Jonathan Zittrain is executive director of the Berkman Center for Internet & Society at Harvard Law School. His essay, "The Un-Microsoft Un-Remedy," appears in the current issue of the Connecticut Law Review.

Editor's note: For more on the Microsoft decision, be sure to check out VoxCap.com's ongoing feature on "Microsoft: The Findings of Fact."

Judge Thomas Penfield Jackson has legitimized the Justice Department's claims regarding Microsoft: The company has bolstered a monopoly on personal computer operating systems through a variety of unsavory means. The findings of fact released earlier this month began a trajectory that will end, it seems, either in settlement or a formal finding of liability under the Sherman Antitrust Act. What should Joel Klein be asking for, and what should Judge Jackson impose if no settlement is reached?

It's hard to imagine a fine large enough to make Microsoft regret its behavior

Even with liability clear, a satisfying remedy is elusive. It is hard to imagine a fine large enough to truly make Microsoft regret its behavior, and the net result -- a continued monopoly on PC operating systems -- would remain unchanged. A series of ongoing restrictions - Microsoft critic Sun Microsystems has already drawn up a wish list including a barrier to Microsoft's entry into cable and wireless, and a requirement that Microsoft invent instead of buy technologies -- would be unwieldy and subject to constant policing. A breakup of the company is as unpredictable as it is ambitious.

Microsoft's unfair -- and well-exploited -- advantage

The idea of compulsory licensing of Windows -- that is, a requirement that Microsoft make the source code available to other software companies at a fair price -- may prove more fruitful. In fact, the concept points the way toward a longer-term solution that could readily prevent any one company from maintaining undue dominance in information technology.

To understand why, first appreciate that would-be competitors to the Windows operating system would find it difficult to gain traction in two ways. First, Microsoft offers a trusted brand and good products backed by a crackerjack set of programmers, designers and marketers. This is old-fashioned competition that should not concern antitrust regulators, who acknowledge that, monopoly or not, some cases of overwhelming market share can be earned and kept fair and square.

The other reason is not so benign, however. Microsoft has a special advantage that comes from the huge popularity of Windows: Many software authors write applications that require Windows instead of a more obscure operating system, and many consumers buy only computers with Windows because only Windows can run all the most popular applications. It is a feedback loop that applies whether or not Microsoft Windows is the "best" operating system in terms of features, quality or price. It is this special advantage that sustains a continuing monopoly not grounded in fair competition, and which gives Microsoft the opportunity for abuse -- its behavior is no longer constrained by the market. But this advantage is itself rooted in copyright, a legal protection that compulsory licensing would partially withdraw.

Not in this lifetime

Without copyright protection, Microsoft could still sell Windows -- people might not trust Windows from other sources any more than they trust generic acetaminophen over Tylenol. But others would be free to make their own variants as well, wiping out that part of the Microsoft monopoly that is based on the desperate need for continued compatibility -- and on a government-given monopoly called copyright that allows only one source for such compatibility -- rather than on features and quality. Indeed, new versions of Windows generated by other vendors would probably still run Windows applications; they would simply compete on other bases: speed, look and feel, added new features that do not break compatibility with earlier software.

In the American legal system, copyright protection is supposed to lapse naturally, precisely to avoid continued monopolization of one's work. The Constitution makes it clear that the copyright monopoly should be for a "limited time." After a term a copyrighted work is to enter the public domain, freely available fodder to seed a new round of innovation and creativity by new authors who draw upon the old to create something new. The intent was to make the copyright monopoly last only long enough to give the author incentive to create; after that, the work would belong to the public.

The current "limited term" of computer software copyright protection is 95 years. That does not make much sense.

Only at the dawn of 2075 will the operating-system code of a later-model TRS-80-a Radio Shack computer (already enshrined in the Smithsonian as a relic) become free. Absent a compulsory licensing remedy, it will not be until 2090 that software architects -- anyone from university students to Microsoft's more traditional competitors -- will be able to legally build upon the code to Windows 95.

Five years and out

The long-term solution? Intellectual-property protection for all computer software, including operating systems, should be drawn much more in line with its shelf life. Before the Smithsonian is preparing a display for a piece of software, the government-granted intellectual-property monopolies,such as copyright and patent, should lapse. That still leaves plenty of time for a software company to recoup its investment in new products -- it will have the exclusive right to sell them for, say, five years -- while preventing it from coasting forever from an early success and automatically parlaying it into future successes.

Once a popular piece of software enters the public domain, other software developers can use it to help build competing products. Imagine: Windows 95 would come free next year, allowing the creative energy of the non-Microsoft programming community to build new operating systems that still run every application Windows 95 supports. Consumers would not have to stick with Microsoft's version of the operating system; they could choose among a whole set of Windows-compatible systems. In the meantime, Microsoft can release and newly copyright Windows 2000, a child of Windows 95 that will have to step out of its parent's limelight and earn its own new base of developers and consumers.

The Microsoft case should serve as the wake-up call. It is time for a multilateral, mutually beneficial disarmament of government-granted intellectual-property subsidies within the computer software industry.