An endless vista

After yesterday’s bad news about Microsoft’s next operating system (Vista testers to Microsoft: Even the bugs aren't stable yet), I’m beginning to think the unthinkable: perhaps Vista will never ship.

Oh sure, something called Vista will ship next year; marketing and licensing imperatives demand it. And sure, it will be but a pale shadow of the vision when the project started; no product ever makes it out the door will all the intended features. I’m beginning to doubt, however, that the platform that ships will be what Microsoft needs it to be. It could become the Big Dig of Redmond. Even though it is eventually finished at vast over-runs, major flaws will continue to appear throughout its life.

There was relief at senior levels in Microsoft when Windows XP shipped. It wasn’t clear, even then, whether a substantial upgrade to such a complex product could be accomplished. Many years have passed since then, and many demands have been added to the wish-list. The code base is a hairball, the result as much of business decisions to lock in customers and preclude anti-trust action as of engineering philosophy. It is so large, interconnected, and poorly documented that any upgrade is a mind-boggling feat.

The important question, though, is not whether and how Vista ships; it’s what happens next. My impression is that the Windows NT code base, as built, is too complex to be the basis for substantial growth. However, there is no alternative in the wings. The NT code base, on which XP and Vista are built, has run out of steam sooner than planned. There is no code which is to Windows NT, as NT was to Windows 95.

The catch is that Microsoft doesn’t have the means to start from scratch, even if it had the time. It doesn’t have the resources to fight a two-front war. Wall Street balked at the investment that it’s making to match Google; it cannot now spend yet another $2 billion to replace its core operating system.

The twilight of the operating system as the engine for innovation may come sooner than I expected. The company may quietly decide – Steven Sinofsky and Steve Ballmer may already have decided – that future OS upgrades will be incremental rather than substantial. Attention is shifting to the network, and hosted applications. Microsoft will be fighting on Google’s ground.