Having read the post, it doesn't seem as dire as you're suggesting. Basically, there was a recognition that WinFS was really a host of different technologies and was split along lines that made more sense. Roughly,

Angel The data API and model is being rolled into the next version of ADO.NET as LINQ and Entities, which were discussed in a recent video with Sam Drucker and Anders Hejlsberg. Seems reasaonable, since it would be cumbersome to have to learn one data API for the filesystem and another for everything else.

(b) The advances in self-management and integration of unstructured data are being put into the next release of SQL Server. Again, good call. It would have made absolutely no sense to offer a self-managing relational data store for free download and then try to sell a separate enterprise-grade RDBMS to go with it. Essentially, you'd be buying SQL Server for the reporting and data mining capabilities while you wondered why the management smarts weren't in SQL Server.

(c) Everything else goes... where? As for the UI and filesystem integration, my guess is a future release of Windows, which seems reasonable in that introducing these changes via a free add-on to existing Windows installs might be somewhat disorienting for many users. Much of the search functionality has mades its way into Vista. The part I was looking forward to was the common data synchronization engine, the need for which seems more obvious in light of all these "live" services; but then maybe that's the niche SQL Server Everywhere is targeting.

In short, it's messier than before, but I think the end goal is more or less the same.

WinFS is dead; long live WinFS! Smiley