U.S. District Judge Thomas Penfield Jackson's "findings of fact" in the Microsoft antitrust case do not constitute the dispassionate analysis one would expect of a federal judge. It is one-sided and betrays the judge's apparent anger toward Microsoft. The document is far less constructive than it could have been had cooler heads prevailed, and it might make finding acceptable remedies in this case more difficult because it leaves Microsoft little room for compromise or negotiation. It also complicates the search for solutions to a larger problem--defining "antitrust" in today's high-technology markets.

I understand why the judge might feel anger toward Microsoft. The company's lawyers and executives, beginning with Chairman Bill Gates, seem to have openly flouted antitrust law while arrogantly dismissing the government's charges as silly. But being angry with Microsoft is like being angry with a child who does not fully understand right and wrong. Microsoft was born less than a quarter-century ago in a cutthroat competitive culture and its management has not yet grown up. The judge's irate opinion won't help Gates and company learn why their practices are monopolistic, or what to do about it.

With 90 percent of the PC operating systems business, Microsoft has a monopoly, which is usually defined as having 70 percent or more of a market. It is not illegal to have a monopoly. Abusing that monopoly is what's illegal, but just what constitutes abuse is ambiguous. It could mean tying a product in one market to a product in a different market, suppressing competition unfairly or overcharging customers. In the case of Microsoft, Jackson's findings of fact make clear that the company bundled Internet Explorer with Windows in order to move into the Web browser market and thwart the efforts of Netscape--its prime competitor. But defining product boundaries and customer benefit or harm in the world of Internet software is not so simple.

First, full disclosure. I have been a Microsoft stockholder for nearly two years, and I want the company to prosper. I also have a pretty good understanding of the company and its people. In 1995, I co-authored the book "Microsoft Secrets" with Richard Selby. We concluded that Microsoft became the most powerful software company in the world because it was unmatched in its ability to set industry standards, roll out a continuous stream of low-priced "good enough" products and then support these with superb marketing and sales efforts.

Bill Gates and Paul Allen started Microsoft on the assumption that PCs would become ubiquitous and would need software to operate. They were right, of course, and deserve to be rewarded for their foresight. Selby and I chronicled this success story in "Microsoft Secrets." But we also wrote that the company, even before the browser wars began in 1995, was right up against the limits of antitrust law by coming dangerously close to tying sales of its application software to its Windows operating system and promoting exclusionary contracts that discouraged alternatives to Windows.

Microsoft's underlying problem is that it has not learned the rules of the game. Microsoft managers seem to believe that a huge market share in software is unlike a monopoly in oil or steel, where assets and market shares are more durable. Some evidence supports that view. For one thing, the monopoly itself is far more tenuous. In 1995, Netscape appeared almost overnight to challenge Microsoft Windows as the dominant computing platform and gained some 80 percent of the browser market. Then Netscape's browser "monopoly" collapsed, as Microsoft counterattacked with Internet Explorer. Sun Microsystems, IBM, America Online, Linux vendors and Netscape complain bitterly about Microsoft's threat to dominate the Internet, but nothing stops them from distributing from their Web sites millions of copies of products that compete with Microsoft's offerings and could someday supplant them.

The judge criticized Microsoft for tying Internet Explorer to Windows instead of treating them as separate products, and claimed that Microsoft harmed consumers because its "free" browser made Windows unstable. But he failed to acknowledge that, while Netscape Navigator was not free to most users until early last year, cost didn't necessarily translate into higher user benefit. Navigator, too, caused Windows to crash frequently--maybe more frequently than the Microsoft browser. Strikingly, the judge also did not seem to care that Microsoft--as well as companies such as IBM, Apple and Sun Microsystems--have been adding networking and communications features to their operating systems since the 1980s. Appending a browser to Windows was not only convenient for users, it was a logical extension of the technology, regardless of whether the browser is now considered a separate product.

These arguments are not well represented in the findings of fact. Nor is there adequate appreciation for what amounts to a new reality: that some kinds of high-tech products gravitate toward a natural monopoly. To be useful, computer hardware and software, and even products like VCRs, require complementary products, such as applications programs or prerecorded tapes designed to the same technical standards. So these markets tend to adopt one standard, often promoted by one company, because consumers demand compatibility, even if it comes at the expense of innovation. It becomes easy for firms such as Microsoft or Intel to overstep antitrust laws when product boundaries blur, dominant standards emerge naturally, antitrust laws are increasingly subject to interpretation and competitors act aggressively to protect especially fragile market positions.

David Yoffie and I studied the complexities of competition in the Internet age in our 1998 book, "Competing on Internet Time: Lessons from Netscape and Its Battle with Microsoft." Among other things, we argued that Microsoft was stepping over the antitrust line with actions such as threatening Compaq with the loss of its Windows license if Compaq featured Netscape's browser. I continue to question the legitimacy of Microsoft's actions in the browser market, and I agree with many details in the findings of fact in this area.

Yoffie and I also noted that the Internet and non-PC computing devices, such as PalmPilots, pose a long-term threat to Microsoft's market dominance. That threat grows larger every day. As a result, since 1995, Gates and company have become increasingly aggressive, paranoid and willing to flout antitrust laws.

I have also experienced Microsoft's arrogance and power firsthand.

In the fall of 1998, Yoffie and I battled Microsoft in court--twice. A team of Microsoft lawyers demanded that we turn over the research notes and taped interviews we compiled in order to write our book. We refused. Microsoft had us subpoenaed. In the antitrust trial, Microsoft lawyers used our manuscript (which it had obtained through a subpoena of Netscape) to argue that Netscape lost the browser wars because of its own blunders. We had documented those blunders, but we refused to turn over information that Netscape had not cleared for publication. Microsoft lost the subpoena battle, appealed and lost again. We have no hard feelings toward Microsoft because we understood the lawyers' motive: They were frantically looking for any evidence that might be useful to their antitrust case.

The problem Jackson faces today is how to get over his hard feelings toward Microsoft as he brings the case to a final resolution. His overwhelmingly negative categorization of Microsoft in the findings of fact document could result in ill-considered, drastic remedies such as forcing the creation of incompatible versions of Windows. Such measures would harm consumers as well as the entire U.S. software industry, especially if the result is years of paralyzing lawsuits as private parties sue to obtain retribution from Microsoft for past wrongs as an illegal monopolist. Every high-tech company with a dominant market share could become subject to frivolous litigation.

Seemingly, the only way out of this bleak scenario is for Microsoft to settle before the court hands down a final ruling. With a settlement, Microsoft would avoid being branded as an illegal monopoly as a matter of law, which could help it in any future cases brought against the company. Judge Jackson himself is encouraging Microsoft to settle and has appointed Richard Posner, chief judge of the 7th U.S. Circuit Court of Appeals, as mediator. Unfortunately, the findings of fact give Microsoft and Judge Posner little room to negotiate.

Whatever settlement or remedies emerge, Jackson should seize this opportunity to clarify antitrust law for murky high-tech markets. He should also insist that Microsoft put in place an education program in antitrust compliance for all its senior managers. Intel is a great example of this kind of effective self-education and self-regulation. It dominates the microprocessor market nearly as completely as Microsoft does its market, but Intel has stayed out of trouble. The company has been able to settle grievances out of court. More importantly, for more than a decade, Intel executives have insisted on having antitrust compliance training for their managers.

Companies that dominate new high-tech markets might behave differently and avoid being branded illegal monopolies if they better understood the rules of the game. The challenge now for Judge Jackson is to start creating a set of effective antitrust rules that Gates and company can live with and respect--not remedies simply designed to punish a misbehaving child.

Michael Cusumano is a professor at the MIT Sloan School of Management.