How about a really bold Windows Vista prediction?

Gartner boldly predicts that Vista is the last big Windows release ever. Oh, now that's daring. I connected a few extra dots and have a bolder prediction. What if Microsoft carved Windows up into pieces that didn't have to run on Windows at all?

I realize I'm a little late to the party on this story, but I'm just now emerging into the world again after three weeks of essentially 'round-the-clock work. Glad that's over.

So, anyway, I missed the original report where Gartner boldly predicts that Vista is the last big Windows release ever. Oh, that's daring. I'm going to go out on a limb and predict the sun rises in the east tomorrow. The difference is you'll know exactly how accurate I was tomorrow morning, whereas we'll have to wait 5-7 years to see whether Gartner was right. I guess after totally blowing their Vista ship-date prediction they decided to issue one with a longer time horizon.

But just because it's obvious doesn't mean it's not interesting. Long Zheng tumbled across a particularly relevant Microsoft patent application. Mary Jo already did a good job of connecting a couple of other dots, rightly noting that Gartner's prediction is "sweeping and vague":

How is Gartner defining "big"? And what will count as an "operating system release," going forward? Will a service pack count? Does a rollup of hotfixes and new features constitute a new Windows release?

What about Microsoft's so-called "Cloud OS" project? Or Microsoft's ongoing moves to modularize Windows (as further evidenced by its patent application for a pay-as-you-go OS)? When elements of what used to be part of a desktop-based version of Windows debut as services, does that actually make an operating system release "smaller," in terms of size and/or importance?

Absolutely right.

Here's one observation I can add from personal experience. Gartner flubbed their Vista ship-date prediction earlier this year because they didn't notice that Vista is being built differently than any previous version of Windows. The modular nature of the project is what allowed it to come together at the end in record time (faster than I thought possible) and with better-than-average quality for a Microsoft OS release. (You may interpret that any way you want.) During the beta testing process, those pieces weren't always in sync, leading to some pretty miserable experiences for testers until the very end.

The point is that the modular Windows we're discussing already exists. The different editions of Vista are really just a base version plus some feature packs. The Media Center add-on is probably the best example, being in its fourth major iteration in just over four years and being developed separately from the underlying OS. Although it's been sold as a separate operating system, it's really just a big feature pack. When you upgrade from Vista Home Basic to Home Premium, you're essentially adding on the Media Center feature pack (and the Tablet PC feature pack and the Aero interface pack, etc.).

If you upgrade from Home Basic to Business Edition, you get what could be called the Advanced Networking and Business Tools package, with support for corporate networks, better file encryption, the ability to host a Remote Desktop session, image-based backup, and fax and scan tools. Plus the Aero pack and the Tablet PC pack, but not the Media Center pack. If you want the whole shootin' match, you get Vista Ultimate, which has all those modules included.

So it doesn't take much insight to predict that a future Windows OS will consist of some sort of basic package (a kernel plus more "other stuff" than most purists would like to see) plus feature packs and services. In fact, within a couple of years Windows Vista could easily be sold the way Dell sells PCs, with loss-leader base versions and a build-your-own-edition that you configure with a mix of features and services, and which can be upgraded at any time, the way you add features to your wireless phone provider's package.

And who says those features have to run on Dell PCs? If I wanted to make a truly bold prediction, I'd try something like this:

Within three years, maybe less, Microsoft will have a version of its Media Center software specifically designed to run on Intel-based Macs, using either native OS X or a Windows virtualization layer.

A Microsoft Media Center for the Mac could use the exact same interface as its Windows counterpart. If it supported all the add-ins that are being developed for Media Center, that would be a huge plus. It would need to support a very limited assortment of hardware and could even lag its Windows counterpart by a few months in the release cycle. The Office team has already proven that Microsoft can sell a lot of software and make a lot of money this way.

What if the Media Center Program Guide and scheduling features came with a monthly charge, just like TiVo, but purchasing this year's Media Center edition got you a free one-year subscription? There's your incentive to upgrade. And what if they found a way to support CableCard devices and conditional-access satellite decoders? That sounds like a TiVo killer to me.

Yes, crazy, I know. But if you're going to make predictions, you might as well be daring. 

And wouldn't it be ironic if the Mac won the battle for the living room but Microsoft wound up making more money?