Programming Windows: Findings of Fact (Premium)

While most histories of Microsoft focus on two obvious areas of general interest—the company’s founding and early years, and its historic antitrust trials in the U.S. and Europe—I’m taking a different approach here. Instead, I am focusing on the history of a specific product, Windows, and on the forces that shaped it and, in turn, shaped Microsoft and the personal computing industry.

For this reason, we’ll examine antitrust only insofar as it impacted Windows. And we’ll start with an existential threat that dominated Microsoft’s attention in the late 1990s, Netscape and its Navigator web browser.

On November 5, 1999, U.S. District Judge Thomas Penfield Jackson issued his Findings of Fact in U.S. v. Microsoft. It is a 183-page, 60,000-word dressing down of Microsoft and its executives, especially co-founder and CEO Bill Gates, and the first major legal determination that the software had a monopoly which it illegally maintained and expanded at the expense of competitors, partners, and customers.

Key to Jackson’s Findings is that Microsoft perceived a very real threat from so-called middleware solutions that ran on top of Windows, and thus used and needed that platform’s application programming interfaces (APIs) while providing their own APIs for developers as well. Microsoft’s fear of middleware was extensive, and it included even minor platforms like an Intel’s Native Signal Processing (NSP), Apple’s QuickTime, RealNetwork’s RealPlayer, and Lotus Notes. But two middleware technologies emerged in the mid-1990s that represented existential threats to Windows: Netscape Navigator and Sun Microsystems’ Java.

From Microsoft’s perspective, Navigator and Java were particularly pernicious because they were cross-platform—they ran on Mac, Unix, and other platforms in addition to Windows—and could thus work to undermine the importance of Windows to PC makers, developers, and businesses and other customers. That Netscape and Sun explicitly collaborated to take on Microsoft’s stranglehold of the PC industry was likewise a clear sign of threat: Netscape announced that it would bundle Java with its Navigator web browser in May 1995, just three months before Windows 95 shipped to an eager public.

And as was proven in Jackson’s court, Microsoft engaged in illegal activity to harm both Netscape and Sun and to undermine their products and technologies.

It started immediately in the wake of the Netscape/Sun deal: Microsoft tried to convince Netscape to divide the browser market with it in June 1995, with Microsoft gaining total control of web browsers for Windows 95; in return, Microsoft promised to promote Netscape’s server products to make Netscape a “preferred ISV” that would gain access to new Windows APIs more quickly than did most developers. When Netscape refused to collude with Microsoft, the software giant withheld the APIs and other technical documentation that the firm needed to make a Windows 95 version of Navigator until that October.

“The delay in turn forced Netscape to postpone the release of its Windows 95 browser until substantially after the release of Windows 95 (and Internet Explorer) in August 1995,” Judge Jackson wrote. “As a result, Netscape was excluded from most of the holiday selling season.”

Microsoft also dramatically sped up its internal efforts to develop Internet Explorer, its own web browser. The IE team grew from five developers in 1995 to more than one thousand by 1999. Microsoft spent at least $100 million directly on IE development and at least $30 million on IE marketing every year from 1995 on.

Despite this investment, Microsoft gave IE away for free, where Netscape had been charging customers for its web browser. And it did so specifically to harm Netscape’s business, by leveraging Windows to distribute software that others were selling.

“Senior executives at Microsoft decided that Microsoft needed to give its browser away in furtherance of the larger strategic goal of accelerating Internet Explorer’s acquisition of browser usage share,” Jackson explained. “Decision-makers at Microsoft reminded those reporting to them that browser usage share remained, as Microsoft senior vice president Paul Maritz put it, ‘job #1.’ For example, in the summer of 1997, some mid-level employees began to urge that Microsoft charge a price for at least some of the components of Internet Explorer 4.0. This would have shifted some anticipatory demand to Windows 98 (which was due to be released somewhat later than Internet Explorer 4.0), since Windows 98 would include all of the browser at no extra charge. Senior executives at Microsoft rejected the proposal, because while the move might have increased demand for Windows 98 and generated substantial revenue, it would have done so at the unacceptable cost of retarding the dissemination of Internet Explorer 4.0. Maritz reminded those who had advocated the proposal that ‘getting browser share up to 50% (or more) is still the major goal’.”

In other words, not only was Microsoft giving away software for free to harm a competitor, it was willing to limit its own revenues to achieve dominance in the web browser market and to harm the current market leader. Additionally, Microsoft “paid huge sums of money” to induce its partners, Internet Service Providers, and others to help increase IE’s market share. Microsoft even bartered Windows desktop real estate so that online services like AOL, CompuServe, Prodigy, and others would push IE over Netscape, even though those services competed with Microsoft’s online service, which was bundled with Windows.

Worse, Microsoft actively worked to prevent Netscape from distributing Navigator as well. It threatened PC makers that wished to bundle Navigator with revocation of their Windows licenses, a business-ending action. It also prevented PC makers from removing or altering IE. And to those who argued that Navigator was freely available online, Jackson countered with something that seems odd today but was obvious to people of that day: Downloads were slow and expensive, and by keeping IE the only web browser option bundled with Windows, Microsoft could ensure that most of its users would simply use that rather than bother going to the trouble of getting another browser that just did the same thing.

Most incredibly, Microsoft worked to “bind the [Windows] shell to the Internet Explorer, so that running any other browser is a jolting experience,” in the words of Microsoft executive Brad Chase. The first version of Windows 95 in which “browsing-specific code shared files with code upon which non-browsing features of the operating system relied” appeared in August 1996.

With this release of Windows 95, called OEM Servicing Release (OSR) 2, “Microsoft placed many of the routines that are used by Internet Explorer, including browsing-specific routines, into the same files that support the 32-bit Windows APIs,” Jackson wrote of Microsoft commingling of Windows and IE. “Microsoft’s primary motivation for this action was to ensure that the deletion of any file containing browsing-specific routines would also delete vital operating system routines and thus cripple Windows 95.” Microsoft did provide a way to hide the IE application, but the commingled code remained in Windows.

But this level of integration wasn’t enough for some senior Microsoft executives. And so Windows 98 was delayed past the crucial 1997 holiday season, harming Microsoft’s PC maker partners so that IE could be integrated even more deeply into the OS. This decision baffled many on the Windows team, who were of course not involved in such top-level decisions. When one asked Paul Maritz if the plan to “hold Windows 98 for IE 4” was serious, Maritz told him it was.

“The major reason for this is … to combat Netscape,” he answered. “We have to position the browser as ‘going away’ and do deeper integration on Windows. The stronger way to communicate this is to have a ‘new release’ of Windows and make a big deal out of it … IE integration will be [the] most compelling feature of [Windows 98].”

In Windows 98, users wouldn’t even be able to hide IE, and even when users chose another browser as the default, the system would in some instances override that default and load IE anyway. This was the “jolting experience” that Chase had described earlier.

The legal issue with this bundling is that it was done solely to harm a competitor, and not because of any perceived value to customers. Indeed, the bundling harmed Microsoft’s partners and customers, in part because the delay of Windows 98 prevented PC makers from shipping PCs that could support new technologies and features and their users were prevented from purchasing them. The resulting experience was likewise confusing to customers and made computers work more slowly. And when users did install the browser they really wanted to use, the experience could be, well, jolting.

But as is the case so often, there is another side to this story.

While the commingling of Windows and IE code is somewhat indefensible, Microsoft’s bundling of the two products did indeed benefit consumers, especially since “the number of reviewers who regarded Internet Explorer 4.0 as the superior product was roughly equal to those who preferred Navigator.” Not having to find, download, and install a third-party browser was thus a boon to the consumers of 1998, especially since the browser that Microsoft included with Windows 98 was as good, if not better than the market-leading browser of the day.

As was the case with Java, too, Microsoft’s embrace of web technology was complete and went far beyond the level of work one might think was necessary to simply defeat an opponent. Indeed, the amount of web integration in Windows 98—artificially commingled or not—is breathtaking.

In Windows 98, Microsoft replaced major portions of the shell with code from its web browser, creating a more seamless experience in which the file system acted like the web, with optional one-click access to files and folders, and browser-like navigation controls such as Back and Forward which persist in Windows today. Many windows now displayed dynamic web views rather than the static views of the past. Windows 98 even brought system updating to the web with the advent of Windows Update, a “web-based extension to Windows 98.”

But the integration of Windows 98 and IE 4.0 perhaps takes its most dramatic turn right on the desktop, where Microsoft provided a dynamic, web-based Active Desktop, complete with a floating Channels bar by which consumers could receive push notifications from popular web services. Windows 98 even included a basic version of its FrontPage website editor so that consumers could create and edit webpages.

And that’s just the web integration: Windows 98 also shipped with dozens, if not hundreds, of meaningful improvements over Windows 95. To this reviewer’s eyes, seeing and using Windows 98 again for the first time in 20 years, I’m struck by how more refined it is than its predecessor. And despite the months-long delay to ensure IE 4.0 integration, Windows 98 arrived just 2.5 years after Windows 95, and Microsoft released several OSR updates for PC makers that provided many of the improvements that later shipped in Windows 98. The speed of the improvements it implemented to Windows during this time period is incredible, even in retrospect.

Today, of course, most of Microsoft’s web integration features in Windows remain only in vestigial form. The Active Desktop and Channel bar are long gone; even Internet Explorer has disappeared. And while IE’s successor, Edge, was originally bundled with Windows 10 similarly to IE in the past, that strategy, too, has finally faded: The next version of Microsoft Edge will ship in Windows, but it will no longer be a deeply integrated, and it will be available from the web, and for other platforms, as well.

This speaks to a central argument of U.S. v. Microsoft, where both sides sought to predict the future of Microsoft’s predatory behavior. If left unchecked, Microsoft’s foes argued, the software giant would continue to dominate the industry, harming innovation, raising prices, and hurting its own customers. This was just speculation, Microsoft argued, and it believed that Windows was constantly in danger of being supplanted by new platforms, perhaps PDAs, phones, set-top boxes, or simpler computing devices like Net PCs.

As it turns out, both were right, and Microsoft’s behavior needed to be curbed in order for new platforms to successfully emerge. And when they did, they managed to supplant Windows and the PC and establish a new age of personal computing in which Microsoft plays a role, but not the dominant role. That we are now dealing with antitrust issues related to these new platforms and their makers isn’t so much ironic as it is short-sighted: Microsoft has certainly learned the lessons of its antitrust travails, but the rest of the industry, apparently, has not.