SQL Server 'Katmai' Lacks Anticipated WinFS Features

When last June, Microsoft decided to discontinue work on its radically redefined relational file system, then called WinFS, the team announced some of its work would be forked off and merged into the next version of SQL Server, code-named "Katmai." After Bill Gates had proclaimed WinFS in 2005 to be the game-changing component for his future vision of Windows Vista, developers were curious to find out how something that monumental would manage to move mountains as part of a database manager rather than an operating system.

How much of Gates' vision survived the exodus? The answer came on Wednesday, and it's not much.

Amid some of the now-commonplace marketing phrases that appear to have emerged face-up from a random set of rolled dice, including "Redefining Pervasive Insight" (we're not sure what the original definition of "pervasive insight" was), Microsoft revealed that one item from the original WinFS vision did survive. It's an esoteric one, to be sure: The new version of ADO.NET (a revision of ActiveX Data Objects to be managed by the .NET Framework) will expose elements of data as simple objects with complex underpinnings, rather than a complex hierarchical network of tables and records.

It's the objective vision of WinFS all right, but it doesn't apply to the file system, at least not yet. For now, it applies to the next version of Visual Studio, code-named "Orcas," in which .NET languages will make greater use of inline queries, using Microsoft's new LINQ lexicon, to query and retrieve large sets of data from networks.

Here's how Microsoft's new marketing brochure now explains this feature: "As part of the next generation of the ADO.NET data access framework, developers will be able to access their data by defining business entities like Customers, Orders and Products instead of tables and columns using the Entity Data Model. Query and retrieve these entities natively within any .NET language with the introduction of LINQ. These services enable developers to work within the logical entity model while administrators can define the physical implementation of the model as tables and columns."

Had Gates led with that back during that now-infamous PDC presentation in 2005, he might not have gotten the same tremendous response. As it stands now, what remains of the WinFS vision is relegated to the back page, paragraph 14, of the company's latest marketing brochure.

The original idea behind WinFS was to create a repository of metadata about files and their contents within systems and networks, and have that repository be centrally managed by the next version of SQL Server. So the decision to move WinFS development to the database division was not an arbitrary one. But desktop search features, including one released notoriously for free by Google, rapidly deflated Microsoft's excitement about the ability to search one's own hard drives using natural-language queries and everyday terms.

In one of two "farewell" messages posted on the WinFS team blog last year, developer Quentin Clark described the transition to a SQL Server project as being in the best interest of the broader body of users. "The Entities features we are now building in ADO.NET started as things we were building for the WinFS API," Clark wrote. "We got far enough along and were pushed on the general applicability of the work that we made the choice to not have it be just about WinFS but make it more general purpose.

"With most of our effort now working towards productizing mature aspects of the WinFS project into SQL and ADO.NET," Clark continued then, "we do not need to deliver a separate WinFS offering. Be encouraged that we are able to get the underlying feature work into Orcas and Katmai. It's great technology and we are super-excited to be productizing this way. And most importantly, it's what people have been asking for."

But what Vista proponents who remember PDC 2005 wanted to see was something close to the pervasive UI improvements that a radically reformed file system might make to Windows - the subject of Bill Gates' spectacular presentation, as commemorated by this video from the show. What would happen to that bigger, bolder vision?

As Clark responded last year, "Hey – we are very busy finishing Vista, and just aren’t ready to talk about what comes next. The vision for a richer storage in Windows is very much alive. With the new tools for searching and organizing information in Windows Vista, we are taking a good step towards that vision."

For SQL Server developers, the continued evolution of what's being called the Entity Framework is indeed extremely important. Conceivably, it could lead to a potentially radical new way of thinking about databases, more similar to a network model than a tabular one, with relations defined more in terms of relevance than presence or absence.

But the bridge between the idea for this technology and its potential everyday consumer application is clearly gone, it's difficult for a good idea to regain not only the momentum but the high-level support it needs from within the company, to finally get off the ground.

In an ADO.NET team blog posting late last month, architect Mike Pizzi summarized the current state of the Entity project much better than Wednesday's announcement: "Microsoft will be leveraging the Entity Data Model in future versions of Microsoft products such as SQL Server. This Data Platform vision enables customers to leverage their investment in a common conceptual entity model across product lines."