WinFS back-burnering could spell death sentence for next-gen file system

Posted by Scott M. Fulton, III

Redmond (WA) - In what appears to have been the final straw for the feature Bill Gates touted in a 2003 keynote as the "holy grail" of Microsoft projects, the Windows File Store (WinFS) has apparently been demoted from its status as an optional download at the time of Windows Vista's release, to a a possible objective storage methodology to be used in future editions of ActiveX Data Objects (ADO.NET) and SQL Server. The news came from WinFS lead developer Quentin Clark, in a melancholy post to his team's MSDN blog last Friday.

The original intention of the project, at the time Gates announced it, was to completely replace the filing system of Windows with an object-oriented, search-ready engine that was queriable using a variation of SQL. A desktop-based command line, Microsoft said, would revolutionize the way users store and search for files. The file allocation table would effectively be replaced with an SQL schema. As the project evolved, however, it became clear that the company's driver development model couldn't easily be based around a file system that was essentially SQL-based...and yet remain compatible with earlier versions of Windows. WinFS became a supplement, and an optional one at that. Now it's a research project.

"There are many great technical innovations the WinFS project has created - innovations that go beyond just the WinFS vision but are part of a broader Data Platform Vision the company is pursuing," Clark wrote. "...These changes do mean that we are not pursuing a separate delivery of WinFS, including the previously planned Beta 2 release. With most of our effort now working towards productizing mature aspects of the WinFS project into SQL and ADO.NET, we do not need to deliver a separate WinFS offering."

Last year at about this time, the rumors of WinFS' cancellation seemed exaggerated. After a news story was propagated throughout the net stating WinFS was history, Microsoft triumphantly posted the Beta 1 version of WinFS on a separate development track. But this separate track meant that WinFS was no longer a synchronized, orchestrated component of Longhorn (Windows Server) or Vista. Like so many other Microsoft grails of varying grades of holiness, WinFS was left to sink or swim.

When Google Desktop was introduced in October 2004, the buzz was that this would be the first product in quite some time aimed squarely at Microsoft's "holy grail." In absorbing Microsoft's buzz, Google may have effectively deflated that company's vision.

Just two weeks ago at TechEd 2006, the WinFS team demonstrated capabilities of WinFS Beta 2 that sadly, now, may never be tested in the field. Around the net this past weekend, Clark's positive adjectives, such as "encouraged" and "super-excited," fell on a disenchanted blogger community. In Web Services Contraptions, Jelle Druyts comments on Clark's announcement:

This article effectively ends Microsoft's second push to move to a relational file system. The infamous Cairo OO-OS in 1991 that was supposed to be built on NT, and then WinFS, as one of the pillars of Longhorn in 2002...My guess is that this whole thing will be completely tabled until after Windows Vienna ships - this would probably make it 2010 until it comes up, add 5-10 years development efforts, so you might have a chance seeing this by 2015.

In Joseph Scott's Blog, the author comments:

Given the last update to SQL Server (2005) took five or six years, is there any hope of seeing WinFS in SQL Server before 2010? Of course if you want to bring up file system announcements from Microsoft you could argue that the basis of WinFS was announced almost 15 years ago with Cairo. I expect that the general concept of WinFS will buried from a PR stand point as quickly as possible by Microsoft.

Incidentally, Microsoft spokespersons had no comment today on this news. Finally, there's this contributor to Tales from the Basement who sees no new holy grails in Microsoft's future:

I don't like to knock Microsoft, I have a lot of respect for what they have achieved, but I can't help thinking that over the last year or so things are not looking so good for them, and although I love Visual Studio, I have no similar feelings for Vista or some of the other technologies that are emerging from Redmond just lately.

"Of course," Clark's Friday post concludes, "there are other aspects of the WinFS vision that we are continuing to incubate...Just as Vista pushed forward on many aspects of the search and organize themes of the Longhorn WinFS effort, Windows will continue to adopt work as it's ready." The keyword there is "incubate," which is a phrase we've seen used to denote projects that are no longer on any official schedule. "As it's ready" or "when it's ready" appears to be an increasingly overbooked date in Microsoft's corporate Outlook calendar.