Microsoft's plans for Next Generation Windows Services (NGWS), now dubbed Microsoft .NET ("Microsoft Dot Net"), represent the most drastic strategy change for the company since its sudden decision to embrace the Internet 3 1/2 years ago. But Microsoft .NET—a set of services and technologies (an infrastructure) that will enable a programmable, next-generation Internet—is more far reaching than anything the company has done so far. Although the cynical part of me cries out that most of the announcements about Microsoft .NET describe little more than vaporware, I also know that the underlying technology is in place for Microsoft to pull it off.

And I'm genuinely excited—with some reservations. Microsoft often evokes such mixed feelings in me: It's hard not to walk out of a Microsoft product announcement with a lighter step, the future all wide-open possibilities. But anyone who has experienced this Redmondian "reality distortion field" also knows the frustration that inevitably sets in when one attempts to reproduce what Microsoft so effortlessly demonstrated just hours before. I've had this up-and-down reaction to Microsoft announcements since the February 1996 alpha of Internet Explorer (IE) 3.0, which was basically unusable, despite its onstage heroics at the demo. The pattern will no doubt be the same with Microsoft .NET.

Not that Microsoft .NET is a simple product or set of products around which we can easily wrap our minds. Microsoft .NET is a complete remake of the company's entire product line. Windows and Microsoft Office, the cornerstones of the Microsoft empire, will become Windows.NET and Office.NET, products that will exist largely in the Internet ether, providing valuable services to PC users and to those who use an upcoming variety of nonPC devices. A pervasive new "user experience" will replace the paltry "user interface (UI)" of our suddenly ancient desktop systems. The user experience enables voice control and feedback, handwriting recognition, and smart links that magically know how to connect previously disparate parts into a cohesive whole. The promises of the Internet come together in a Jetsons-like future that's suddenly upon us.

No, Microsoft .NET won't all happen at once. Microsoft admits that the transition to its Dot-Net future will require at least 2 years, which makes me wonder if a 3-to-4-year transition is more likely. We'll glimpse this connected future first in Visual Studio 7 (excuse me, Visual Studio.NET), due in alpha this summer, with a final release in the spring. Also due in the spring is Windows Whistler, the follow-on to Windows 2000, which will ship in Windows.NET form as well as legacy versions for the unconnected. And Office 10 could quite possibly be Microsoft's last desktop version of Office.

Although it's easy to get caught up in the excitement of this initiative, note that Microsoft .NET represents the one thing that the company's detractors most feared: a Microsoft tollbooth on the Internet. With its Dot-Net plans, Microsoft isn't changing only how the company ships software but also how you pay for that software. Instead of yearly upgrades to Windows and Office, you'll pay a monthly subscription for access to Microsoft's superior superset of the Internet. Microsoft has been eager to move to this model for some time now. Once the company moves to a service-based subscription model, you'll receive the benefit of constant but minor upgrades rather than steep upgrade curves (and costs) each time a new version of its products is released. You won't be a Windows user per se, but rather a subscriber to the Windows service. Presumably, Microsoft will offer varying levels for each subscription service in the same way that cable-TV companies offer different plans.

It's baffling, isn't it? Although I have a hard time reconciling today's world with the Microsoft .NET world, surely we'll look back one day at the quaint, shrink-wrapped software we now use and shake our heads. Did we ever really live in such times?