Microsoft Windows was left off InformationWeek's list of the greatest software ever written. That's a debatable move. Windows meets the criteria in certain respects. It's hard to argue against something that's used everywhere--but does that mean it should have been included?

One of our standards for great software is that it can't be just programming that appears brilliant in a research or laboratory setting. It has to hold its own in the competitive marketplace, stand the test of time over many years, and be useful to many people. Windows meets those requirements and more when it comes to holding its own in the marketplace. Microsoft has been able to expand Windows capabilities at a rapid pace, making it difficult for competitors to catch up. It continues to do so, and Windows has had a huge impact as the practical operating system for the non-guru PC user.

But Windows owes its dominance to something other than programming breakthroughs or technical innovations. Other operating systems, including MS DOS, Unix, Digital Equipment's VMS, and Apple Macintosh, were the first to introduce many new features. Microsoft popularized them to a wider audience.

No, Windows owes its dominance in large part to a series of agreements Microsoft negotiated at the start of the PC revolution with computer manufacturers. They called for the placement of Windows exclusively on new machines as they were shipped. Those agreements weren't made public in the embryonic PC market of the 1980s, but disclosed in later antitrust proceedings. In return for default placement of Windows, Microsoft discounted its charge per copy to the PC manufacturer. To not play along with this discount practice meant loss of competitive advantage in a market with razor-thin margins.

In the 1990s, Windows even had a feature that checked for other operating systems on the hard drive and reacted if it detected one, prompting the user to disable it before Windows would boot. It could be argued that this was a necessary feature in an operating system that can't risk colliding with a second system using the same disk space. But a cottage industry grew up turning out products, such as V Communications' System Commander, that could get around this restriction and allow a second operating system such as Linux to coexist.

Great software doesn't rely on such maneuvers to win in the marketplace, and consequently great software isn't always guaranteed success. At the same time, big successes may have as much to do with shrewd business distribution as programming greatness. Windows has passed along too many problems to end users and ducked too many hard issues, such as meeting system security requirements, to qualify as great software. For years, Windows' Internet Explorer browser was defenseless against all kinds of riff raff on the Internet.

To be great, it's supposed to do the job intended. Windows comes close, tries hard, gets an A for effort, but just when we think it might be great, along comes that blue screen of death and we're rebooting all over again.