When it comes to understanding IT, groupthink only gets you so far

Earlier Friday I was speaking with a CIO when the conversation turned to the subject of Microsoft. There's been no small amount of reactionto the publication of Gartner Group's gloomy report on Windows. But this exec was not buying into the notion of a future tech landscape where Web browsers elbow aside client operating systems as the preferred software development platform.

"We're still on XP. I'm not going to move to Vista for a while. We'll let other people be the early adopters," she said, asking to remain unidentified in case Microsoft happens to read her quotes. Still, she added, "I just don't see a world in which Web apps make the OS obsolete anytime soon."

Anytime soon is the operative phrase.

One quote doesn't a trend make--except when the person delivering the lines agrees entirely with me. (Just kidding.) In their quieter moments, I suspect that most of the participants in the bloviation-fest which attended dissemination of the Gartner report conclusions would agree. But the groupthink around this topic is leading to entirely off-the-wall prescriptions--especially when it comes to promoting online advertising as the magic answer to Microsoft's troubles. I'm on record writing that Microsoft CEO Steve Ballmer would only be wasting billions of dollars to buy Yahoo, but enough on that topic for now.

Microsoft has a big challenge figuring out how Windows will thrive in a world where more client apps are operating system-agnostic. But that's not the same as proclaiming the demise of the client OS. I don't know how many people attended the speech or read the report, so consider the following from the report:

"There's lots of discussion on how Web 2.0, Ajax and open-source products will make the client OS unimportant and unseat Microsoft as the dominant desktop software vendor. Some even insist that the client OS already doesn't matter. The client OS may be less important today than it was 10 years ago, but that's a more-accurate description for application developers trying to decide what OS they want their applications to run on. New applications are increasingly OS-agnostic, but legacy applications were very often developed for a specific OS, usually Windows. Legacy applications remain installed and important for years, meaning that for enterprises, the client OS is still a critical choice and will be for years to come."

Gartner is more concerned with the changing definition of an operating system if a virtualization hypervisor takes care of the interface between hardware and application. (Hasn't that been the traditional role of an OS?) The report then geeks out on what Microsoft needs to do to retrofit and rescue Windows from oblivion, etc., etc. But nowhere does it make the silly leap of logic mentioned above.

Microsoft's seemingly getting attacked from all sides these days, but give the company some credit for not being completely clueless about the changes taking place around it. Apropos, ZDNet's Mary Jo Foley offers several good points to consider:

•  Windows currently contributes one-third of Microsoft's revenues and two-thirds of its profits, I've heard company officials claim. Windows is installed on more than 90 percent of consumer and business desktops combined. That market share isn't going to disappear overnight, no matter how much Web 2.0 pundits and online-services vendors want that to happen.

•  Windows 7, from all accounts, is going to be a minor upgrade to Vista. It is not going to be the start-from-scratch, slimmed-down operating system that many believe Microsoft is building in a back room as a "Plan B." Singularity, the Microsoft Research microkernel OS, also is not that brand new Microsoft operating system. However, I do believe Singularity is the core of what ultimately will become a brand-new distributed OS platform from Microsoft. Unlike Gartner, I'm not going to pick a date out of a hat (by 2011!) and claim that's when such a platform will be announced.

•  As has been reported previously, Windows 7 is likely to include a feature that, at least at one point, was called the "Component Delivery System" which is expected to allow users to install the pieces of Windows that they want and need in a more user-configurable way. This may not be identical to the modularized role structure offered in Windows Server 2008, but it is similar in its intention. This should help, to some extent, with Windows' bloat--as should Microsoft's expected move to use Windows Live to deliver non-core pieces of functionality to users.

•  Windows 7 also is likely to include some kind of virtualization layer that will help ease backward compatibility, I've heard from various sources. Microsoft isn't likely to a port of Hyper-V to Windows client. But it could take the form of a virtualization service like SoftGrid (Microsoft's application virtualization offering) and/or hosted desktop virtualization (the new name/positioning for Microsoft Terminal Server, I hear).

That's the sort of granular analysis that won't flag much attention on Techmeme. But you know as well as I that headlines often don't tell the full story.