Molly Wood, senior editor of CNET.com, recently asked why people would bother to upgrade to Longhorn when Avalon and Indigo will be available to WinXP/2003 users, and WinFS has been deferred entirely. In her article, she concludes that "Windows is weak," leaving the door open for competitors to eat Microsoft's lunch.
It won't be an easy sell, convincing people to give up XP for something new, but that's the sell that Microsoft is hoping to make with Longhorn. As long as Microsoft is going door-to-door, hat in hand, why shouldn't someone else beat it to the buzzer?
On the one hand, providing early information about Longhorn helps to make Longhorn a product that better fits consumer needs. Microsoft's decision to push WinFS back was predicated on the problem of enabling the transport of WinFS metadata in distributed environments. The fact that it was largely tied to one computer isn't acceptable in the modern networked world, and alpha users told them that. Similarly, developers aren't going to write to the WinFX API if it means their creation can't run on the large installed base of Windows XP systems. Hence, Avalon and Indigo are available for older Windows versions.
On the other hand, people start to fixate on the "gee whiz" features (in particular, the big three--Avalon, Indigo, and WinFS) and use them as the tea leaves within which to read the future of the Longhorn operating system. If the amusement park moves in a few blocks away, why go to the amusement park across town?
I think this misses the essential reason people upgrade their operating system. It's a rare thing for a new OS version to involve a "revolutionary" change. Windows 95 was "revolutionary" in the sense that it inaugurated the shift from 16 bit to 32 bit. Windows 98 and Windows ME were "evolutionary" changes, in that they were simply OSes that offered more features than a previous version.
The "Tiger" release of Mac OS X is also an evolutionary change. I tracked down the feature list for Tiger, and I can't say that any of the changes make me open up the door and shout "wow, this is the most amazing operating system ever built." These are feature improvements, and they are good. They are reason to upgrade, but they are not revolutionary.
Longhorn features are revolutionary, a fact noted in the article I wrote shortly after the Professional Developers Conference. The complete refactoring of the huge Windows API, the use of .NET as the primary API for Windows, and all the unique features of Avalon and Indigo are revolutionary. Those features are being made available on Windows XP and Windows Server 2003, but that doesn't make the features themselves any less revolutionary.
It just makes the upgrade decision less revolutionary, since you aren't buying into the product, which will have exclusive rights to Indigo and Avalon. You will, however, have a whole slew of feature improvements...just like Tiger. Many of those feature improvements don't neatly fit into the "Avalon, Indigo, WinFS" categories.
I bet lots of Apple users will upgrade to Tiger for the new features. I bet lots of Windows users will upgrade to Longhorn for the same reason, and Longhorn's features are far more revolutionary than Tiger's features.
So, I don't think Microsoft will have trouble getting people to upgrade, particularly as more of the Windows software world starts to use the Longhorn APIs safe in the knowledge that their creations can run on Windows XP and Windows Server 2003. Microsoft is likely to do the same thing Apple is doing to promote its new OS version, showing all those nifty new features and showing how WinFX software runs better on Longhorn. Furthermore, they will have an easier time of it, because the core features developed for Longhorn are revolutionary, even if some of them are available on Windows XP.