The Mosaic Code

Microsoft consolidated its dominance of the PC desktop before the days of the Internet. But even as Microsoft worked feverishly to integrate Windows into DOS, a team of students at the University of Illinois busily developed a graphical Web browser, Mosaic.

We saw the influence that the Netscape IPO had on the Internet’s emergence from a sleepy research project to a global information infra­structure and how all those initial hopes were dashed somewhere between Bill Gates’s Pearl Harbor Day speech in 1995 and AOL’s acqui­sition of Netscape in 1998 (chapter 4). The browser wars tell the story of what happened in between, of how Sun Microsystems was dragged into the mess, and of how, above all, Microsoft stayed true to its nature and squelched this most dangerous of all middleware threats.

All middleware threatens platform owners because middleware is inherently ephemeral. The natural process of software evolution will eventually migrate selected middleware into the platform to define a next-generation translation frontier. But middleware has yet another feature that scares platform owners. Middleware can translate platform APIs into a foreign language. What’s more, if the middleware developer decides to learn several different platform APIs, she can turn her mid­dleware into a universal translator. If an ISV learns the middleware’s APIs, and the middleware knows how to speak to multiple platforms, the middleware’s potential market will necessarily be larger than those of any of the platforms. Rational application developers will learn the middleware’s APIs—and the middleware will define the largest, most valuable network.

Netscape’s Navigator was a Web browser that ran on top of any plat­form—Windows, Mac, Unix, or other.

Microsoft met the challenge with some intense strategic thinking. The first stage was tried-and-true. Microsoft threatened OEMs who offered consumers access to the new exciting middleware. But OEMs were only one of Navigator’s distribution channels. So Microsoft expanded its strong-arm tactics to ISPs, Web developers, and virtually anyone else in the software or computing industries. Strong-arming alone helped Microsoft foreclose most of Navigator’s obvious and attractive distribu­tion channels. None of this was possible in Manifestoland.

Microsoft knew that stage one was only temporary. After all, if con­sumers really wanted the new middleware, they’d find it somewhere; Microsoft couldn’t keep Navigator off the market entirely. What more could Microsoft do with the wondrous IP rights granted by our Con­gress but denied to the citizens of Manifestoland? Microsoft could intro­duce incompatibilities to prevent the middleware from running on its platform, but that easily could backfire. Anything that confuses middle­ware developers is also likely to confuse ISVs—precisely the souls and minds that Microsoft was fighting to keep. Furthermore, consumers may chase truly exciting middleware all the way to a competing platform; exciting new technology is precisely the sort of disruption that induces consumers to eat their switching costs. If consumers all migrate to com­peting platforms running the middleware, Microsoft would risk not only losing its monopoly, but also the bulk of its customers. That’s how a market leader slides into oblivion.

Microsoft would have to do something bolder with the short time it gained by blocking distribution channels. Microsoft needed to develop a competing product.

But what if, somehow, Navigator was better than Microsoft’s com­peting Internet Explorer? What if it were so much better that people were willing to pay a premium for it—a large enough premium for Netscape to stay in business. Well, then Microsoft would have to override the gospel of software engineering and forego modular, evolutionary design. Microsoft would have to elevate marketing strategy over the imperatives of sound product design. In violation of all the basic tenets of software development, Microsoft could launch this complex new function as an integrated part of its platform.

Any consumer who purchased Windows—which would be almost everybody—would already possess Navigator’s basic functionality. Few would be willing to search out the few odd distribution points to get even a superior realization of these same functions, sold at a price capable of sustaining a small, standalone company with high product­development costs. Internet Explorer didn’t even have to be good when Microsoft launched it. It only had to be passable. At trial, the govern­ment established that Microsoft employed all of these leveraging strate­gies to destroy Navigator and Java, two of the most exciting and innovative software products of the 1990s.

Meanwhile, back in Manifestoland, poor Musoft couldn’t even foreclose a distribution channel.

The story of Musoft’s development in Manifestoland demonstrates how Microsoft, doing nothing but being true to its nature and using the rights that congressional IP policy gave it, maintained its monopoly posi­tion at the expense of consumer welfare. It shows that the wrong IP rights in the wrong hands can truly be dangerous. And yet, neither the DoJ, Judge Jackson, nor anyone else spent much time talking about Microsoft’s IP arsenal. Microsoft had devised a tactically brilliant strat­egy designed to maximize its profits. And it was able to enact this strategy in large part because of the IP rights that Congress had authorized—allegedly to motivate innovation.

These rights have led to an unfortunate public strategy. We’ve effec­tively decided to offer software developers a powerful combination of patent, copyright, and trade-secret protection, to instruct those develop­ers not to misuse those rights by violating the antitrust laws, to watch them behave like rational corporations, and then to conduct massive, expensive, politically charged, time-consuming trials. Is such a strategy necessary? Or might a Musoft, given only the weaker rights of Mani­festoland, have equaled—or even surpassed—Microsoft’s levels of inno­vation? If so, the Manifesto’s IP bargain would probably have given us a software industry—and a broad information sector—even more vibrant than the one that we have today, and we the consumers could all be enjoying even greater levels of entertainment and productivity at a fraction of the cost.