It used to be an accepted truism that it was worth waiting for Service Pack 3 before adopting any new Microsoft operating system. In the last few years, they've managed to knock that back to SP1 and still satisfy most corporate IT buyers' comfort levels. But in a recent Gartner research note, analyst Michael Silver claims that Windows 7 will be so stable you won't need to wait for SP1 to come out, even though you probably will anyway.

The truth is that this rule of thumb has always been based on a rearview mirror perspective anyway, and that the reality has been different for almost every operating system released since 2000. But Silver is correct that most organizations aren't particularly well set up to implement a new base OS release as soon as it comes out anyway, and the interval in which they may be preparing for deployment is likely to coincide roughly with the first service pack release anyway.

From my observations, 7 has been easily solid enough to substantiate this advice. But it brings up a larger question, which Vista can be credited with exposing more broadly in many organizations: what are your motivations to upgrade, if your assessment of the target platform is narrowed so simplistically to its stability? Is that really all you are looking for? And if so, doesn't your existing platform already fill the requirements?

I've been as guilty as anyone of assuming that an upgrade to Windows 7 is a foregone conclusion. As much as I decried Vista, my main argument boiled down to "Wait for something better." But what if you already have something better? Don't you? It's paid for (at least in some cases... if you're in a subscription model, given everything you've already seen Microsoft pull, then I have nothing further to say to you), it is probably already solid and working, and everyone is familiar with it. How often do you actually rely on Support? Does EOL really mean that much to you? Maybe, instead of adopting "Skip Vista" as my mantra, I should have been saying "Skip the Desktop."

In five years, when XP finally hits Extended Support EOL, what are the odds you're going to be doing significant processing on the desktop still? It's a serious question; I'm terrible at predicting this sort of thing, I either underestimate adoption rates or overestimate organizations abilities to make rational economic choices. But a lot of the trends that would tend to move the industry back to a server-centric paradigm are moving faster than I would have imagined. In five years, Microsoft itself is going to be well-positioned to serve a lot of your existing desktop needs via virtualization or web-based services. Of course they would love to charge you for a desktop OS license in addition to the server-side licensing you'll need, but at that point, why would you continue to play the game? Every study I've ever seen proves what a massive maintenance drain smart clients in general and Windows in particular is on IT department overhead. If you are incurring all those costs for the privilege of having the user fire up a terminal session or similar and do all their real computing elsewhere, what exactly are you gaining in the bargain?

Maybe you should wait for 7 SPx before making your call after all....