SQL Server 2005 shipped last November and has been a bright spot for Microsoft lately, making rapid gains market share. The next version of SQL Server, codenamed Katmai, will receive some WinFS bits. It's planned for release some time in 2007.

ADO.Net is Microsoft's object library of choice for data access. It will be updated to include WinFS technologies for data access in Orcas, the codename for the next release of Microsoft's Visual Studio development environment, also planned for a 2007 release.

WinFS was originally meant to be a new file system with the Longhorn operating system, but that idea was jettisoned two years ago. It was supposed offer a new method for storing data, getting away from the whole file and folder structure we're used to. Instead, Windows applications would display the files associated with a particular request.

The first beta came out at last year's Professional Developer's Conference and a second was promised for this year. Five sessions covering WinFS were held at TechEd 2006 in Boston this month. Internally, Microsoft had been discussing making just such a decision to splinter WinFS prior to the show.

So what happened?

"No decision had been made yet when TechEd came about, so we were faced with a decision; to cancel sessions based on a hypothetical, which didn't make sense, or go forward with the discussions and put the word out later," said Corey Thomas, group product manager for SQL Server at Microsoft.

The move to separate the pieces of WinFS to individual products will serve both well, said Thomas. SQL Server has done well as a relational database with data stored in rows and columns, but it's lacking in support for unstructured or object-based data. WinFS will help plug that gap.

"Up to now, SQL Server has not been an unstructured data store," he said. "What we're embarking on now is expanding and helping customers solve another set of problems around unstructured data and helping leverage that with structured data."

As for ADO.Net, the WinFS API will provide developers with a higher level of abstraction over a greater variety of data structures and allow programmers to work with entities and not underlying data structures, said Thomas.

Splitting WinFS among several products was a necessary move because WinFS was trying to solve too many problems with a single solution, said Peter O’Kelly, research director for collaboration and content at The Burton Group.

"There's probably a lot of people who are relieved Microsoft isn't trying to do the all in one solution," he said. "The reason Microsoft got into this situation is they put four or five different facets of data aggregation into one thing. Now they are saying it makes sense to have them go out with the applications the different facets are more closely aligned with."


Finding Open Source Bugs Like Klocwork
What's The Top IDE?