Dr. Dobb's | Microsoft On Killing WinFS: End-User Value Not Ready

Microsoft On Killing WinFS: End-User Value Not Ready

Quentin Clark, WinFS product unit manager for SQL Server, said in a blog that many end-user features of the much-anticipated WinFS file system for Windows aren't ready, and the project has been shelved once again.

June 27, 2006
URL:http://drdobbs.com/windows/microsoft-on-killing-winfs-end-user-valu/189602206

If you ask Microsoft whether WinFS is dead, the answer is both yes and no. But it's clear the planned Windows File System is kaput, at least for now.

Quentin Clark, WinFS product unit manager for SQL Server, acknowledged Monday--in an update to his controversial blog, posted last week--that many end-user features of the much-anticipated WinFS file system for Windows aren't ready, and the project has been shelved once again.

Still, Clark insisted that many of the capabilities planned for WinFS have been integrated into Windows Vista and other technologies are designed for the next versions of Visual Studio, code-named Orcas, and SQL Server, code-named Katmai.

Microsoft made WinFS Beta 1 available last August, a year after the Redmond, Wash., software giant shocked the industry with news that the file system--one of three major, new subsystems planned for Vista--was being pulled from the blueprint.

At Tech Ed 2006 just weeks ago, Clark announced that WinFS Beta 2 would be delivered this fall. That plan, however, abruptly changed in a period two weeks, he admitted.

"Is WinFS dead? Yes and no. Yes, we are not going to ship WinFS as a separate, monolithic software component," Clark said in his latest blog, published Monday evening in response to the outcry that followed his first blog post. "But the answer is also no. The vision [of WinFS] remains alive, and we are moving the technology forward. But some of the technology, especially the end-user value points, are not ready. And we're going to continue to work on that in incubation."

This is the second time that Microsoft has cut WinFS from Windows feature list.

Microsoft upset many by announcing in 2004 that WinFS would not make it into the Vista client and the "Longhorn" Windows Server code. But then the company promised that WinFS would be delivered as an add-on to those Windows versions. Microsoft reneged on that promise this week, industry observers noted.

"WinFS, as we understood it, as a new file system is essentially gone," said Rob Enderle, principal of the research firm Enderle Group. "WinFS, as we knew it as a file system, is back in incubation, which means it may never show up in a form consistent with how we originally anticipated it."

Despite the such disappointment about WinFS, some industry analysts and solution providers were forgiving. They believe Microsoft's claim that many of the WinFS technologies have trickled into Vista and the next-generation ADO.NET 3.0 and LINQ integration layer for Visual Studio and in the Katmai edition of SQL Server. "SQL Server's role in WinFS is far from dead. Microsoft is just changing how they go to market with it," said Douglas McDowell, director of business intelligence at Solid Quality Learning, a Microsoft partner in Atlanta. "There's tons of great innovation coming in this arena."

Peter O'Kelly, research director at Burton Group, said Microsoft has made adjustments to accommodate key market shifts since plans for WinFS were touted heavily at the software vendor's Professional Developers Conference in 2003. The rise of Google's search engine and Microsoft's acquisition of Groove Networks, for example, have rightfully impacted the WinFS plan, he said.

Microsoft's Search and Organize feature in Vista, LINQ integration layer for Visual Studio's ADO.NET 3.0, Windows Workflow Foundation in WinFX, "Katmai" SQL Server and Groove's offline, peer-to-peer file system each encompass a piece of WinFS functionality or deliver on the promise of the WinFS vision, according to O'Kelly.

"To be fair, there's a tactical mistake, but strategically they're on track. It's the death of WinFS but the resurrection of a lot of ideas in a lot of places, and Microsoft will get to market through other means than through a monolithic chunk of code," O'Kelly said. "It's called WinFS, but there's always been a sense that WinFS was a compelling vision with a big question about how it would go to market."

Microsoft's Clark said he has no regrets about changing the blueprint but acknowledged that the marketing of WinFS has been less than ideal.

"The original goals of WinFS have never changed, but the technology we are building isn't easy. So we did take a number of internal design changes and rewrites," he said. And I am not going to apologize for that. Getting the relational engine to behave and perform like the Windows file system isn't a matter of a few lines of code. It has to be done very carefully and architected right," Clark wrote in his blog Monday.

Then why did Microsoft promise WinFS Beta 2 at Tech Ed 2006?

"When we were at TechEd, we had not made the decision. Sure, it was under discussion, but we did not have all the information we needed. And we had not made the call yet," Clark explained. "We did share the news as soon as we had the final word. We could have waited longer to disclose the information and made the change in plans less of a contrast, but we chose to notify people as soon as we could. This is why we used the blog and didn't fire up the big MS PR machinery. That takes time."

Others say the resurrection of WinFS also comes in the form of Project Orange, a group within Microsoft whose mission is to develop a next-generation Information Explorer based on WinFS and WPF.