Jesse Berst: The hidden failure of Windows 2000

He who attempts to be strong everywhere... ends up strong nowhere.

That old adage refers to military campaigns... but it could have been written about operating systems. Take Windows 2000, for example.

Due this autumn, Windows 2000 is the update to Windows NT 4.0, Microsoft's "serious" operating system for business workstations and servers. Despite the new name, Windows 2000 carries over an old NT problem: It is too much for consumers. And too little for mission-critical business environments. By trying to be strong in both places, Windows 2000 is strong in neither.

At one point, Microsoft planned to make Win2000 the underlying platform for consumer and low-end business PCs, thereby abandoning the Win98 code base. Unhappily, Win2000 remains too big, too clunky, too complex and too incompatible with older apps to play in this space.

After testing Windows 98 SE and Windows 2000, ZDNet's sister publication PC Magazine concludes that Windows 2000 is more reliable and secure. But Windows 98 SE is more compatible with older apps and entertainment-oriented hardware..

Ironically, Win2000 is also deficient on the high end. After years of rapid growth, Dataquest research indicates NT's market share is in a holding pattern. As Dataquest analyst Kimball Brown puts it, Windows 2000 will lift NT from a "reasonably good low-end server environment" only to the "midrange." Those environments that demand high reliability and high availability -- financial applications, big Web sites and so on -- must still turn to Sun, Linux and other Unix variants, or even to mainframes.

Win98? WinNT? Win2000? Customers have been in this same dilemma for five years. Microsoft has spent many years and many millions of dollars... and left us just as stuck as we were before. That's why I hereby declare Win2000 a failure. Does it have lots of great new features? Absolutely. Is it an improvement over the previous version? Of course. Should you move to Win2000 if you already own NT? Yes, without question.

But does it achieve Microsoft's original goals? Is it best of breed in any one area? Is it the dominant operating system of the future? To those questions, I say no.

And you know what? Microsoft secretly agrees with me! Recently the company has:

  • Scrapped plans to abandon the Win98 code base. In fact, it just released another Win98 retread.

  • Licensed a "modular" version of NT for set-top boxes.

  • Announced a plan to create a new handheld version of NT.

  • Pulled some of the file features out of NT and gave them to the Exchange team to produce instead

Quite simply, Microsoft doesn't know itself which operating system is right for which market. In desperation, it is allowing product groups to compete with each other in a process of Darwinian selection.

But anyone who studies Darwin knows that a species is most successful when it adapts to a particular environment. Windows 2000 failed because it tried to succeed in every market. As a result, it's a jack of all trades but a master of none.