates told Intel CEO Andy Grove to shut down the Intel Architecture Labs. Gates didn't want IAL's 750 engineers interfering with his plans for dominating the PC industry." More damning still were a slew of Gates' own emails, which Boies entered into evidence in rapid-fire succession. "We are trying to convince them to basically not ship NSP," Gates wrote after a dinner with Grove in July 1995. "We are the software company here and we will not have any kind of equal relationship with Intel on software." A few months later, after Microsoft had pressed computermakers aggressively to reject Intel's multimedia software, Gates wrote, "Intel feels we have all the OEMs on hold with our NSP chill … This is good news because it means OEMs are listening to us."

By the end of his first day in the dock, McGeady had made so many incendiary allegations that Boies feared the Intel brass would intercede – either pressing him to clam up or pulling him off the stand altogether. By the end of the second day, his testimony had taken on the flavor of a software-world Scenes from a Marriage. The Intel-Microsoft coupling had always seemed a union of equals. But in the picture McGeady painted, it was Microsoft that clearly wore the pants in the family, while Intel played the part of the long-suffering spouse, sticking with the relationship because, as one Intel memo put it, "divorce will be bad for the kids." ("The kids," McGeady explained, were the OEMs and other industry players.)

The S&C lawyer charged with McGeady's cross, Steve Holley, knew he faced an uphill slog. He started off well enough, using the depositions of McGeady's immediate superior and of other Intel executives, as well as a raft of emails, to sketch a coherent counter-explanation of why Microsoft had torpedoed NSP: Rather than being tailored for the forthcoming Windows 95, Intel had targeted the technology at Windows 3.1. "In retrospect, a mistake," McGeady allowed.

But Holley ran into trouble with his next move, a venomous and voluminous attack on McGeady's credibility. McGeady was arrogant. McGeady was biased. McGeady was, in the words of one of his colleagues in an email Holley brandished, a "prima donna." ("I've been called far worse," McGeady said with a grin.) He was also a fabulist and a fabricator, argued Holley. Harking back to Barksdale's testimony, the lawyer accused McGeady of cribbing the air-supply quote from a Larry Ellison biography. He accused McGeady of being in league with Jim Clark. He even accused him of being rude about his boss, citing an email in which McGeady referred to Intel's chair as "mad-dog Grove."

"What is the point of this?" Judge Jackson demanded. "Are you just trying to embarrass him?" In an act of perjury as great as any ever committed in Courtroom No. 2, Holley denied it.

Still, Jackson was curious about Steve McGeady. He had a question of his own, a question that everyone in the courtroom was, in fact, dying to ask. So when the cross-examination ended, Jackson said, "Mr. McGeady, to what extent do you understand that you are a spokesman for Intel Corporation here as distinguished from speaking for yourself?" As ears pricked up and eyes widened at the lawyers' tables and in the spectators' pews, McGeady hemmed and hawed. Jackson tried again: "Are you here with the blessing of your CEO?"

"'Blessing' would be a strong word," McGeady mumbled. "I'm not trying to be evasive, Your Honor. It's a difficult question…. I believe that in certain circumstances, Dr. Grove and other executives might share some of my opinions. In some cases they would share them privately. They may not agree with my expression of them."

"Are you aware of any instances which are actually at variance with what you understand to be corporate policy?" Jackson asked.

"Perhaps only the most dramatic, Your Honor," McGeady answered. "It's important to Intel to maintain a positive working relationship with Microsoft. My appearance here, obviously, creates a problem there."

And with that, McGeady got up and went back to Oregon.

Eight weeks later, the moment that filled him with more trepidation than any court appearance ever could finally arrived. At an annual black-tie dinner for Intel's senior executives, McGeady came face to face, for the first time since being quarantined the previous summer, with Andy Grove. Clutching a cocktail, surrounded by a boisterous crowd, McGeady made small talk for a few minutes and then gingerly tiptoed into the danger zone: "Hey, Andy, um, about that other thing, you know, no hard feelings, I hope …"

Grove's eyes twinkled. "Vell," he replied in his Hungarian accent, "I vould have done it a different way. But I guess it vorked out OK in the end."

For sheer drama, nothing in the rest of the government's case approached the quality of its first four witnesses; the next two months were up and down. John Soyring of IBM rehashed the controversies around OS/2. James Gosling of Sun, a long-haired, pot-bellied, bushy-bearded Buddha figure with so many forms of RSI he's officially handicapped in the state of California, testified with such low-key candor that his testimony kicked up little dust. Edward Felten, a Princeton professor, contended that he'd devised a small software program that could remove IE from Windows 98 – something Microsoft claimed was impossible. William Harris, the new CEO of Intuit, stumbled badly on the stand by wandering out of the land of fact and into the realm of speculation, and by offering half-baked ideas about remedies, which allowed Microsoft's lawyers to suggest, not without reason, that he was calling for a National Operating System Commission. Finally, the MIT professor Franklin Fisher, a giant in the field of antitrust economics who had worked with Boies on the IBM case, argued that Microsoft had created high barriers to entry in the operating-system and browser markets, and that the company had the ability, even if it didn't use it, to raise prices almost at will – two key tests of monopoly power. As the first half of the trial came to a close, an air of confidence tinged with cockiness filled the hallways of the DOJ.

In public, at least, Microsoft's lawyers displayed an almost commensurate degree of confidence. The facts and the law were on the company's side, Neukom told me. As Warden had argued during his opening statement, "The antitrust laws are not a code of civility in business," and although Microsoft had played tough, its actions had only benefited its customers. Indeed, even Professor Fisher, when asked by a Microsoft lawyer if consumers had been harmed, had said, "On balance, I would think the answer was no, up to this point." And while it was undeniably true that Microsoft possessed a high share of the OS market, Neukom believed that the company had demonstrated conclusively that the software business was fiercely competitive, and that Microsoft's position in it was forever under siege.

Neukom's point had been underscored in late November, when AOL announced, in effect, that the "grand alliance" Steve Case had dreamed of in 1995 was about to become a reality. In exchange for $4.2 billion in stock, AOL planned to acquire Netscape and then to team up with Sun Microsystems to create an Internet powerhouse aimed squarely at challenging Microsoft. On the courthouse steps, Neukom declared, "From a legal standpoint, this proposed deal pulls the rug out from under the government. It proves indisputably that no company can control the supply of technology. We are part of an industry that is remarkably dynamic and ever-changing."

Yet for all of Neukom's proclamations to the contrary, the mood behind the scenes on the Microsoft team was considerably more sober. Judge Jackson had rejected almost all of the defendant's motions. He had repeatedly upbraided the S&C attorneys. He had rolled his eyes, shaken his head, and giggled conspicuously (along with the press) every time another piece of billionaire vérité had flickered on the courtroom monitors. In late November, at a conference in chambers with lawyers from both sides, John Warden had made one of numerous pleas to have the judge stop Boies from showing the tapes in "bits and pieces" and instead to have the whole thing shown. Jackson had again shaken his head. "I think the problem is with your witness, not with the way in which his testimony is being presented," the judge said. "I think it's evident to every spectator that, for whatever reasons, in many respects Mr. Gates has not been particularly responsive to his deposition interrogation."

Within a couple of weeks of the trial's opening bell, Neukom and S&C's lawyers had started tailoring their approach for the appeals-court case that seemed increasingly inevitable. Jackson had given them plenty of grounds for complaint, from the trial's unique procedures (the 12-witness limit, say) and the broadening of the case to his decision to admit what Warden called "multiple layers of hearsay" as evidence.

"Vindication will be bittersweet," said a Microsoft official. "Now, either the decision stands, and people think we're criminals, or it's overturned, and people think we somehow got away. No vindication will erase that stain."

Jackson's court wasn't the only forum in which Microsoft was faring poorly. To any reporter who was willing to listen, Microsoft PR specialists flogged polls indicating that the company's image remained in fine shape. Privately, though, one of them told me, "We knew we were losing the PR war, and badly."

In early December, a decision was made to roll out the big gun: Gates himself appeared, via satellite hookup, at a hastily arranged press conference at the National Press Club. "In the software industry, success today is no guarantee of success tomorrow," he said. And, "The government is trying to increase the cost that consumers have to pay for browsers." And, "Three of our biggest competitors band together to compete with Microsoft, yet, amazingly, the government is still trying to slow Microsoft down." Then Gates did something unusual: He turned to the topic of his deposition and vented his spleen at David Boies. "I had expected Mr. Boies to ask me about competition in the software industry, but he didn't do that," Gates said. Instead, "he put pieces of paper in front of me and asked about words from emails that were three years old." Asked about Judge Jackson's recent criticism of his performance, Gates snapped, "I answered truthfully every single question … but Mr. Boies made it clear … that he is really out to destroy Microsoft … and make us look very bad."

On TV that night, and in the papers the next day, "destroy Microsoft" would be every reporter's lead. With just two words, Gates had provided ringing confirmation of what many in the media already suspected: that he was paranoid, self-pitying, and quite possibly delusional.

And maybe he was. Soon enough, I would see for myself.

VII. IN THE BUNKER

The weather when I arrived in Redmond was filthy: the sky soupy gray, the roads slick with rain, the landscape draped in a fog thick as porridge. It was January 1999, midway through the trial's courtroom phase. After three treks to Microsoft's campus in as many months, I had started to think of it as a mushroom colony – a damp, leafy mulchpile where spongy-beige coders multiplied in the dark. There were 45 buildings on campus, and a new one seemed to spring up every week. Many of the buildings were connected by a labyrinthine series of hallways and passages, so that employees could shuffle from their offices to the company food courts and back again without ever encountering even a dewdrop of moisture. On days like this, you could drive around campus for hours without seeing a soul – and often, you had to. Even on a holiday (in this case, Martin Luther King Jr.'s birthday), the parking lots were jammed to capacity with Acuras, BMWs, and SUVs.

The official line at Microsoft was that the trial was mere background noise; that no one was distracted by it; that they were all too busy cranking out the next great chunk of software. Yet in reality the topic was inescapable. All over downtown Seattle, some renegade artist had plastered up posters featuring a macabre caricature of Gates under the headline "Trust Me" – the first word overlaid with a blaring red "Anti-." One day, in one of the Microsoft cafeterias, as my designated PR handler went on about how surprised she was that nobody ever talked about the doings in Washington, DC, an Indian programmer to our left regaled his friends with a detailed assessment of the government's technological ineptitude, while a German to our right called Joel Klein a socialist. (My handler smiled sheepishly and picked at her stir-fry.) Even the hallways were papered with protest. BOYCOTT THE GOVERNMENT. BUY MICROSOFT read a bumper sticker on one office door.

Among the Microsoft executives I spoke to, the sense of persecution was pervasive and acute. The only question had to do with the government's motives: Was it acting out of malice or plain stupidity? Brad Chase, a close Gates consigliere, blamed the "Alice in Wonderland" culture of Washington, and suggested that Klein was driven by (unspecified) political pressure. Charles Fitzgerald, who was Microsoft's one-man "truth squad" on Java, saw the culprits in Silicon Valley, and postulated the existence of shadowy meetings between McNealy, Ellison, Barksdale, and Doerr (four men whose combined egos were barely containable within one state, let alone one room) to plot twin conspiracies against Microsoft in the courts and in the marketplace. Nathan Myhrvold preferred a psychoanalytical take, attributing the government's crusade to the impulses of a collection of "very successful people whose deepest regret is that they're not as rich as Bill."

Other executives, and especially those who had already been touched directly by the trial, were deeply embittered. In 1995, as Paul Maritz' 26-year-old technical assistant, Chris Jones had been among the Microsoft contingent that attended the infamous June meeting at Netscape. Jones claimed that nothing untoward had happened there. Indeed, he told me that the very idea that he'd been part of some "Microsoft mafia" trying to intimidate Netscape into dividing the browser market was "ludicrous" on its face. The Microsoft team was made up mostly of junior-level staff like him. The Netscape side was led by Barksdale, an "impressive guy who'd been doing business for a long time." Jones said, "I think the perspectives on who was being intimidated in that meeting differ." Taken at face value, the comment was a telling reflection of the insularity of the Microsoft culture. Regardless of Barksdale's age and experience, Netscape was a money-losing startup, and Microsoft was – well, Microsoft. When Jones walked in the door, what the Netscape people saw wasn't some 26-year-old kid; they saw a 26-year-old kid who spoke for Maritz, one of the most powerful executives in the software industry.

And that was how the DOJ saw Jones too. In a deposition in April 1998, Jones had made statements that the government believed supported its case, a number of which had turned up in its court filings and in Boies' arguments in court. The statements were damaging – and, in Jones' opinion, taken flagrantly out of context. From a 45,000-word deposition, he said, the DOJ had lifted a few isolated, ambiguous comments that served its purposes while ignoring numerous straightforward denials that didn't. Microsoft had taken pains to point this out, but the press had run with the DOJ's interpretation anyway. For months, Jones' friends and family had been asking him: Is it true? Did you really do this, say this? By the time I met him, Jones was shaken. "It's been disillusioning, because it's a case where being really honest and answering questions fully did not serve me well," he said. "I'd be happy if there was a trial on the merits, but there's so much other bullshit going on – the PR, the leaks, the Gates video – you can't even tell what the merits are."

Listening to Jones describe his sense of being violated by the DOJ, it was impossible not to think of Microsoft's chair. The transformation of his deposition into a kind of televised water-torture – drip, drip, drip – had been one of the most severe public humiliations inflicted on a CEO in recent memory. In Microsoft's executive ranks, the traditional reverence toward Gates was now tinged with a new emotion: protectiveness, even a touch of pity. "I feel sorry for Bill," Greg Maffei, Microsoft's then-CFO, told me over a late dinner the night before one of my meetings with Gates. "This poor guy. Look at all he's accomplished, look at all he's done. Now he's being vilified. Not exactly a happy resting place." I mentioned Gates' depression at the end of the consent-decree case. "That was bad, but the videotape thing has been worse," Maffei said. "The fact that it goes on and on, that it feels like it's never going away. Every day they play a new snippet and make him look bad, and there's no way to punch back. It's tough on him because it makes him second-guess himself, which is not" – Maffei chuckled – "what Bill generally does."

I asked Maffei if he thought the clash with the government had changed Gates. "How could it not?" he said. "He's human. No human could go through what he's gone through and come out the other side unchanged."

Once upon a time, not very long ago, interviewing Bill Gates was one of the great pleasures in journalism – assuming you had a mild streak of masochism. From Microsoft's earliest days, he dispensed with the standard CEO patter and established a rapport with the media that was decidedly more frank. Though he could charm and flatter as proficiently as anyone, he would also badger, mock, and harangue. His favorite riposte to Microsoft subordinates – "That's the stupidest fucking thing I've ever heard!" – was one he never hesitated to hurl at a reporter who happened to ask him something silly or obvious. But the flip side was that, if you coughed up a question Gates considered sharp, he tried hard to answer it with equivalent insight. "Right! Right!" he'd yelp, jumping up to his feet, pacing around the room, engaging in an act most other public figures would regard as dangerously rash: thinking out loud. Despite the abuse, interviewing Gates was exhilarating.

By the time of our meeting in January 1999, that Gates was long gone. With the release of Windows 95, a milestone in the history of high tech hype; with his ascension to the pinnacle of personal wealth; with the construction of the 37,000-square-foot, $30 million lakeside compound he called home; with all this, Gates had transcended the software business and become a celebrity in the broadest sense possible. This had taken its toll. It had sanded down his rough spots, leaving him smoother, more polished, but infinitely blander. Now, under assault by the government, Gates seemed increasingly schizophrenic, vacillating in public between bursts of outrage – his attacks on Boies, say – and excretions of saccharin. In the space of one month that winter, he managed to appear on both Rosie O'Donnell's and Martha Stewart's TV shows, where he avoided all topics of controversy and rabbited on about the joys of parenthood.

The Gates I encountered that cold misty morning was guarded, distant, and defensive. He wore brown slacks, brown loafers, and a white dress shirt with faint brown stripes and his initials monogrammed on the breast pocket. His hair was freshly washed and parted vaguely on the side; an unabashed cowlick shot up from the back of his head. We sat at right angles from one another on Breuer chairs positioned next to a small maple coffee table. The tabletop had nothing on it but a jar filled with a dozen identical black ballpoint pens, which Gates would use every so often to draw diagrams for me on a yellow legal pad.

We talked for a while about the mid-1990s, the timeframe around which the trial revolved. That Gates had been late to grasp the significance of the Web, and had then turned Microsoft on a dime to embrace it, was a fact no one disputed – until the court case, that is, when the company suddenly, and for obvious reasons, started peddling the revisionist history that its plans for the Web had taken shape before Netscape's founding. I noted that the first edition of Gates' book, The Road Ahead, which was published in the fall of 1995, had barely mentioned the Internet.

"That's not true! This is a book …" he started to say, then caught his irritation and trailed off. "Certainly there were things we missed. We did our big mea culpa in December '95 in terms of realizing the importance" of the Web. "But the Internet, you could still say, Do people get it? Did people know six months ago that Amazon was worth $20 billion? How many people got it? I didn't happen to get it. I didn't go out and buy it, so, darn, that's another thing I missed."

The night before, Maffei had pointed out that, before the consent-decree case was filed, "in the public eye and most influential circles, Bill sort of walked on water; he could do no wrong." I wondered how it felt to have seen the tide turn so dramatically. "Eighteen months ago, you were universally admired," I said to Gates. "Hardly anything really bad had ever been written about you."

"That's not true!" he protested again. "Let's live in the real world for about half a second here."

I asked him if he felt like a victim of what Bill Clinton had memorably described as "the politics of personal destruction."

"It's overwhelmingly true that the case is misguided," Gates replied calmly. "Was Netscape able to distribute their product? Is that tough to decide? Was Netscape able to thrive in terms of being able to get advertising revenues &##91;from its Web portal]? Well, they were purchased for over $4 billion. Those are the two questions the complaint in this case raises. And that's it. So, clearly, if they've got a tough time with those, they're going to go try and throw as much mud as they can. And there's going to be competitors who are going to show up and participate in that."

Not only competitors, I interjected. Had Intel's participation in the case put a strain on its relationship with Microsoft? "That has no effect on the relationship whatsoever," Gates replied. "You're asking very Hollywood-type questions. These are companies that have to keep innovating in their products. We don't make chips. We're dependent on Intel."

Maybe so, but to see Intel on the witness stand was still rather striking.

Gates' face turned the color of claret. "No, it's not Intel up there – it's Steve McGeady! Don't say Intel! Intel was not up there! Steve McGeady was up there. Was I surprised that Steve McGeady does not like Microsoft? No."

Considering how things were going in court, I asked if Gates regretted having not settled the case in May 1998. "I would have been glad to do a settlement," he said. But, "when it comes to giving up the ability to innovate in Windows, that was something that, whether it's for Microsoft's shareholders or consumers at large, was not something I felt was right to give up."

I asked Gates if he believed it was possible to have a monopoly in the software industry. "In operating systems, no," he said.

Impossible?

"It's not possible."

Why?

"Because people's expectations of what they want out of the operating system are constantly changing. They want something better. Why have I increased our R&D from a few hundred million a year to $3 billion? Because it's a very competitive business…. A monopoly is where you don't have competition. The notion that this is a market without competition is the most ludicrous thing I have ever heard in my life."

Monopoly or no, Windows was unquestionably an enormous asset for Microsoft. ("An asset of the shareholders of Microsoft," as Gates put it.) And it was one over which the company had claimed total freedom – the freedom to add a ham sandwich, for instance. Was there any limit to how far he was willing to press the advantage of owning the dominant operating system?

"I don't know what you mean by 'advantage,'" he said, inspiring in me the brief fantasy that I was David Boies. "It is one of the more proven things that just because we put something in the operating system doesn't mean people will use it," Gates went on, citing the early, failed versions of IE, as well as the MSN client software. "Putting new features in the OS is a very, very good thing. Some of those features will end up being used heavily and some won't. All you have to do is look at the growth of the software industry to say this is an industry that's delivering for consumers in a fantastic way. So, yes, innovation is OK."

Gates hadn't answered the question, so I asked it again, this time more precisely: "Is there any limit to what you regard as appropriate to put into the operating system?"

"Let's say a piece of software is free and it's distributed on the Internet. Then it's available to everyone, friction-free. Is that software part of every PC? Well, logically, it is. They can just click and download it and get it on the PC. So, if we decide essentially to have a piece of software that's free, many, many companies can do that."

I repeated the question once more.

"Understand, anybody can give any piece of software away for free. That's just a fact."

"They can't integrate it into the operating system," I said, "because they don't own the operating system."

"Anyone who owns a product, like AOL, integrates new capabilities all the time. Netscape integrated massive new capabilities into their browser – mail and conferencing and dozens of other things. The fact that companies innovate in these products and put new features in, that's a good thing. I can't even think of a scenario where that would be negative."

But AOL doesn't own the operating system, I said.

"They own their online service."

And so it went, round and round like that for 15 minutes or so. Six times I asked Gates the question; six times he ducked and dodged. It was truly depressing. The idea that Microsoft had the unfettered right to add anything it wanted to Windows was an extreme principle – but it was a real principle, and it was one that was arguably worth defending. The old Bill Gates would have defended it forthrightly. The new Gates wouldn't, or, at least, didn't. After a year of withering press and tribulations in the courts, Gates may still have retained the courage of his convictions. But he was flaccid, lifeless; all the piss and vinegar seemed to have been drained out of him. In more than an hour, he didn't call me stupid even once.

Yet for all his dismay about how the case had been going, Gates still seemed to hold out a sliver of hope. "Are you going out to DC for the rest of the trial?" he asked as I got up to leave. I said that I was.

"I'm really looking forward to our witnesses," he said. "Now people will finally hear the other side of the story." For the first time all morning, Gates actually looked pleased. "You know, you've got to have faith that the facts will come out in the end. And the facts, in this case, are all on our side."

VIII. SHOWTIME

On the other coast, in the other Washington, the DOJ had just rested its case, and was readying itself to prove Gates wrong. "The next three weeks are critical for us," a senior government lawyer confided to me. "Microsoft is putting its three most important witnesses on first. The pressure's on David to make some headway right off the bat. If he does, we're in good shape. If he doesn't, we could be in trouble."

First up was Richard Schmalensee, an MIT professor with wavy gray hair and a well-tended mustache. Like his colleague Franklin Fisher (who, by an irony, happened also to be his academic mentor), Schmalensee was one of the country's most formidable economists, the dean of the Sloan School of Management. In antitrust cases, economists are crucial. While executives can testify about particular events and decisions, the economists, as Dan Rubinfeld puts it, "weave the facts together and explain why a company's business practices make sense and are legitimate."

Rubinfeld and Boies considered Microsoft's choice of Schmalensee as its inaugural witness a risky one. "When your economist goes first, he sort of lays out in advance a justification for what the other witnesses are about to say," Rubinfeld explains. "If he does a good job, everybody else looks good. If he doesn't, it casts a shadow on everything that follows." It was this reasoning that led Boies, despite his faith in Fisher from their work together on the IBM case, to put him last on the government's witness list.

The DOJ was doubly surprised that Schmalensee was the only economist in Microsoft's quiver. (The government had two.) In an antitrust trial, there are two big issues for an economist to address: Does the company in question have monopoly power? And has it abused that power? Both Boies and Rubinfeld told me that if they'd been advising Microsoft they would have counseled the company to concede the first issue, as Boies did in the IBM suit, so as to strengthen its hand on the second; both believed Microsoft hadn't done this out of fear that the concession would be used against it in future antitrust litigation. But by asking Schmalensee to advance both claims, Microsoft placed him in a vulnerable position.

Boies punched holes in Schmalensee's attempts to assert that Linux, BeOS, and the Palm OS posed a significant threat to Windows. But the coup de grace came on Day Two. In preparing for Schmalensee's cross, Rubinfeld and his team of economists were startled to discover a 1982 Harvard Law Review article in which the witness argued that "persistent excess profits" indicate monopoly power – an argument that contradicted his current position. The DOJ had no doubt that Schmalensee would have an explanation ready if Boies asked about this. How could he not? Yet when Boies confronted him with his own writing, Schmalensee was dumbstruck. He said, slightly slackjawed, "My immediate reaction is: What could I have been thinking?"

From that point on, Boies believed Judge Jackson saw Schmalensee as Microsoft's "what-could-I-have-been-thinking expert witness."

When Boies asked Schmalensee whether he'd tried to determine how much of Microsoft's profits came from operating systems, the economist said he had, but was told by the company that it didn't have that data.

Boies: "And did you accept that explanation at face value, sir?"

Schmalensee: "I was surprised, but I will be honest with you … Microsoft's internal accounting systems do not always rise to the level of sophistication one might expect from a firm as successful as it is."

Meaning?

"Mr. Boies, they record operating-system sales by hand on sheets of paper."

"Your honor," Boies said, grinning madly, "I have no more questions."

At the lunchtime recess on the afternoon Paul Maritz was to take the stand, Boies sat alone in the empty courtroom, staring up at the ceiling for a very long time, then down at the documents spread out before him, like a surgeon contemplating his instrument tray. Boies was aware, like everyone else, that his cross of Maritz was the trial's most high-stakes operation. As Microsoft's group vice president for platforms and applications, Maritz was generally regarded as the company's number-three man, and, in the absence of numbers one and two, he would be the seniormost executive to appear in court. His fingerprints were all over virtually every strategic decision under scrutiny; indeed, it often seemed as if his name was on more of the email evidence than Gates'. Anticipating a showdown, Joel Klein arrived and took his occasional seat in the front row, just behind the government lawyers' table. The courtroom was packed; the atmosphere, electric.

For the next four days, Boies and Maritz tangled like a pair of scorpions in a sock. When it was over, Neukom would declare the Microsoft executive victorious, trumpeting the fact that Boies had left untouched most of the claims in Maritz' 160-page direct testimony.

Boies, meanwhile, was convinced that refuting every jot and tittle of Maritz' testimony was unnecessary and maybe unwise. Better to drill down on a few crucial points and throw a haze of doubt on the witness' credibility.

One of Boies' prime targets was the 1997 Apple deal. In his testimony, Maritz denied that Microsoft had used the threat of canceling Office for Mac to induce Apple to adopt IE. He claimed that the browser was but a minor part of the negotiations, whose overriding concern was settling a potential patent dispute between the two companies. The problem for Maritz was the email trail. Boies presented one message after another, many from Gates himself, in which the browser issue featured prominently while the patent issue was mentioned only in passing or not at all. Maritz stuck to his guns. He insisted that Greg Maffei, Microsoft's CFO, who negotiated the agreement with Steve Jobs, had assured him that the first time Maffei brought up making IE the Mac's default browser was during a long walk around Palo Alto with a barefoot Jobs in July 1997 – well after the "primary deal terms," including the continuation of Office, were settled.

Watching this, I could only shake my head. I had covered the Microsoft-Apple negotiations. On the day after Jobs announced the deal at the August 1997 Macworld trade show in Boston, I interviewed Apple's top executives about how it had come together. The dickering had gone on until 2 am, just a few hours before Jobs' keynote. What was the hangup? The default-browser issue, the Apple guys said; if they hadn't given in, the deal would have fallen apart, and Apple would have lost Microsoft's commitment to Office. A week later, I called the very same Greg Maffei whom Maritz was now citing and put the question to him. Yes, he said, the browser had been the late-night sticking point. I asked what leverage Microsoft had used to secure IE's status as the fallback browser. "I don't want to comment on that," Maffei said. I pressed the point. Was it fair to assume that, at the eleventh hour, Apple had reason to fear the cancellation of Office? "Yeah, that's fair," he said.

(Three years later, Maffei also admitted to me that although Microsoft had bought $150 million in Apple shares as part of the deal – "We invested in the company when people had lost faith," Gates would boast – he had hedged Microsoft's bet by simultaneously shorting the stock.)

Boies then turned to Netscape's air supply. After a bit of thrust and parry, the adversaries deadlocked. Maritz denied ever having said anything of the kind, and Boies could produce no smoking email; Steve McGeady remained Maritz' sole accuser.

But the truth was that, among the several Intel officials who attended that meeting, at least one could have corroborated McGeady's account. Frank Gill, a former top executive, now retired, was no Microsoft-basher, and his opinion of McGeady was nearly as harsh as Gates'. Yet when I spoke to him, Gill's memory of the meeting was identical to that of Intel's chief troublemaker. I asked Gill if Maritz uttered the fateful phrase. "He said it," Gill replied. "When you're in business meetings, you often hear people say, 'Let's kill the bastards,' when they don't literally mean either 'kill' or 'bastards.' I really didn't think it was a big deal." But he was sure he'd heard Maritz say it? "Yes, I did, firsthand. I was there."

The third member of Microsoft's putative power troika of leadoff witnesses was Jim Allchin, a grade-A geek with a shock of white hair who was responsible for development of the company's core products. Allchin called himself "the Windows guy."

"We expected him to come in and claim that software was an arcane science, show a slick demo, and run circles around us technically," a DOJ lawyer told me. Instead, as media accounts have amply recorded, Allchin became the victim of the goriest and most celebrated evisceration of the Microsoft trial. Yet for all the drama of the tape-doctoring fiasco, Boies scored a more significant legal triumph two days earlier, when he walked Allchin through a different portion of the Microsoft video. It was a segment that enumerated the benefits – 19 of them, in all – of the "deep integration of Internet technologies" in Windows 98. Stopping at the first benefit, Boies asked Allchin: If a user took a PC running Windows 95 without an integrated browser and simply added a standalone retail copy of IE4, wouldn't that user get exactly the same benefit shown in the video? "Yes, I believe that's correct," Allchin replied. Boies moved on to the next benefit: same question. Eighteen more times he did this. Eighteen queries that began, "And again, sir." And 18 times, Allchin, his tone shifting from frustration to despair, answered in the affirmative.

Laborious though it was, this sequence of questions went "right to the heart of the Appeals Court decision" in the consent-decree case, Boies told me. The Appeals Court had said that tying two products together was legitimate only if doing so "offers advantages unavailable" from purchasing the products separately. With Allchin's 19-fold admission, Boies believed he had proven that Windows 98 didn't meet that test.

The humbling of Allchin – and of Schmalensee, and of Maritz – left the DOJ jubilant. Microsoft "put their home-run hitters at the top of their lineup," a government official told me. "And they all struck out."

The defense, meanwhile, was officially in disarray. The Wall Street Journal said so on its front page, in a blistering analysis by the reporter John Wilke, who quoted a number of economists – and not just economists but pro-Microsoft economists, culled from a list provided by the company itself – who flayed the firm for not conceding the obvious: that it did indeed try to eliminate competitors; that it was indeed a monopoly. In the Washington antitrust bar, Sullivan & Cromwell's performance was criticized as bordering on incompetent. Yet the question that remained was whether the fault really lay with S&C or even Neukom or whether it lay with Microsoft's chair. Some months after his testimony, Dick Schmalensee told a fellow economist privately, "The lawyers are not in charge. All the shots are being called by Gates."

After Allchin stepped down on February 4, it took Boies three weeks to dispatch the nine remaining Microsoft witnesses. A few of them, notably marketing executive Brad Chase, emerged relatively unscathed. For most, though, the trip to Courtroom 2 was like strolling through hell in a suit soaked with gasoline. Dan Rosen, the employee sent to testify about the June 1995 Netscape meeting, uttered such patent falsehoods that Boies felt no compunction about calling him a liar outright: "You don't remember that, do you, sir?" he asked at one point. "You're just making that up right now, aren't you sir?" As an email showed, Rosen didn't and he was. Robert Muglia, Microsoft's designated Java witness, prattled on so incessantly and nonsensically that, with no help from Boies, he drove Judge Jackson into a blind rage. "No, no! Stop!" Jackson roared, as Muglia attempted for the umpteenth time to explain that a Gates email didn't mean what it said. "There is no question pending!" the judge concluded, stalking out of court for a 10-minute recess.

Jackson's disdain for Microsoft's defense, never exactly a secret, became more and more evident as the trial wore on. One February afternoon, before gaveling court into session, he offered some words of wisdom he claimed were directed at no one in particular, but whose target in fact could hardly have been clearer. "The code of tribal wisdom says that when you discover you are riding a dead horse, the best strategy is to dismount," Jackson said. But lawyers "often try other strategies with dead horses, including the following: buying a stronger whip; changing riders; saying things like, 'This is the way we've always ridden this horse'; appointing a committee to study the horse; … declaring the horse is better, faster, and cheaper dead; and, finally, harnessing several dead horses together for increased speed." Jackson smiled and then turned to Boies. "That said, the witness is yours."

On February 26, after Microsoft's last witness, the judge recessed the trial for six weeks (in the end, it would be 13) before the start of rebuttals. "Use this time wisely," he told the lawyers for both sides, who were in no way confused about what that meant. For some time, Jackson had been quietly encouraging the parties to reopen settlement talks. Now he took steps to prod them in that direction. At a routine status conference on March 31, Jackson informed Microsoft and the government that he was imposing another novel procedure: After rebuttals were finished, he would divide the conclusion of the case into two phases. The first would be devoted to "findings of fact" and the second to "conclusions of law." By separating the facts from the law, Jackson was, in effect, ratcheting up the pressure on Microsoft to settle. Even a one-eyed tea-leaf reader could divine that he was going to come down hard on Microsoft when it came to the facts; that, for a start, he was almost certain to declare the company a monopoly, which could on its own inflict a fair degree of what Microsoft's lawyers called "collateral damage." If Microsoft was going to cut a deal, the time to do it was now, before Jackson had shown any of his cards.

Settlement talks took place sporadically that spring. They went nowhere. Though Microsoft was willing to contemplate some of the behavioral changes it had rejected in May 1998 – giving OEMs a real measure of control over the first screen, for instance – that was no longer enough for the DOJ and the states. Indeed, it was during these off-and-on talks that Klein first told Neukom the government was considering a structural remedy, perhaps even a breakup. Microsoft refused to address the subject, which seemed, Neukom said later, "ridiculous." In the last round of talks that spring, in June, Klein put forward a proposal for a broad set of conduct remedies, touching on everything from the pricing of Windows to opening up its APIs. In Microsoft's opinion, even that proposal was far too draconian to merit discussion. Yet the fact that the remedies were still only behavioral reinforced in Neukom the impression that Klein's earlier threat of structural relief was mere posturing.

Neukom had no idea how wrong he was. Since the fall, Rubinfeld and his economic team had been digging into the question of remedies, and the more they dug, the more attractive the idea of doing something structural became. Rubinfeld told me he was "very intrigued" by the notion of compelling Microsoft to auction off the Windows source code to its rivals – thus creating competition between, say, IBM Windows, Oracle Windows, Sun Windows, and so on. Many of the state AGs were keen on this idea, too.

And then there was Klein. Each time we met, he seemed to be moving a tiny bit closer to dropping the Big One. He had first suggested to me that a structural remedy was within the realm of practical possibility back in November 1998. By the spring, the word "divestiture" was popping up in our talks with increasing frequency. Klein had started down the path of suing Microsoft with extreme reluctance, yet here he was, entertaining a solution so hawkish it would make Kissinger blush. At one of our Saturday morning meetings, I asked him how he explained his conversion.

There was no mystery to it, Klein replied. "The nature of the problem and the pervasiveness of the practices in their corporate culture are far worse than I thought," he said. When the case began, all he could see was the tip of the iceberg; it was only after the evidence had piled up, first in discovery and then during the trial, that the full dimensions of the thing were clear. "That's what happens when you try a case," Klein explained. "You have instincts, you have views, then you go out and just rip it apart. And only then do you finally understand it."

He got up from his chair and walked over to his desk. "You wanna know how you try a case?" he said, picking up a small pewter mug that sat next to a paperweight. "This is how you try a case."

Taped to the side of the mug was a piece of scuffed white paper with a quote from T. S. Eliot's "Little Gidding":

We shall not cease from exploration And the end of all our exploring Will be to arrive where we started And know the place for the first time.

I handed the mug back to him. "So, if you win the case …" I started to say.

"If we win?" Klein laughed. "Get outta here!"

The spring settlement talks broke down in June, just as the trial's rebuttal phase was coming to an end. The rebuttals – highlights included another IBM exec, who had a diary full of details of threats Microsoft had allegedly made against Big Blue, and a command performance by AOL's David Colburn, who was as snarky as ever in answering Microsoft's questions about the AOL-Netscape-Sun alliance – were moderately entertaining, but did little to alter the trial's dynamics. By the time the rebuttals concluded on June 24, Jackson was openly using the dreaded M-word, monopolist, in reference to the beleaguered defendant.

A few weeks later, a new law clerk, fresh out of Harvard Law School, reported for work in Jackson's chambers. His name was Tim Ehrlich, and his first assignment was a daunting one: to write the initial draft of the findings of fact, and, in so doing, throw the book at Microsoft. Jackson made clear to Ehrlich that he wanted the findings to be unremittingly harsh. And so they were. Of the 400-plus paragraphs that composed the 207-page epistle, only one or two were remotely favorable to Microsoft, while the remainder of the document could have been written by the DOJ. Released on November 5, it was, Klein said to me, "the most damning document in the entire case."

The findings of fact reflected Jackson's assessment of the evidence he'd heard. But they were also designed to serve a tactical goal: to create the most powerful incentive yet for Microsoft to settle. On November 18, Jackson summoned the lawyers to his chambers and surprised them all by announcing his appointment of Judge Posner from the US Court of Appeals as a mediator.

In March, as the mediation Posner oversaw hurtled to its conclusion, the state attorneys general began to fret. Locked out of the negotiations taking place in Chicago, they feared the DOJ would sign on to a settlement based on conduct remedies that were weak, unenforceable, and riddled with loopholes. To assess the true impact of the proposals being swapped back and forth, they turned to Silicon Valley. In particular, they turned to Eric Hahn, a former Netscape executive who sat on the boards of several startups, including Marc Andreessen's new company, Loudcloud. Enlisted in March by California's attorney general, Bill Lockyer, Hahn would serve in secret as the states' unofficial technical adviser. It was Hahn who helped them devise the set of demands they sent to Posner in the final week of mediation. And after negotiations collapsed and Jackson issued his verdict, it was Hahn who helped them tackle the question of remedies, serving in the process as a crucial conduit to the Valley.

All along, the states had been more hardline about remedies – or at least more publicly so – than the DOJ was. A year earlier, in March 1999, at their annual convention in Washington, the AGs had presented a plan to force Microsoft to auction the Windows source code to rival firms. But when Hahn began looking into the feasibility of such a scheme, he quickly realized it would never get off the ground. The Windows code was constantly evolving, so what precisely would the licensee get? Making sense of the code would require Microsoft's assistance; how likely was that under the circumstances? Plus, whoever bought the code would be in competition with the firm whose programmers wrote it in the first place – not a tremendously appealing proposition. "I spent a week calling around, trying to find someone who'd want to bid on Windows," Hahn told me. "And I couldn't find a single company."

Given what they knew of the mediation, the states believed the Justice Department would never ask for a breakup. Like Neukom, they'd misjudged the DOJ. For months, Dan Rubinfeld, who'd left his post at the department but still worked as a consultant, had been advocating that Microsoft be chopped into two companies: one containing Windows, the other containing its applications and Internet businesses. Before the mediation died, Klein agreed in principle; now he agreed in practice. In a conference call on April 20, he informed the AGs of the DOJ's plan. Surprised and pleased in equal measure, 17 of the states signed on. (Only Ohio and Illinois dissented, asking exclusively for conduct remedies.) A week later, the government presented the breakup proposal, along with a list of conduct remedies to be implemented in the meantime, to Judge Jackson.

Microsoft's response was apoplectic. In the days leading up to the government's proposal, the company's honchos had adopted a defiant stance. In an interview with the editorial board of The Washington Post, Steve Ballmer said, "I do not think we broke the law in any way, shape, or form. I feel deeply that we behaved in every instance with super integrity." On television, Gates declared, "Microsoft is very clear that it has done absolutely nothing wrong." Now, with the breakup officially on the table, Gates decried it as "unprecedented," "extreme," "radical," and "out of bounds." Then came the ultimate insult: "This was not developed by anyone who knows anything about the software business."

One month later, on May 24, the lawyers from Microsoft and the DOJ gathered once more in Courtroom 2. It was brilliant spring day – bright, sunny, unseasonably hot. All along, Jackson had indicated that, if the government prevailed, there would be a separate process to deal with remedies. This was the hearing to begin that process. Everyone wondered what Jackson had in store. Microsoft asserted, given the severity of the government's proposal, that somewhere between several and many months were required to depose more witnesses, gather more evidence, hold more hearings. The DOJ disagreed, but assumed that the process would last at least a few weeks. But Jackson was determined to put this case on its path to appeal as quickly as possible. He'd decided that holding more hearings, in which eminent experts would offer conflicting predictions about the future of an industry that was inherently unpredictable, was a waste of time. And he was fed up with Microsoft: with the disingenuousness of its witnesses; with its failure to settle the case; and with the recent public comments of Gates and Ballmer, whose lack of contrition was so bald, so galling, that it would play no small part in his decision a few weeks later to cast aside his qualms and affirm the government's call for a breakup.

So when John Warden asked at the end of the day what the next step in the remedy process might be, the courtroom fairly gasped when Jackson said, without missing a beat, "I'm not contemplating any further process, Mr. Warden."

Five minutes later, the Microsoft trial was over.

IX. WHISTLING IN THE DARK

Early in August, eight weeks after Judge Jackson decreed that Microsoft be rent asunder, I went back to Redmond to see Gates again. In the software heartland of the great northwest, the spring and summer had been two mean seasons, and not simply because of the judge's rulings. The long-awaited launch of Windows 2000 in February had proved lackluster. The company's revenue growth was flagging, especially in its core OS business. Analysts had slashed their forecasts for the coming year by $1 billion or more – another reason the stock was taking a drubbing. In June, Microsoft unfurled its master plan for the age of the Net with the blaring of trumpets and the rolling of drums. Dubbed .NET, the initiative was, Gates said, a "platform for the next-generation Internet." But while everyone agreed that .NET was bold and ambitious, they also concurred that it was not fully baked. For the press, .NET was a one-day story; for much of the industry, a one-day shrug.

On Microsoft's campus, frustration had turned to a sense of defeat. "At the club, in the steam room, people who used to talk about the great things we were doing, all they want to do now is give you opinions about the trial," senior executive Craig Mundie told me. "Even family members are like that. It's discouraging." After talking for an hour about the challenges of taking on AOL with young marketeer Yusuf Mehdi, who'd switched over from Windows to work on the MSN portal, I asked him if the trial had affected morale. "There's been disruption, for sure," he said, "but there's also been a circle-the-wagons mentality, which is good, in a way." Mehdi paused. "My mom asks me, though, 'Yusuf, is Bill really that bad?'"

Then there was the exodus: For the first time in Microsoft's history, the company was hemorrhaging talent. The bleeding went from top to bottom, from high-profile pashas like Nathan Myhrvold, Greg Maffei, Brad Silverberg, and Tod Nielsen, to browser warriors like John Ludwig and Ben Slivka. In several cases, executives who had testified in the trial – Eric Engstrom, say, or Nathan's brother, Cameron Myhrvold – left the company almost the instant they stepped down from the stand. By Microsoft's count, around 50 employees were peeling out every week, but by some estimates the number was three times that high. Some left in search of dotcom riches, others for the thrill of running their own show. Some had grown weary of Microsoft's size and proliferating bureaucracy. When Paul Maritz announced his resignation three months after my visit, it drove home vividly the brutal truth: Microsoft was no longer the place to be.

Microsoft's reaction to the departures was stunning. In an industry spurred by the sparking of synapses, Gates had long recognized that the most precious raw material was gray matter, and Microsoft prided itself on acquiring only the best. But now I kept being told that many if not all of the big names who had left – men who'd run large swathes of Microsoft when the company was at its zenith – were in fact dead wood; that Ballmer had merrily chucked them out the door. When I asked the new CEO if this was true, he shrugged and smiled. "We've lost senior people who I wish hadn't left, and we've lost senior people where I'm fine, I'm happy, it's OK," Ballmer said. "We've got both categories, and we may have more of the latter than the former." Microsoft's marketing chief, Mich Mathews, remarked to me, "We could lose 40 percent of the IQ in this company and still be the smartest." She said, "All we really need are three smart guys."

When the smartest of the smart guys announced in January that he was handing the CEO reins to his best friend and taking up the title of "chief software architect," some observers wondered how meaningful it was; surely the buck would still stop with Bill. Yet Gates relinquished more control over the company than even many Redmondites expected. In short order, Ballmer began instituting new processes and disciplines on Microsoft's operations. And he began systematically replacing Gates' team of top managers with his own posse. "Bill and Steve have different attitudes about people," an ex-Microsoft exec says. "Bill likes smart people – plain smart. Steve likes people who get shit done."

Officially, the rationale for Gates' decision was twofold: First, the company had become too unwieldy for one person to be both chair and CEO; and second, Gates yearned to get back to the role he'd played in the company's early days, when he was intimately involved in the design and development of its key products. Yet many of Gates' friends and colleagues believe that the antitrust suit played a part here, too; that it had worn him out, beat him down, and induced him to seek a less strenuous role. "It's all been very hard on Bill – I mean physically; it literally made him sick," Greg Maffei says. "I think the reason he's no longer CEO is directly attributable to this experience with the courts and the government."

Gates had given no published interviews since Jackson's breakup order, so I had little idea what to expect when I came trundling into Building 8 this summer. Outside, the soft morning sun sent a stream of pale-gold light through Gates' picture window. The first thing I noticed was that he looked as if he'd been spending some time outdoors; his skin tone was closer to ecru than its customary eggshell. He seemed thinner. His greeting was warm and full of good humor. As we settled into our chairs for our hour together – which would wind up being closer to two – it soon became clear that whatever had drained the juice out of Gates before our last meeting, his tank had been refilled during the 18 months since.

Gates was clearly relishing his new role as chief software architect. In abundant detail and with avid enthusiasm, he described the genesis of .NET, its technical underpinnings, and his role in its concoction. He sang the praises of XML, of distributed computing; he lectured me with verve about "probabilistic input APIs" and "loosely coupled message-based programming." Back in 1995, Microsoft had embraced the Internet, he said, but only as a feature. "It was the most important feature – but it was still a feature," he explained. Now everything would be different. With .NET, the embrace was total; the Net was all.

When I'd mentioned Sun in our previous interview, Gates' response had been as banal as it was disingenuous: "Every comment I've made about Sun has been positive – Sun's a good company." Now I raised the subject again, pointing out that Sun's software wizards, who'd invented Java and Jini, had been talking for years about many of the ideas Gates was discussing today; they contended that .NET was, at bottom, an endorsement of their corporate motto: "The network is the computer."

Gates, who'd been pitching back and forth in his chair, like a hummingbird at a feeder, dug his heels into the carpet, propelled himself bolt upright, and flapped his arms. "The most nonsense I've ever heard!" he exclaimed. "But it's not unexpected. The business model of Sun is to sell overpriced hardware." When it came to solving the complex software problems that .NET was addressing, he said, "Sun's not involved in that. Sun has never had anything to do with that."

At the launch event for .NET, Gates had called the initiative a "bet-the-company thing." Didn't it worry him to be undertaking such an ambitious project at precisely the moment when so many of his best and brightest were flying the coop? "Look at the top of this company," Gates shot back. "We've had more continuity of management leadership than any technology company ever." Maybe so, I said. But doesn't it hurt to lose a Nathan Myhrvold? To lose a Brad Silverberg? "It doesn't diminish our ability to do .NET, absolutely not," he said. "We have a team here that is the best software development team in the world. It just shows the embarrassment of riches Microsoft has had, that even without those two guys, we can go and do phenomenal things. But those are great guys. If they want to come back and work here, I'll take them in a second."

But not many of the others, apparently. Did Gates, like Ballmer, regard some of the senior people who'd left as dead wood? "I won't name their names, but certainly," Gates said. "Come on, give me a break. This is not simple stuff."

I mentioned that Craig Mundie told me, "The trial has significantly diminished our ability to attract and retain people of the highest caliber." Between the shadows cast by the DOJ and the allure of Internet startups, did Gates think it would be increasingly difficult for Microsoft to replenish its pool of human capital?

"It's a very competitive environment for getting smart people," he replied. "People think, 'I'll go do an IPO and be rich tomorrow.' I don't promise them anything like that. I promise them more impact." Gates went on: "So many startups are doing the same things and terribly short-term things. B2C? That fad is gone. B2B? That's in the fad stage right now." Yet for those with limited interest in fads, Gates said, Microsoft retained a powerful allure. "The things we care about are long-term things, tough things. We can afford expensive things. We build 747s. We don't build Cessnas."

After a while, we turned to the trial. To many observers, the most inexplicable of all of Gates' and Ballmer's actions had been the unrepentant poses they'd struck in public in the period after Judge Jackson issued his verdict but before the DOJ or the states submitted their proposed remedies. In private, they were even more strident, with Gates telling a gathering of Microsoft employees that the company was a victim of a "travesty of justice," that "we are absolutely confident we will win on appeal," and that they would "never allow" Microsoft to be broken up. In the aftermath, a number of state AGs had cited Gates' and Ballmer's public comments as a "slap in the face" and said that they'd figured in the decision to ask for a breakup. Judge Jackson himself told The New York Times the comments "astounded" him and helped make a breakup "inevitable."

With employee morale dismal and the stock price plummeting, Gates and Ballmer must have figured that anything less than an adamant stance would have sent a terrible message to the troops. Still, I asked Gates if, on reflection, he thought those remarks were a tactical blunder.

"You can accuse us of having put Internet support into Windows," he answered. "You can fault us for contributing significantly to the PC market and what that's meant for the software industry and prices and all of those things. We believe that what we've done is absolutely pro-competitive, and it's our right to stick up for that."

I understand you have the right, I said. What I'm asking here is a tactical question. It was a moment of great political sensitivity. Wouldn't it have been better to keep your mouths shut?

Gates shot me a look of contempt. "We are defending principles of great importance," he said. "Our right of appeal. Our right to innovate. Our right to have an appeals court sit and judge that." Even to mention tactics, he seemed to be suggesting, would sully those principles with the grubbiness of politics.

Another thing Jackson told The New York Times was that he didn't think Microsoft had taken the case seriously enough. Had they? "Hey, you should see our legal bill – are you kidding?" Gates quipped. "Of course we took this seriously."

The conventional wisdom was that Microsoft and its lawyers had made hash of the case from start to finish. They had failed to settle before the trial started and after it was over. In between, they had waltzed into a federal court and tried repeatedly to claim that day was night and night was day, that up was down and down was up, that words with clear meanings were somehow ambiguous – or even meant the opposite of what they plainly said. They had defended a position – that Microsoft was not a monopoly – even pro-Microsoft economists regarded as untenable.

With the benefit of hindsight, are there things that you regret? I asked. Where you look back and think, We made a mistake?

"Understand," Gates said, "that this is an attack on our ability to add new features to Windows, so it's not the kind of thing where you can say, 'Oh, oh, that? Oh, sure. We'll give that up.'" In the end, he believed, the law was on their side. "Every action we took that's been attacked in this case is Microsoft working on behalf of consumers, working in exactly the way we should work."

There was not much more to say. In the face of overwhelming evidence to the contrary, this was Gates' bottom line: They'd done nothing wrong. They'd made no mistakes. In the end, they'd be exonerated. And everything at Microsoft was A-OK. There was no hint of artifice in any of his statements. I believed he believed every word he was saying. It was one of those moments where you question yourself. Is this man hallucinating? Or does he glimpse a reality that I'm too blind to see?

Either way, it raised another question: Given what you believe about yourself and Microsoft, how does it feel to have the US government calling your company crooked and calling you a lawbreaker?

Gates stared out the window and thought about the question a good long time. Still gazing at the trees, he began, "There's a certain irony to being in a situation where we literally have to bet the company on an unknown business framework and a new set of technologies just to stay in any type of position at all, that we have to do that, that this is the most competitive market the world has ever seen. The notion that somebody could come in and say (a) we're a monopoly, (b) we shouldn't be able to add features to our product, and (c) throw a little mud in the process – the irony is deep. Very, very deep."

Has the whole thing left you cynical about the legal process?

"No," Gates said simply.

I said I found that hard to believe.

"The law is interesting," he mused. "The US judicial system, like, 98 percent of the time works extremely well." For the first time in a while, Gates looked me in the eye. "This case, in the final analysis, will be part of that 98 percent."

X. THE VERDICT

As the old economy gives way to the new, some of the most profound questions in public policy revolve around how a legal regime conceived and enacted in the industrial era applies to the information age – if, indeed, it applies at all. Whatever the ultimate outcome of United States v. Microsoft, the case promises to yield a historic precedent, one which will shape fundamentally the terms of competition in the dynamic high tech markets at the center of our emerging postindustrial order. "I cannot imagine a more important case for the future of antitrust," Dan Rubinfeld says. "If our victory is upheld, it will set the rules of the road for years to come. If it gets overturned, almost anything goes."

In June, the government asked the Supreme Court to bypass the federal Appeals Court and take the case as soon as possible. Microsoft opposed the request, arguing that the company is entitled to pursue the full appeals process through its normal course. There is no mystery about either side's motives. In light of the Appeals Court ruling in the consent-decree case, the DOJ wanted badly to avoid that venue, while Microsoft wanted just the opposite. When this article went to press, the path to a final verdict remained uncertain. But whether or not the Supreme Court decides to take the case now, most experts assume that the matter will wind up there eventually. Although some pundits have speculated that a new presidential administration, particularly one headed by George W. Bush, might drop the case, that would be highly unusual. It would probably also be irrelevant, for the states would be likely to pursue the lawsuit to the bitter end.

Microsoft's arguments on appeal will be many and varied. In its filings to the Supreme Court, the company accuses Judge Jackson of "an array of serious procedural and substantive errors." On the substance, Microsoft's strongest arguments involve tying, where the law is fuzzy – especially when it comes to software – and the weight of precedent favors defendants. On the procedure, the most glaringly questionable of Jackson's actions was his handling of the remedy phase. Even among antitrust scholars who believe the government clearly proved its case, there is widespread incredulity that Jackson devoted only one day to the complex question of chopping Microsoft in two. "I have nothing good to say about it," comments William Kovacic, a law professor and antitrust specialist at George Washington University. "It was a tremendously glib way to handle a very serious process."

Glib or not, Jackson's refusal to entertain further testimony or examine further evidence was rooted, at least, in one home truth: Nobody really has a clue what a breakup would mean. In Silicon Valley, there are plenty of sensible, intelligent executives who think the government is right; that halving Microsoft would unleash competition and let innovation reign. But there are also plenty who believe the opposite; that they'd simply be saddled with two Microsoft monopolies instead of one. There are those who argue that the applications company would flourish while the OS company withered; and there are those who say that both would be doomed. Would consumers benefit from a breakup or would they suffer? Would shareholders prosper or would they get hammered? For every two questions, there are at least five theories. Putting aside prognostication, only one thing is certain: A breakup would mean the end of Microsoft as we know it.

Yet all the speculation about the effects of a breakup obscures a simple but staggering fact: We are already witnessing the end of Microsoft as we know it. For the past three years, Gates and his company have been caught in a pincers. Pressing in from one side has been a technological shift more sweeping than any since the rise of the PC: the Internet. And pressing in from the other has been a menace more threatening than anything Microsoft ever encountered in the world of business: the United States government. For a lesser company, either of these forces alone might have spelled ruin. But it took both, working in devilish harmony, to put Microsoft on the path to a new identity.

On Microsoft's campus they can sense the transformation, but they struggle for the words to describe it with precision. When I visited in June, the people I spoke to were more apprehensive about the future than I had ever seen them. With the company on the verge of turning 25 years old, middle age was encroaching; could Microsoft stay vital? "The question is: Do we recede or do we maintain our leadership?" asked Craig Mundie. "Or are we superseded by another company that rises up and takes leadership? People say .NET is a 'bet the company' thing. But companies don't roll over and die. The question is whether we become just another company."

When Nathan Myhrvold still worked at the company, he had an expression for just such occasions – "putting a name to the nameless dread." After talking for a while with Mundie, though, it began to come clear that the dread Microsoft was feeling had a name after all. "Either Microsoft will stay Microsoft or it will become IBM," he said. "That's just my opinion. But I think those are the stakes involved in this transition."

At the dawn of the PC era, when Big Blue's power was as yet unchallenged, the new technology had presented the company with a choice: Resist it, ignore it, or get with the program. IBM opted to get with it – or at least to make the effort – and for several years it dominated the market. But the forces of change unleashed by the PC were too swift, democratic, and decentralizing to contain. Today, IBM is still the biggest manufacturer of mainframes in the world. It has a stock market value of more than $200 billion. It has happy shareholders, happy customers, happy employees. Yet few people fear it, or follow it; nobody considers Big Blue a leader anymore.

Microsoft today is on an eerily similar trajectory. Just as IBM embraced the PC, so Redmond embraced the Internet. Yet with the dotcom boom, Microsoft's position seems, if not tenuous, then increasingly peripheral. The real estate it controls, the PC desktop, remains the most valuable territory on the digital map. But, as everyone can see, with the rise of the Net, the universe of computing is expanding and exploding, while the desktop seems only to be shrinking in strategic importance.

Andy Grove finds the parallel compelling. "For a long time in the 1980s, IBM was everything to Intel," he told me. "We thought about them constantly, lived and died by their whim. Then around 1990, I woke up one day and it wasn't so anymore. It wasn't some momentous event. And now it was Microsoft who we thought about all the time. Maybe this is happening again – only this time, instead of Microsoft being replaced by another company, it's being replaced by the Internet, by a whole bunch of things happening all at once."

Creeping giantism has begun to take hold at Microsoft, too. Microsoft's goal has long been to retain its agility even as it grew – to be "the smallest big company around," as MSN executive Brad Chase puts it. Yet Microsoft has become a very big company, with 40,000 employees worldwide. Though that 40,000 includes the largest concentration of skilled coders anywhere on the planet, the culture of the company has lately begun to smell as much of marketing and sales as it does of technology – a distinctly IBM-ish aroma. At the same time, the sheer scale of the software endeavors into which Gates has plunged Microsoft's programmers has a certain whiff of old IBM as well. The Gates who boasted to me about how Microsoft "builds 747s" is the same man who, in the 1980s, used to mock Big Blue's programmers by saying that IBM's motto was: "Building the world's heaviest airplane."

Meanwhile, Microsoft's well-known insularity has taken on a new dimension. In their heyday, Gates and Ballmer were relentlessly in touch with the industry they sought to rule. On the floors of trade shows, in the hotel ballrooms at high tech conferences, they picked brains, probed for clues, and tested their assumptions against the prevailing wisdom. No longer. Hemmed in by his wealth and fame, Gates attends few industry events these days, and when he does, his appearances are scripted; spontaneous exchanges are strictly verboten. And even among his fellow info-tycoons at Herb Allen's annual schmoozefest in Sun Valley, Gates is known to keep largely to himself. (Kay Graham and Warren Buffett are the only guests with whom he routinely socializes.) As for Ballmer, when the new CEO was invited this summer to speak at one of the Internet industry's preeminent conferences, the organizers were rebuffed with a message from his handlers: "Steve says he doesn't speak at conferences where he doesn't have any customers."

There is one other parallel between Microsoft and IBM, and the irony here is thick. IBM's entanglement with the government had paralyzed the company. By doing everything in his power to avoid such paralysis, Gates brought the government slamming down on Microsoft. The demoralized employees, the slumping stock price, the cloud of uncertainty hanging over Redmond – in a way, all of it was due to Gates' IBM phobia. By trying to avoid Big Blue's fate, Gates had instead done much to guarantee it.

Not surprisingly, the suggestion that Microsoft might wind up as the new IBM is something that Gates and Ballmer aren't willing to countenance. When I asked Ballmer if it would be a bad fate to be perceived that way in five or ten years – as successful and solid but no longer dominant – he nodded his head in violent agreement. "Yes," he said. "Terrible? No. Bad? Yes." When I asked Gates the question, he answered as emphatically: "Absolutely."

Gates imagines for himself a rosier future. Though he told me he could envision a day – in his fifties – when he would no longer be Microsoft's chair, he was "excited," he said, that "in these next couple years, I'll get to do some of my most interesting work." To the extent he admits his reputation has been muddied, he assumes, like John D. Rockefeller, the plutocrat with whom he is so often compared, that he will be vindicated. But, where Rockefeller believed his vindication would be dispensed by history and in heaven, Gates expects to receive his very shortly – and here on Earth. According to polls, he remains one of the most admired figures in the world of business. And his $21 billion charitable foundation has made him a hero in the world of philanthropy. The only thing missing is the higher court reversal he so manifestly considers his due.

Yet even if that reversal comes, it might provide Gates with less satisfaction than he hopes. "Vindication will be bittersweet," a Microsoft official told me. "The company has suffered too much. Before, people thought the world of us. That we were great innovators. That we were this great engine of the new economy. Now, either the decision stands, in which case people think we're criminals, or the decision is overturned, and people think we somehow got away with something. No vindication will erase that stain."

Truer words were never spoken. Before the Microsoft trial began, Gates was more than a high tech hero; he was the pristine embodiment of the high tech myth. At an impossibly young age, he'd come out of nowhere, consumed with ideas and a pure burning passion. He had launched a company that unleashed an industry, and then led that industry as it transformed an economy. For a long time, Gates represented everything that was inspiring about this protean phenomenon taking shape in our midst – its freshness and its ambition, its sense of possibility and its connection to the future. But like a figure lifted from classical tragedy, Gates sowed the seeds of his own undoing. He created a company that reflected his image and fostered a culture that fed his sense of omnipotence. He mastered a business that rewarded farsightedness, but failed to develop his peripheral vision. In his arrogance he lost whatever perspective he once had, and in his monomania he was unwise to the ways of the world. He began his journey as an aspiring god, an illusion his universe nurtured and sustained. When his reckoning came, it was shocking and final – and it seemed somehow ordained by the ages. For the wreckage of the trial revealed that Gates was mortal.

• magazine-8.11