WITH THE MICROSOFT P.R. machine spinning madly, those of us who have been involved in the public policy issues raised by the Justice Department's six-year ordeal against the software giant can only sit back and smile in astonishment. The current Redmond P.R. mantra, the newest in a long line of pseudo defenses from Gates & Co., is that breaking up their firm is a "radical" remedy for the U.S. v. Microsoft Corp. antitrust case.

Wake up and smell the Starbucks, Seattle! Nestled in your little Northwest world, you suffer from an "Inside the Sound" myopia that makes Washington, D.C.'s Beltway mentality seem downright accomodating. Microsoft is going down. It's going down because it broke the law, blatantly. And a breakup is not just the best way to deal with the situation, it's the only way.

With its lobbyists swarming on Capitol Hill (the D.C. version), Microsoft says the Justice Department's April 28 proposal to cleave the company into two parts (separating the Windows Operating System from Office and other applications software) is overkill, stifles software innovation, and goes "beyond the scope of the case." What Microsoft forgets is that, for nearly 100 years, antitrust law has held that in monopolization cases like this one, a simple injunction ("Thou shalt not do that again") is too little, too late. When competition has been stifled, telling the defendant to walk straight and act right is not enough.

There are three reasons why a breakup, known as a "structural" remedy in antitrust parlance, is a far better solution than the other option -- a "conduct" remedy -- which relies on constant government monitoring of behavior.

First, the purpose of antitrust remedies is to "pry open" the market to competition and to prevent suppression of competition in the future. A conduct remedy only constrains the monopolist's ability to act on its (quite rational) business incentives in a marketplace that's already rigged to its advantage. (Windows' domination of desktop operating systems and Office's control of applications creates a marketplace where, if Microsoft were a Hollywood movie studio, it would also control 85 percent of the nation's movie theaters.) A structural solution goes right to the heart of this problem by dismantling the playing field and thus changing those incentives from the get-go.

Second, conduct remedies do not work because there's an imbalance in resources between the government enforcers and huge corporate defendants. No matter how complex and detailed the rules become, they ultimately depend on the defendant's own good faith for enforceability. Microsoft has already shown itself to be creative at best, and recidivist at worst, in finding ways around the most finely crafted government-imposed conduct remedies. Specifically, Microsoft took the position that its 1995 consent decree with the Justice Department did not apply to Windows 98, thus setting the stage for the second government antitrust lawsuit against the company.

Third, a conduct remedy inevitably leads to a perverse kind of "regulation by decree." The remedy becomes a hostage to the enforcement process. Judge Jackson (like Judge Greene before him in the AT&T case) would become a sort of software czar, sitting in judgment of Microsoft's every business decision. This sort of intrusive, long-term judicial regulation of the marketplace is inconsistent with traditional American notions of competition, particularly in a rapidly changing industry such as high technology. It would be the Vietnam of software regulation.

So, to answer Microsoft's latest line, a structural remedy is not only not radical, it is in fact the most conservative approach to fixing the problem. Instead of subjecting Microsoft to a decade or more of detailed judicial regulation of its conduct, a breakup allows the courts to act once and walk away. By limiting the need for judicial involvement in defining, limiting, and constraining behavior, a divestiture (breakup) changes incentives so that the behavior goes away. If constructed properly, structural remedies are the surgical strikes of antitrust law.

The Justice plan to divide Windows from Office would permit either new company to innovate in any area, and do anything except re-acquire the other. It would eliminate the need for the government to define the boundary between operating systems and applications by permitting integration without constraint.

Furthermore, trustbusting is not beyond the scope of the case. Microsoft likes to think that the antitrust case was only about its conduct against Netscape. That's sophistry. Netscape and the browser market were illustrative of potential competition for operating systems, thus threatening the Windows OS monopoly. When Marc Andreesen announced in 1996 that he would make Netscape into a "platform" competitor to Windows, he aroused the sleeping giant. Netscape, Java, NSP, and other "middleware" (software that can work as both an application and an operating system) presented the potential to evolve into direct competitors to Windows. So, by its predatory, exclusionary conduct, Microsoft blunted that threat, pushing back the prospects for OS competition by a decade or more. The case was never about browsers, it was about protecting the Windows OS monopoly.

The degree of hypocrisy coming from the Microsoft spin machine is incredible. In 1998, their defense was that AOL's acquisition of Netscape made the case moot, because Netscape would suddenly be revived competitively. It wasn't. In 1999, their defense was that Linux and the Palm V made the case irrelevant, because alternatives to the desktop PC were about to swamp the Windows OS. They haven't.

But that pales behind the direct contradictions that permeate Microsoft's case strategy. While his lawyers were telling Judge Jackson that "Internet appliances" constrained Microsoft's power over the PC desktop, Bill Gates told Newsweek that none of these devices would ever replace the PC. While the lobbyists were arguing that Linux presents a competitive alternative to Windows, Microsoft CEO Steve Ballmer was telling the press that Linux is confined to the server market and is only a blip on the screen.

In the final analysis, there are only three possible ways to deal with Microsoft: do nothing, regulate them, or break them up. The first is foreclosed by the trial judgment. The second is unconscionable. That leaves divestiture. Breaking up is hard to do only politically. In contrast to fashioning and enforcing a conduct decree, a divestiture is a piece of cake. And while the politicians bluster and prance, they forget that 20 years ago, they said the same thing about AT&T. Breaking up Ma Bell, though loathed by consumers and belittled as government tinkering with the "best phone system in the world," unleashed a torrent of innovation and price decreases in the telecommunications marketplace. We would not have the Internet of today if the Bell System still ruled with an iron hand.

So, in another 20 years, we will all look back on Windows as Bill Gates' version of the Princess Phone: the innovation that wasn't. Microsoft can kick and scream and complain, but the only legitimate solution to its monopoly power is to change the structure of the marketplace, and thus to unleash the competitive forces that Redmond now has the incentive and ability to subjugate. Breaking up may be hard to do, but it's the only thing to do.

(Glenn Manishin is an antitrust attorney at Patton Boggs LLP in Washington, D.C. He has represented the Software & Information Industry Association and other critics of Microsoft, and served with the Antitrust Division of the U.S. Justice Department from 1982-85.)