Making Microsoft Safe for Capitalism
       click here to go back to the home page

also Before he installed Windows 95, John Dodge connected to the Internet using software from a Microsoft competitor, CompuServe's Internet in a Box. Not anymore—Windows 95 silently disabled a key piece of his setup and made it too difficult for him to reinstall it.
  Dodge is no novice. He is senior executive editor of the trade journal PC Week and so had access to the highest-level support engineers. But life is short and even software professionals learn to take the path of least resistance—in this case, the path leading to Microsoft. He has become a regular user of the new Microsoft Network, though he has trouble with its Internet features.
  Still, he believes Microsoft executives when they deny trying to gain market share by sabotaging competitors' software. He just wonders whether Microsoft "has a full appreciation of its actions in the market place."
  There is reason to believe that Microsoft does.

Microsoft vs. the World

The Government's lawyers are engaged in the third major phase of an investigation that may prove to be the most important, and the most difficult, in the century-old history of antitrust law. Its target is a scrappy, young, fast-moving company with a mere 18,000 employees—a fraction of the size of I.B.M. and A.T.&T., the last great subjects of antitrust action. Microsoft does not control a manufacturing industry (as I.B.M. did), a natural resource (as Standard Oil did) or a regulated public utility (as A.T.&T. did). Microsoft's strategic monopolies—for it does possess and covet monopolies, despite vehement denials from its lawyers—are in a peculiarly subtle and abstract commodity: the standards and architectures that control the design of modern software.
   In a historical eye blink, as the technologies of computing have come to pervade the world's economic life, Microsoft has turned 20 years old. When Ronald Reagan became President, Bill Gates's new company was an unincorporated partnership with accounts kept in handwritten ledgers. Apple was a big new personal-computer company, worth $3 billion; I.B.M., the mainframe giant, was cobbling together its first personal computer out of parts from outside suppliers. By 1990, just a decade later, Microsoft had become the world's richest software company, though it had no leading product in any important category but operating systems. Today nearly half of the world's total P.C. software revenue goes directly to Microsoft.
   "I personally believe that Microsoft is the most powerful economic force in the United States in the second half of the 20th century," says Eric Schmidt, chief technology officer of Sun Microsystems—a minicomputer and networking company whose business used to be remote from Microsoft's but now finds itself under direct competitive pressure. Some of Microsoft's control over computing, at all levels, is obvious. Much, however, is invisible. Even longtime insiders are just beginning to understand the nature of that power: how Microsoft acquired it, preserves it and exercises it.
   "The question of what to do about Microsoft is going to be a central public policy issue for the next 20 years," says Mitchell Kapor, the founder and former C.E.O. of Lotus Development Corporation—once the leading P.C. software company. "Policy makers don't understand the real character of Microsoft yet—the sheer will-to-power that Microsoft has."
   The vast majority of the world's personal computers—estimates range from 80 percent to more than 90 percent—run on Microsoft software from the instant they are turned on. Yet, pervasive as P.C.'s are now, Microsoft has made clear that they are only the beginning. The company is working toward wallet computers that carry digital signatures, money and theater or airplane tickets; toward new generations of fax machines, telephones with screens, and car navigation systems; toward Microsoft-run interactive television boxes, office networks and wireless networks, and, most potently, toward an aggressive Microsoft role in the Internet itself.
   By making connections among all these levels of modern computing, and by exerting control over the architectures that govern those connections, Microsoft is in the process of transforming the very structure of the world's computer businesses. "Microsoft is imposing a new verticality on the industry," says Gary Reback, a Silicon Valley technology lawyer who represented a group of anonymous Microsoft rivals in the antitrust proceedings. "Bill's been able to exploit the market far better than anybody else has, and I think that's because he intuitively under stands what enormous power he has and how to exploit that power."
   It is a software company with the broadest possible understanding of software: not just computer code but books, news services, music, movies, paintings, maps and directories of people and businesses. It believes that you will buy all these on line, and it intends to deliver them. With its new Microsoft Network, providing both an on-line service and Internet access, it is focusing on electronic financial-transaction processing—which is to say, all electronic commerce; which is to say, at least in some visions of the future, pretty much all commerce. "Basically what Microsoft is trying to do is tax every bit transition in the whole world," says a senior executive of a competing software company. "When a bit flips, they will charge you."
   Its profit margins are staggering by the standards of manufacturing companies—it salts away about a quarter of every dollar that comes in, compared with about 3 cents for Apple. It sits on an enormous reserve of cash. Among modern corporations it has been an unparalleled generator of personal wealth. Never mind that its founder and chairman may on any given day be the world's richest person; the third-richest Microsoft executive, Steve Ballmer, owns close to $3 billion in Microsoft stock, and 2,000 or more of its employees have be come quick millionaires, creating a remarkable new class structure in Seattle's social and political life. In a less-charged era, Gates and Ballmer both occasionally joked about their goal of world domination. Now they are more careful. Microsoft's people are taught to avoid using the word dominate in public discussion of the company's role in any part of the software business; the preferred word is lead.

There are many, many articles that say Microsoft is about to fail," Gates tells me in a hasty interview on the eve of a vacation in China. (Only later does it occur to me that he must have Satan's own clipping service.) "Those two extremes are silly beyond belief. We won't fail tomorrow, and we don't have a guaranteed future. That's just logical." It has become an article of faith—with considerable help from Microsoft —that no credible threat exists to its monopoly in operating systems for personal computers or its rising dominance in all P.C. software. This summer, during the orchestrated build-up for Windows 95, Wall Street found that Microsoft is the company that drives the American financial markets as only I.B.M. and General Motors could in previous eras. The closing months of 1995 see competitors and potential competitors—I.B.M., Apple, Lotus Development, WordPerfect, Novell—fading back from Microsoft's businesses or bracing fatalistically for the next onslaught.
   Gates is back now from his vacation: a personal trip, but he did find time to meet formally with President Jiang Zemin, and Beijing announced—no Antitrust Division there—that it was declaring Windows to be the country's official software standard. Days later, Gates revealed that he had bought the world's greatest storehouse of historical photographs, the Bettmann Archive, adding to his already unchallenged collection of visual images. Two more markets cornered, it seems. The Government must ask now, as the computer business is asking, whether a dangerous threshold has been crossed—whether a single force has taken control of the most tempestuous, inventive, unpredictable industry of our time.

Cult of Bill

By now it's well known that where other companies have offices, Microsoft has a huge and verdant campus—low-slung steel-and-glass buildings set amid stands of evergreen trees in a Seattle suburb, with softball fields and basketball courts and an artificial pond called Lake Bill. Most employees still have private offices, and soft drinks are still free, but the campus has lately taken on an air of relative maturity. It's full; a new set of buildings are on the rise across the freeway. A soccer field was torn up this summer to make way for the extravaganza of the Windows 95 launch.
   Everywhere, though, is a sense of the forceful influence of the company's 40-year-old leader, who at One Microsoft Way is always referred to simply as "Bill." Bill's so smart, says a character in Douglas Coupland's new novel, "Microserfs" (another Microsoft first: popular fiction inspired by its wondrous corporate culture). Bill is wise. Bill is kind. Bill is benevolent. Bill, Be My Friend . . . Please! A heightened casualness does strain the voices of Microsoft middle executives when they drop mention of "face time" with Bill.
   Sometimes people at Microsoft say that they are a mere surfboard and Bill is the man who rides it. The company went through a series of short-lived presidents before finally realizing that a president in the presence of Bill was an impossibility; now there is just an Office of the President, occupied by a group of vice presidents. "One of the things that makes us work today is the incredible brain capacity, memory capacity that Bill has," says the most senior of them, Ballmer, striding energetically around a tiny conference room.
   Microsoft wears the personality of its leader like a wet suit. Gates's mind-set might be described as a blend of ruthless competitiveness and planned paranoia. He chooses to be scared; he wants his company to be scared. At the moment it is the explosive rise of Internet that scares him most. At similar critical moments in history -- "discontinuities," as he accurately puts it -- he has watched most of his competitors stumble and fail, beginning with I.B.M. He goads his employees with fear of failure. It may help that Microsoft is the company that the est of the industry loves to hate.
   Accusations that Microsoft's people lie, cheat and steal information are as much a part of the company's lore as its cadre of millionaires with FYIFV (". . . I'm fully vested") buttons. Microsoft knows it has clout, and it uses what it has: to pressure small competitors, trade-show operators, journalists, retailers (shelf space for non-Microsoft software will be at a premium this Christmas) and everyone else.
   "Can you name anybody that's happy about being in the same industry with Microsoft?" Mitchell Kapor asks.
  Microsoft lives according to a "thin ethics," as he sees it: "Anything not a direct lie or clearly illegal is O.K. to do and should be done if it advances Microsoft's tribal cause. This licenses the worst sorts of manipulations, lies, tortured self-justification and so on." Microsoft is hardly alone, of course; plenty of its competitors would play as rough, if they only could. Others in the industry suggest that Microsoft's small-company scrappiness has kept it from facing the issue of corporate ethics: behavior that people will forgive, or at least understand, in a start-up looks considerably less attractive when David grows into Goliath.
   Microsoft stumbles, but less often than its competitors; and when its competitors make mistakes, Microsoft has historically managed to take advantage. It has cultivated an aura of inevitability. It has failed so far to overcome some rivals, but it has never lost an important franchise once gained. And if Microsoft people are now openly contemptuous of the Government's multiphase investigation of its trade practices, it is Gates who sets the tone. This spring, when the second phase ended with Microsoft's dropping a proposed acquisition of the financial-software maker Intuit, Gates said sarcastically, "In the future we may wait a week or two before we decide to do something like this again."

None of the above appears in the entry on Gates, William Henry, III (1955- ), in the world's best-selling multimedia encyclopedia: "Much of Gates' success rests on his ability to translate technical visions into market strategy, and to blend creativity with technical acumen. . . ." There is a picture, too, with a sound clip: "Microsoft was founded based on my vision of a personal computer on every desk and in every home. We've never wavered from that vision."
   Needless to say, that's not the Encyclopaedia Britannica, now struggling for its life. The leading encyclopedia in the multimedia world is Microsoft's own Encarta—a glossy retread of the old Funk & Wagnalls, updated with pictures and audio bits. Microsoft is rapidly accumulating best-selling entries in every reference category: general desk reference; movie guides; music guides; cooking and wine guides. Most of these were licensed or bought outright, but Microsoft's consumer division is gearing up to produce more and more of its own material for CD-ROM's and on-line information products. Its new Digital Cartography Lab alone employs 15 highly trained cartographers and geographers, working on a new generation of digital maps. (Hammond, Rand McNally—are you ready?) Over at the Microsoft Network, a fledgling news staff produces a sort of electronic front page every day.
  And by the way, the unabridged version of that famous Gates motto is: "a computer on every desk and in every home, all running Microsoft software."

Microsoft vs. the Internet

Not only is the new Microsoft Network software automatically set up for every Windows 95 user; its icons—"MSN" and "The Internet"—are an astonishingly persistent feature of the "desktop" that stares at you from your screen.
   "Does anyone know how to get rid of the Internet Explorer icon so that I can put my Netscape Navigator icon in its place??" asks a Windows 95 user on the Microsoft Network. Over on CompuServe, a user says, "I want the MSN icon to go away, but I don't seem to be able to delete it. How do I get rid of the thing?"
   That's what Steve Case wants to know, as president of America Online, the most popular commercial on-line service and one of the companies with the most to lose. "The tens of millions of existing computer owners who are expected to upgrade to Windows 95 won't be offered choices built into their operating system other than MSN," he says. "The operating system for 85 percent of all personal computers is about to become an exclusionary marketing and distribution tool."
   He has sent the same message to the Department of Justice. He argues that the operating system is to a computer what the dial tone is to a telephone: the thing you have to use to go anywhere at all. Just as the Antitrust Division eventually prevented A.T.&T. from using its local-telephone monopolies to perpetuate a monopoly in long-distance service, so it should prevent Microsoft from leveraging its operating-system monopoly into the new territory of Internet and on-line services.
   The Internet has forced Microsoft to make a late change in its on-line strategy. As little as three years ago, when MSN was a vigorously leaked secret code—named Marvel, the on-line landscape comprised thousands of hobbyist bulletin boards and just three giant commercial services: America Online, CompuServe and Prodigy. The Internet, meanwhile, was obscure, academic and seemingly irrelevant to any vision of electronic commerce. Marvel was designed as an America Online-CompuServe-Prodigy killer: a private service that would host proprietary content from newspapers, television networks, Microsoft's own consumer-product sources and a wide range of businesses with information and products to sell.

Microsoft was not the only company caught by surprise when the Internet burst into public view, and it was one of the quickest to begin a recovery. It took the unusual step of buying a minority stake in an Internet access company and building a nationwide network that customers will be able to use for dialing into the Internet—paying, of course, by the month or by the hour. Gates insists that Microsoft will remain strictly a software company—"We're not in the connectivity business; we're not in the business of owning wires"—but by last year it was clear to Microsoft, as well as the big on-line services, that Internet access was essential. And Microsoft determined to provide it by means of a single button on the Windows 95 desktop.
   But that button is only the beginning of Microsoft's strategy. In a confidential memo to 14 senior executives last year, Gates described the rise of electronic communication as a "sea change" and warned that in one category, the sharing of documents among groups of co-workers, "embarrassingly we find ourselves some what behind one of our old rivals"—Lotus.
   It is a revealing document, with a mixture of goading and exhortation, of futuristic vision and rock-hard attention to Microsoft's singular economics. Nothing matters more than persuading users to pay for upgrades to their software. In mature product categories like word processing, he notes accurately, users will not upgrade or switch products merely for the sake of a few extra features, but they will if the new software takes advantage of a sea change. "It takes even more guts," he wrote, "to bet on the Sea Change when you are the market leader but it is the only way to position yourself for massive upgrades."
   Every software division at Microsoft is now redesigning its products to take advantage of a world in which every computer can talk to every other. The next version of Microsoft's CD-ROM encyclopedia can be updated live through the connection to MSN or the Internet. For word processors, integration with the Internet means thinking not in terms of personal documents at home or even work-group documents on your private office network, but in terms of browsing, searching and publishing on line. For spreadsheets, it means viewing and manipulating data that comes across private and public networks, interchangeably. "Excel must blow away the competition," Gates urged in the memo. "The basic point, however, is that users' expectation of what Office applications will do is changing and three to four years from now anyone forced to use the software we have today would find it completely inadequate for dealing with the electronic world."
   Nathan Myhrvold, one of Microsoft's chief strategists, sums up the attitude now driving every company division: "The Internet is an example of a revolutionary shift that, if we forgot about it, would eventually kill us. The notion that you would do a task on the desktop with desktop software in a few years that didn't involve the Internet is just ludicrous."
   Microsoft has already tightly integrated its Internet access into the new Windows 95 environment. Addresses for all kinds of Internet resources can be dragged onto the desktop, where they appear as colorful icons of their own; dragged again into E-mail messages to be shared with friends; and clicked on to begin an automatic dialing process. The Microsoft Network as an on-line service has its problems -- performance is sluggish and the content thin—but as new computers stream into the marketplace with Windows 95 already installed, millions of newcomers will find their way to the Internet by clicking that Microsoft icon.
  Hence the extra annoyance of its competitors over the little matter of Windows 95's disabling their users' existing Internet access. Many users who had installed the widely popular Netscape browser and then tried Microsoft's Internet Explorer discovered that Netscape would no longer work. The same problem affected users CompuServe's Internet in a Box software.
  "Windows 95 includes a process that disables your Internet account," says David Pool, a top CompuServe executive. "And that's just the tip of the iceberg of the inappropriate things Microsoft does from a networking standpoint. It's a clear extrapolation of their operating system monopoly into the network application market."
   Microsoft is characteristically unrepentant. "This guy makes me laugh," says Brad Silverberg, head of the personal operating systems division. In the Microsoft version of events, Windows 95 does not "disable" anything. It just happens that some companies' applications cease functioning -- they "use nonstandard components" and "need special configuration." Those companies violated Microsoft's published guidelines, he says; they have realized their error and are preparing new versions of the software to repair the problem.
   The truth is not quite so innocent. Most Internet dial-up software written for Windows relies on a piece of software called winsock. Everyone's winsock is supposed to be more or less interchangeable with everyone else's, but differences do exist. Many vendors put their winsock into the Windows directory of the user's computer—a friendly practice, since it is then available to other software that might need it, but a risky one, too. If Windows 95 sees a non-Microsoft winsock, it carefully and explicitly replaces it.
   "It's not like we blow it away and it's gone forever," Silverberg says, beaming with sincerity. "I think we do a very honest and responsible thing. It's admirable, really."
   He acknowledges that the specifications for using the operating system's new dialer were slow in coming but says they are now available to all who want them. And for that matter, he asserts, if Microsoft chose to keep such specifications private, to give a competitive advantage to its many software departments, that would be the company's privilege. It does own the operating system, after all.

More Windows, Bigger Windows

It is conventional at Microsoft to say that success comes from making good products. Microsoft does devote extraordinary resources to improving its technologies. It has effectively stressed "usability" and crisp design. It has recently created a 100-person research laboratory that resembles a leaner and harder-driving version of A.T.&T.'s Bell Laboratories and I.B.M.'s Thomas J. Watson Laboratory. But at least to date, the quality of its products has been incidental to Microsoft's triumphs over its competitors.
   Even Windows 95 shows more awkwardness and instability than the personal operating systems that have long been available from Apple, I.B.M. and Next. It adopts virtues of all those systems, but many users will still struggle with obscure techniques for allocating memory to their old DOS programs, or find that they regularly crash the entire system. "In many ways this is an edifice built of baling wire, chewing gum and prayer," wrote Stephen Manes in assessing Windows 95 for The New York Times.
   It is conventional in the industry to say that Microsoft cannot make great products. It has no spark of genius; it does not know how to innovate; it lets bugs live forever; it eradicates all traces of personality from its software. This view, too, misses the point. Microsoft knows that the technologically perfect product is rarely the same as the winning product. Time and again its strategy has been to enter a market fast with an inferior product to establish a foothold, create a standard and grab market share.
   Designing the ideal laboratory operating system and competing in the real world are two problems that have little to do with each other. Apple has had the benefit of a closed battlefield; it could design its software for a limited set of hardware that it controlled. That was a huge advantage for developers and, ultimately, a fatal disadvantage in the marketplace. I.B.M. created in OS/2 an operating system clearly superior to Windows 3.1 in most important respects; yet it failed to persuade the hundreds of crucial manufacturers of P.C. hardware and the thousands of independent software developers to fall in line with compatible products. Windows 95, despite its "32-bit" fanfare, contains so much vestigial 16-bit code that it makes Intel's new Pentium Pro processor look bad. But that ugly old code means that users who make the switch will not have to throw out their old software too quickly. Microsoft's genius has been in navigating—and controlling—the fantastically complex ecology of the computer business.
   Microsoft's launch of Windows 95 in August, kicking off a planned $150 million marketing blitz, will live in history as a pinnacle of public-relations showmanship in a public-relations-driven year. When thousands of onlookers and journalists gathered under the big top on the Microsoft campus or watched nearby on giant screens, the subliminal message was, We can buy anything: Jay Leno (emcee and vaudeville partner for Gates), The Times of London (an entire day's run of a once-great newspaper), the Empire State Building (colored lights usually reserved for national holidays). The press made fun, but it was taken in, too, giving weeks of extensive coverage to what amounted in essence to a product introduction—and an upgrade, at that.
   Three months later, Windows 95 boxes are stacked high on store shelves, and Microsoft refuses to re lease sales figures. Anecdotally, it is clear that millions of high-end users have bought the upgrade but that millions of corporate customers have chosen to delay the inevitable headache, particularly when most existing hardware lacks the speed and memory to run it well. It doesn't matter. In the long run virtually every desktop computer will run Windows 95 and its successors. New computers shipping now have Windows 95 preinstalled by default. Applications developers have either stopped developing for DOS and Windows 3.1 or soon will.
   Windows has long since stretched the definition of operating system past the breaking point. The original DOS was little more than a thin (and clumsy) layer of hooks that applications could use for reading and writing data to memory, screen and disks. Windows 95 not only provides a rich environment for controlling many programs at once; it also offers, built in, a word processor, communications software, a fax program, an assortment of games, screen savers, a telephone dialer, a paint program, back-up software and a host of other housekeeping utilities and, of course, Internet software. By historical standards, you get a remarkable bargain.
   Some companies used to live by selling such things. Every time Microsoft adds a new feature to the operating system, ripples flow through the software business. When it added a built-in backup program, it instantly destroyed what had been a modest, competitive market in backup utilities; the only customers left were those with highly specialized back up requirements. And when Microsoft asks to license your technology, you may not always find it easy to say no.
   One company that tried was Stac Electronics, which had developed software that used a compression technology to effectively expand the capacity of users' disks. Microsoft wanted to build Stac's technology into the operating system and negotiated in its usual scorched-earth style, demanding a worldwide license for a one-time flat payment and threatening to move ahead with or without Stac's license. Stac refused, Microsoft acted on its threat and unlike most small companies that brush up against Microsoft, Stac sued. A jury, finding that Microsoft had stolen Stac's property, awarded $120 million for patent infringement. Microsoft then swallowed its pride and acquired the technology by settling with Stac, buying a 15 percent stake in the company. Stac now exists as a happy Microsoft partner and the disk-compression business is no more. There are pilot fish that manage to swim with sharks, and there are fish that get swallowed.
   A new cycle is beginning: with Windows 95 out, new groups of software companies are struggling to rethink their place in the market. Fax software companies are one example; and if Microsoft has its way, Internet software companies may become an other. The Netscape Navigator leads the market now, but after all, Microsoft's Internet Explorer is almost as good, and it's free.
   So the operating system has become, from the consumer's point of view, a useful pack age of software. From a different point of view, however —the point of view of the essential underlying structure of modern computing—the operating system Microsoft owns has become something else altogether: a collection of standards.

Walk Softly, Carry a Big API

The age of mass production could not begin until the world agreed on standards for the dimensions of nuts and bolts. The tire and automobile industries coexist be cause there are standards for wheel sizes. Standards development acts as a catalyst in economic development; the Internet itself emerged when, from the grass roots, open and free standards were created to allow different types of computer networks to talk with one another. All these standards were set by Government or international organizations or by industry consortia. No one must pay a royalty or license fee to manufacture a Class 3 fax machine or a keyboard with keys arranged QWERTY-style.
   From the point of view of standards, no form of machinery rivals software for the complexity of its interlocking parts—the number of jigsaw-puzzle interfaces between one element and an other. In understanding the two-decade history of Microsoft's increasing control over the computer software industry, nothing matters more than its strategic management of these points of interconnection: the creation, marketing and then manipulation of standards.
   Let's say you are an expert at a small company in the infant field of speech recognition, creating technology to turn the spoken word into stored text. You probably got an invitation from Microsoft during the past year to attend a series of meetings. You and your competitors, under Microsoft's guidance, helped create a standard set of hooks into the operating system, a so-called "application program interface," or API. No single company in the field had the clout to produce an API that the others would agree on, so there was danger of conflicting standards. But Microsoft did have the clout.
   The result: Microsoft, in cooperation with virtually the entire speech-software industry, will release early next year a "Microsoft Speech Software Development Kit," containing "all the necessary tools." Problem solved. Incidentally, in the course of the meetings, Microsoft received and filed away an enormous body of intelligence on the speech-software state of the art and even the specific product plans of your company. That's a risk you had to take.
   "I think it's a good thing," says Bathsheba Malsheen, general manager of technology at Centigram Communications, one of the speech-software companies. "To integrate voice and speech into applications is a costly problem." These standards are open, in the sense that they are publicly available.
   But in the long run, who actually owns them? "I guess they really are the property of Microsoft," Malsheen says.
   Microsoft has a mail standard, called simply MAPI (mail application program interface). It has a new telephone standard, for letting software interact with telephone equipment: TAPI. It is belatedly but feverishly working on a proprietary on-line multimedia document-publishing standard code-named Blackbird. Microsoft abhors industry-wide standards-setting: its pattern, with increasing consistency, has been to refuse to cooperate with any standards procedures but its own.
   "At one time it may have been, hey, the gang's all here and let's have a consortium blah blah blah," says Ballmer derisively. "You can't have things that thrive and get moved forward aggressively if it takes a consortium."
   Money on the Internet will require standards. Visa International and Mastercard International managed to set aside their rivalry long enough this summer to announce that they were creating a joint standard for processing credit-card charges across the Internet. Every major player in electronic commerce needs such a standard; until money can flow across the public net work in securely encrypted form, on-line shopping malls and information services remain more experimental than real. Then, a few weeks ago, the alliance broke apart.
   Mastercard, along with Netscape and I.B.M., charged that the standard, created by Microsoft and published as an "open" set of specifications, was actually proprietary, designed to give Microsoft a powerful advantage, perhaps enabling it to take a slice of every transaction. Microsoft responds that the specifications are freely available; its own Windows implementation of those specifications, however, is proprietary and available for those who wish to pay for a license, possibly on a per-transaction basis. It has become a familiar scenario: Microsoft claims an architecture is public and open; its competitors say the crucial details are reserved to Microsoft alone.
   Microsoft is by no means the only company that seeks to exploit private standards. Netscape itself is playing a dangerous game with the standards that gave rise to the World Wide Web: creating proprietary "extensions" that work only with its own software and hoping that its market dominance will be enough to make them stick. The history of I.B.M.'s downfall in the P.C. industry is a history of failed attempts to impose standards by fiat. I.B.M. took its clout for granted. Microsoft gives top priority to its standards-setting; it "evangelizes" its standards, using every possible form of persuasion to bring the industry in line.
   Ultimately, only one kind of company can play the standards game risk-free: a company with a monopoly. The risk for everyone else is that the company that owns the standard can change it without warning, can give its own programmers special advantage and can freeze innovation elsewhere.
   "We've lost this notion of a public standard as good," says Alex Morrow, general manager of architecture and technology at Lotus. "Instead we have this new thing, a quasi-open private standard that's controlled by one company. That's where innovation is going to suffer."
   The ultimate standard—the ensemble of standards—is of course the operating system itself: the power spot in the digital ecology. The case against Microsoft, in the eyes of its rivals, comes down to one central issue: leverage, using the operating-system as a fulcrum to gain power in new markets.
   The market in big desktop applications is a much-disputed case in point. Not long ago, WordPerfect led the word processor market with a much-loved product and a toll-free customer support service (something Gates has never authorized at Microsoft); Lotus 1-2-3 dominated the spreadsheet market, and Borland International's Paradox led the P.C. database market. In 1991, Mike Maples, a senior Microsoft executive, described the company's goals in the aggressive style that its top executives used to favor: "If someone thinks we're not after Lotus and after WordPerfect and after Borland, they're confused. . . . My job is to get a fair share of the software applications market, and to me that's 100 percent."
  For all three companies, the fatal "sea change" was the transition from DOS to Windows, particularly Windows 3.0, the first widely popular version. Microsoft notes with considerable justice that its rivals made a strategic blunder in not releasing Windows versions of the software more quickly. Microsoft's applications group and its system group were able to "fly in formation," as Ballmer puts it (zooming his hands cheerfully through the air). Microsoft critics have said that flying in formation included sharing technical information that gave Microsoft's own programmers an advantage over outsiders trying to write fast and well-integrated Windows software, and there is some truth to that. But there is also no question that WordPerfect, Lotus and Borland were late by choice—in part because, caught up in the Catch-22 of the operating-system wars, they knew that their Windows versions would help Microsoft by cementing the establishment of Windows.
  The flow of inside information will remain a critical issue for the antitrust investigators. In the 1980's, Microsoft executives often spoke of a "Chinese wall" between the systems group, responsible for DOS and Windows, and the applications group, responsible for the programs that ran in those operating environments. Ballmer himself once said there was "a very clean separation" -- "It's like the separation of church and state." Competitors were dubious, knowing that all neurons at Microsoft led to Bill Gates; these days Microsoft executives take a different tack. They deny that the concept of a Chinese wall ever existed. They admit that their own developers sometimes get an edge in knowing how to take advantage of new Windows features before the knowledge spreads to competitors, but they insist that the knowledge does spread sooner or later—because it is in their interest to make sure that everyone writes for Windows—and they say that's as level as the playing field needs to be.
   The final blow to the applications market came with the emergence of "office suites"—packages of word processors, spreadsheets and data bases bundled together. Again, Microsoft saw the opportunity first and made sure that its package was more tightly integrated than its competitors' could be. It announced a new standard, called OLE (for "object linking and embedding"), that allowed, say, a word processor document to display and even work with a spreadsheet. Again competitors charged, and continue to charge, that Microsoft manipulates the OLE specifications to its advantage-changing them to suit its applications programs. Almost as an afterthought, Microsoft also added its not well regarded Powerpoint presentation-graphics software to the package, effectively cutting the price to zero and transforming that business over night. Though transforming may not be the perfect word. "Microsoft didn't transform the market, but strangled it," says Karl Wong, director and principal analyst at Dataquest, a research company.
   Today, Microsoft says it "leads" the market in office suites. Yes, indeed: its market share is estimated at 90 percent, closer to Mike Maples' target than he could have dreamed four years ago.

For Its Own Good

The essence of antitrust law is an American view that the public has an interest in preventing excessive concentration of economic power. In the 1960's, two companies appeared to have such power, in the two industries with the greatest grip on the future, computing and telecommunications. The investigations of those companies, through several Presidencies, formed an era in antitrust law that ended abruptly on a single day: Jan. 8, 1982. The Justice Department dropped its long-running case against a jubilant I.B.M. but announced at the same time that A.T.&T. had, with bitter reluctance, agreed to a historic break-up.
   Today, I.B.M. has lost sway over every business it participated in. It allowed the P.C. industry to emerge at its feet, and it turned itself from a paragon of financial reliability into a company that for several years was losing money at a frightening rate. It has become a stagnant noncompetitor, looking for ways —its only hope—to break itself up into smaller business units.
   At A.T.&T., meanwhile, it is now an article of faith that the court-imposed break-up was a brilliant turning point in the company's fortunes. It was the event that freed it from its own hamstrung indolence and enabled it to compete in new arenas. A.T.&T. is continuing what the Government began, breaking itself up into smaller and, it hopes, more agile companies.
   Monopolies become their own worst enemies—particularly in businesses that live or die by technological innovation. They get soft. They make poor research choices. They bleed both profit and invention. They poison the marketplace that created them. In the rarest cases, like A.T.&T.'s, an outside force can save a monopoly from itself, but Government interference is always frightening and never popular.
   It's certainly unpopular with many politicians—witness the "pinch me" statement, a comment by Senator Bob Dole that Microsoft rushed into its legal briefs and news releases: "Let us understand what is going on here. A company develops a new product, a product consumers want. But now the Government steps in and is in effect attempting to dictate the terms on which that product can be marketed and sold. Pinch me, but I thought we were still in America."
   Microsoft's lawyers encourage an ideological view of United States v. Microsoft, employing not just "free-market capitalism" arguments but also a quaint form of red-baiting, assailing would-be "commissars of software," and insisting: "Such thinking should have disappeared with the Berlin Wall. Fortunately for American consumers, we do not have a centrally planned economy."
   "It's like a throwback to the 1950's," says Case at America Online: `What's good for General Motors is good for America."'
   For her part, Anne K. Bingaman, assistant attorney general in charge of the Antitrust Division, bridles at suggestions that the political climate could affect the investigation—and also at a widespread industry view that, in the end, the high-technology business will prove too fast-moving and too technical for the non-nerd lawyers in Washington to keep up.
   "We have a much better handle on the industry than people realize," Bingaman says. "The group of people that work on these matters have long and deep experience. We keep up. We understand it. We have sources."
   Bingaman is proud of achieving the consent decree in phase one, in which Microsoft agreed to end a set of licensing practices without admitting any wrongdoing or suffering any penalty. The most blatant was an arrangement in which P.C. manufacturers paid Microsoft the same royalty for shipping a computer without DOS as with DOS—meaning that, if you were one of the few people who bought a non-Microsoft operating system, you paid its manufacturer and then you paid Microsoft on top of that, a huge disincentive. Microsoft was "locking up the market with practices which every computer manufacturer despised and which the competitors despised," Bingaman said in July 1994. "To get these low prices you had to sell your soul and never leave Microsoft." And she also said: "I hope consumers, within a short period of time, will have more choice of operating systems."
    It has not happened. The practices Microsoft agreed to forgo had already served their purpose. Gates was right when he summed up the effect of the consent decree in one word: "Nothing."
   And each month brings new issues, all variations on the same theme: Microsoft's use of not-quite-public standards as a sword and a goad. The Microsoft Network shipped with every copy of Windows 95 before the Government's lawyers could decide whether to act. Now they must consider the new Microsoft-Visa agreement on standards for financial-transaction processing—open standards, according to Microsoft; closed standards, according to Mastercard and Netscape -- and as of this month Visa has already shipped its Windows software implementing the standards. "We are not giving away our implementations of those specifications, just as we don't give away Windows or any other software product that we make," says Craig Mundie, senior vice president of Microsoft's consumer systems division. Microsoft is well along in the creation of proprietary software to handle every stage of the process, from customer to merchant to bank.
   Meanwhile, the stores are filling with third-party software boxes displaying the official Windows 95 logo. To get Microsoft's permission, the manufacturers had to demonstrate not only that their software runs under Windows 95, but also under the more advanced version of the operating system, Windows NT—a version that so far, despite all Microsoft's evangelizing, does not have the support of many popular applications. That logo is a powerful lever, applying power from one product line to another, and it deserves the Justice Department's attention.
   So does Microsoft's new campaign on behalf of not-yet-available on-line development tools, like the one code-named Blackbird, for companies that want to publish news, design games, build shopping malls or deliver entertainment over the Internet. Designers of competing tool sets—Netscape and Sun—see Microsoft's as attempts to gain control of another key choke point in the pathways of electronic commerce.
   So does an odd bit of language in Microsoft's contracts with the computer makers who bundle Windows 95 with their hardware: a forced promise not to sue Microsoft or anyone else for patent infringement. It happens that Microsoft is building up a strategic portfolio of software patents, both home-grown and licensed.
   And so, of course, does the intimate connection between Windows 95 and Microsoft-brand Internet access: the bundling of the Microsoft Network software; the persistence of the desktop buttons; paradoxically, all the features that make on-line access easiest for new customers. As Microsoft vehemently points out, every other big on-line service manages to gets its software into your mailbox and bundled with your new computer. Still, the Antitrust Division should, if nothing else, see a natural analogy with the bias A.T.&T. created for itself be fore the days of equal access. Customers could use M.C.I. —but only by dialing a slightly inconvenient code.
   Microsoft retorts angrily to all the telephone analogies by noting that A.T.&T. was a Government-regulated monopoly. The folks at One Microsoft Way are merely . . . successful. They are big. If they are linking together pieces of the hardware-software-network chain, they are doing it in a way that has lowered prices, added value and made life easier for consumers. It is not their fault if the economics of scale in the software business, combined with tactics that press antitrust law to its limits, brings them huge benefits.
   Is Windows an open standard? Yes —when and only when that suits Microsoft. "We could say, hey, we're not publishing any A.P.I.'s to our operating system," Ballmer says. "Or we could pick five guys and tell them what's in this operating system—we're not going to tell other people."
   And that is where the Government should draw its line.
   There was a moment in history, just a few years ago, when any number of operating systems, real and imagined, could have emerged to run the world's personal computers. That moment is past. The Microsoft architectures have established them selves so deeply in every segment of the computer business that they cannot be displaced, not even by Microsoft. Those standards are an essential facility—to use antitrust jargon—like the 60-hertz AC current that flows to every American household. To date they have remained mostly open and mostly public, because that served Microsoft's business interest. Now the Government could, and should, declare a public interest in open standards in computing.
   The Department of Justice does not need to break Microsoft apart. It need only—a far-reaching step in itself—require Microsoft to make its operating system, and the web of standards surrounding it, truly and permanently open. Other companies should be allowed to clone it if they could; Microsoft should be restricted from taking internal advantage of new changes until they were published to the rest of the market.
   For that matter, Microsoft should open its standards voluntarily. It will not, but it should: end the painful cognitive dissonance that comes from proselytizing for open standards and then threatening to close them at will.
   "It's not like everyone and their brother is going to go out there and beat them," says Eric Schmidt at Sun. "They'd probably have 95 percent of the market any way. Then all the arguments about their behavior would stop. If they really did open interfaces, it would change the dynamics of the industry in a positive way." It would be for their own good, he says: "They could get back to work and try to build great products and compete."
   Gates is right about one thing: Microsoft's future is no more assured than was I.B.M.'s. The Internet does pose a threat—a new set of open standards that, so far, Microsoft cannot control. And Microsoft's own power poses a threat, too—the threat that comes with the self-fulfilling destiny of any monopolist. Microsoft could fail to drive consumers to new waves of upgrades; it could stagnate financially even as it retains its grip on the neck of the market. "The company in some sense is a captive of its own history of voraciousness," as a former Microsoft executive says. It is a captive of shareholders who have come to expect nothing less than Microsoft-style profit margins and growth rates. It is a captive, to its own horror, of lowest-common-denominator design, the inevitable consequence of serving a market of 100 million.
   The rest of the industry is captive, too. No company has Microsoft's power to place bets; few companies can afford to chance a new approach in a product category near the ever-advancing boundary of Microsoft's Windows package. No quantity can be harder to perceive and harder to measure than innovation that never occurs—the absent pioneers, the fading of vitality in a still-comfortable industry.
   No monopolist wants to be relieved of its burden. To Microsoft, it would be nothing short of theft. They own that operating system—they sweated, invested and fought for it. If they can put a computer on every desk and in every home, all running Microsoft software—and all connecting to the Internet —consumers should be grateful.
   "You click a button and it's so easy!" Silverberg says, grinning again. "How could there be anything wrong with that?"


This article first appeared in the New York Times Magazine on November 5, 1995, just as Windows 95 was entering the marketplace. Many readers—maybe most readers—disagreed with it. But it seems that my explanations of how Microsoft gains and exercises power--controversial then--have gained mainstream acceptance.

The Antitrust Division began inquiries into Internet-disabling incident described in the opening section, among other issues; it has also asked questions about Microsoft's tactics in the battle against Netscape (a battle that, where browsers are concerned, I believe Microsoft has already won). As of 1998, it is trying to halt Microsoft's tying of the operating system to its Web browsing software. I took another look at one of the issues in the case, here.

The Internet continues to shape and reshape Microsoft's strategy. The company has reorganized itself, redirected its online-tools development, and rethought its marketing plans along the lines Gates had indicated. More than anything else, Microsoft is striving to control the standards for access to content on the Internet; I hope readers of this article will understand why.

Meanwhile, the full extent of Microsoft's triumph over Apple has become even clearer. Apple thought it would be enough to make good products that its customers liked, without necessarily trying to control everything. Microsoft knew better.

Reaction to this article continues to be voluminous and intense.

Many Microsoft supporters -- including, I take it, happy customers—responded with outrage. In some cases I had the impression that they were reacting more to a general sense that I was attacking Microsoft (as published in the Times, the article was accompanied by a possibly inflammatory cover picture of the proverbial 800-pound gorilla) than to the specific arguments. A sample of these letters was published in the Times; in terms of flavor, it was as if I'd stumbled into a Windows v. OS/2 flame war circa 1994.

Many others, ranging from industry insiders to academics to journalists with Microsoft experience, have responded appreciatively. Thank you for your comments—quite a few of which have given me new insights into the history and the dynamics I tried to describe.

Microsoft itself appeared upset by the article, which did constitute an unusual and critical look at its practices. Bill Gates wrote a long reply, also published in the Times, repeating company positions that he and others had expressed in the article. Microsoft has declined to permit his reply to be included here.


Copyright 1999 James Gleick