A Microsoft Decade: Writing And Rewriting History

By Stuart J. Johnston

This Monday, February 16, I marked an anniversary: exactly 10 years of covering Microsoft. Unfortunately, some of the history of that period has been mythologized by conspiracy theorists. Not least is the story of how Microsoft captured the Windows applications market. There is what I call the "X-Files version" of history and then there's what occurred, as I reported it.

The "X-Files" version of events goes something like this: Microsoft used its operating systems monopoly to illegally leverage the applications market. Microsoft's applications programmers used undocumented programming interfaces to create applications that ran "better" than competitors' applications. Microsoft also tricked competitors into developing for OS/2 instead of Windows while its own developers secretly wrote for Windows.

Here's a more historically accurate version:
Microsoft shipped Windows 1.0 in late 1985 and immediately began working on applications for it. Several more Windows versions shipped over the next four years, but all were utter failures. Additionally, despite Microsoft's dominance of desktop operating systems with MS-DOS, it never became dominant in any major DOS applications category.

However, Microsoft's applications did become dominant on one platform--the Apple Macintosh, where Microsoft did not own the operating system. Although most major DOS applications ven dors had promised to ship Mac products, only Microsoft actually delivered. Lotus got as far as beta-testing a spreadsheet named Jazz, but then killed it.

It was on the Mac that Microsoft introduced a desktop productivity applications suite--Office--and originated the concept of selling several applications in a bundle for less than the cost of purchasing them separately.

In 1983, Microsoft signed an agreement with IBM to jointly develop OS/2, which IBM planned as DOS's replacement. Over the next seven years, Microsoft spent at least $100 million on OS/2 development and double that amount on marketing the ill-starred system. This was during a period when Microsoft's sales grew from $50 million per year to $1.2 billion. It is ludicrous to suggest that any company--even the "evil Satan" Microsoft--would spend roughly 10 cents out of every dollar on a feint.

In mid-1989, Microsoft began beta-testing Windows 3.0. The immediate and deafening buzz from smaller applications vendors like Micrografx, Co rel, and Ami, and from the user community was that this system would finally answer their needs. Unlike OS/2, it could run more than one DOS application at a time; it ran on top of DOS, which system administrators already understood; and it provided much of the ease of use of the Macintosh. Windows also required fewer hardware resources than OS/2 and it cost less. It quickly became obvious that Windows 3.0 was going to be a blockbuster.

But when I asked major developers such as Lotus why they weren't writing Windows 3.0 applications, the condescending response was that IBM was the "kingmaker." Since IBM had already resolved that we would all soon be moving to OS/2, Windows development was "unnecessary." Windows was just a "transitional" environment.

Despite the fact that Microsoft provided beta copies of Windows 3.0 to all the major vendors by mid-1989, none deigned to write applications. It took more than a year after Windows 3.0 shipped before Lotus and others finally released Windows versions. But even then, they didn't really take advantage of the platform--and most magazine reviews reflected that.

In effect, by ignoring a rapidly developing market, the dominant applications vendors--who had apparently assumed that if they didn't release versions of their applications on Windows the operating system would fail--effectively ceded the market to Microsoft. After all, platform choices always revolve around a "killer application," don't they?

The blunders continued. IBM belatedly created a Windows applications division, only to kill it just before the first application was set to ship, preferring not to further wound sluggish OS/2 sales. Borland paid handsomely for Ashton-Tate's dominant DOS database, dBase--then took several years to ship a Windows version.

What about Microsoft's alleged "bait and switch" tactic of telling developers to write for OS/2 while secretly developing for Windows? Oops. Microsoft was the first vendor to ship OS/2 applications--a spreadsheet and a word processor. Bo th were terrible flops.

Additionally, on my first day on the job, Feb. 16, 1988, I covered a joint IBM/Microsoft OS/2 developers' conference in San Francisco where my first interview was with Steve Ballmer, now Microsoft's second-in-command and third-largest shareholder. That day, Ballmer said repeatedly--characteristically hammering his fist into his open palm---that Microsoft strongly suggested developers write to Windows first, and then port their code to OS/2. In fact, the first Microsoft OS/2 development languages, delivered that day, included a copy of the Windows software developers kit.

So what about Microsoft's applications developers using undocumented Windows calls, a story I first reported in September 1992? That story too has been conveniently revised to fit the conspiracy theorists' version of events.

All that my stories proved was that, contrary to what senior Microsoft executives had said, some Microsoft developers were using undocumented calls in their code. But all of the ot her major applications vendors were using those same undocumented calls. Though they were indeed undocumented, the calls were well known. No unfair advantage was gained since everyone was using them.

To my mind, Microsoft became the dominant applications vendor primarily because it took its own advice while its competitors took a blindered, contrarian approach that cost them market share. Competitors ignored an obviously shifting market, and whistled away their dominance through complacency.

If there is any valid argument that Microsoft leveraged its operating systems dominance to take over the applications market, it is that it was able to use DOS sales as the cash cow to pay for applications development and acquisition.

Now if you were to ask whether I thought that makes Microsoft invulnerable to attacks by the U.S. Department of Justice over the question of integrating its Internet Explorer Web browser into Windows 98, I'd say probably not. Indeed, the results of two recent InformationWeek s urveys, conducted three months apart, indicate that only a small minority of IT managers--about 25%--want Microsoft to merge the browser into the system. A larger percentage--around 37%--don't want Microsoft to merge the two. That weakens arguments that users demand that integration of the operating system and browser.

So draw your own conclusions about whether the Justice Department should try to stop Microsoft from shipping Windows 98 as is, or even take stronger actions. But nix the historical revisionism and conspiracy theories. That's not the way it happened.

AuthorITies Archive

Send Us Your Feedback

Rich Levin: Run Time Rich fills you in on all of the latest products, issues, and trends in application development.		Stuart J. Johnston: Redmond Watch As our eyes and ears in Redmond, Stuart gives his perspective on the latest events at Microsoft.		Charles Pelton: Eye On IT Charles explores IT management issues and strategies that business and technology managers face.

Top of the Page