The Windows File System (WinFS) is key to two of Microsoft's biggest next-gen products in the pipeline: Longhorn and Yukon.

While Microsoft released Yukon Beta 1 to testers in July, there have been few analyses of the WinFS-specific portion of it by folks who have actually played with Yukon. So we decided to call on Yukon expert Randy Holloway to get the inside, pre-PDC scoop on WinFS. Holloway is the coauthor — with Andrej Kyselica of an upcoming book "Developing Solutions with Yukon: Beyond Transact-SQL" (Addison Wesley) — on the ins and outs of Yukon. He also is a prolific blogger who writes about Yukon and the Common Language Runtime.

Read Our Previous Yukon Q&A with Holloway

Check Out Holloway's Blog

We grabbed a few minutes of Holloway's time this week to get a pre-PDC update.

Microsoft Watch: You've had some time to play with the first beta of "Yukon" for a couple of months now. What are your first impressions of the product?

Holloway: I think that the Yukon team has done a great job. The amount of progress that has been made with the product this year alone is tremendous. At the PDC next week, people will finally get to judge for themselves.

Similar to most early product releases, the core functionality is pretty strong. Sure, there are some issues with the software that need to be worked out, but that is to be expected. There is a lot of work left to be done around the management tools for the product. The SQL Server team has set some very ambitious goals for the functionality of the new SQL Workbench tool. That tool is designed to replace the tools that people are accustomed to using currently, tools such as Enterprise Manager and Query Analyzer. I'm sure that they'll be refining the SQL Workbench features up through the next beta release as well.

"All PDC Roads Lead to WinFS, Page 2

Microsoft Watch: We've heard WinFS described as "The new Windows File System (WinFS) will be able to store structured, file and XML data in a way that optimizes it for organizing, searching and sharing." Does it live up to this description?

Holloway: The interesting thing about WinFS is that it has really morphed into something different than was originally expected. WinFS is essentially a searchable metadata system that will function on top of existing file systems such as FAT or NTFS. Although this technology is based on technology that was developed by the groups focused on the SQL Server product, WinFS isn't really part of the Yukon release per se.

There will be more information about this released at the PDC and people will finally figure out what "that other thing" — as some of the SQL Server folks have referred to WinFS over the past year — actually does and how it is actually implemented in Windows.

Microsoft Watch: We know, from Microsoft statements, that they are planning to run WinFS "on top" of NTFS. But do you think it will be possible to take all of WinFS and just move it wholesale into the OS? Or would you guess that Microsoft will take elements of WinFS and move them into Longhorn? (I know you are just giving your "educated guess" on this, as we have yet to see the latest Longhorn builds.)

Holloway: I'll be interested to see how WinFS will be positioned, since it is essentially a "service" that will run on top of existing file systems.

Theoretically, this type of service could be integrated into Windows XP to improve the support for searching on the current version of the OS. My guess is, however, that Microsoft will preserve WinFS for Longhorn only.

The reason for this I think is going to primarily center around the user interface considerations. The Longhorn user interface (UI) will need to be substantially different from the current UI for XP when it comes to viewing files and searching for data. My thinking is that these kinds of substantial updates will not be made available as a "bolt on" feature to the current Windows versions.

Microsoft Watch: From PDC session materials, it sounds like other Longhorn technologies, such as the Avalon graphics/presentation application-programming-interface (API) set, and the Indigo Web services stack, will integrate directly with WinFS. What will this do for developers/customers?

Holloway: Any new API that is well designed should be consumable by a host of applications. I think that Microsoft is really trying to tie these technologies together in a way that will be compelling to developers at the PDC. Their goal will be to have developers walking out of that conference thinking about the new applications that they can build and bring to market. Applications that really leverage these new features will be part of the criteria that customers will use when deciding whether or not to switch to Longhorn.

With WinFS and its programmability features, there will be ways to more efficiently search for relevant files on the systems that connected to the network using the metadata supported by their file systems under Longhorn. Using the types of services that Indigo may support, the data could be aggregated and stored on network storage devices while also being catalogued and associated with relevant data in your business systems.

The possibilities are really endless, but making it easier to find information, share it with other people or other computers, or to aggregate the data for use later are all functions that will be supported in Longhorn much more effectively than they are today.

"All PDC Roads Lead to WinFS, Page 3

Holloway: WinFS has to be one of the most important features in the product. I would rate it right up there with the UI features that will be provided through Aero and Avalon, maybe higher.

I think that in terms of potential to boost user productivity, WinFS has to be the number one feature in this product. I'm sure that Aero and Avalon will serve to attract Windows customers into upgrading from Windows XP, but WinFS will be one of the core features that really makes Windows a better computing experience for the average user. Simply put, WinFS will enable you to do things with Windows that you cannot do today or that you can't do today in any reasonable amount of time. That will be compelling.

This is an edited version of an article that ran in the October 23, 2003, issue of the Microsoft Watch newsletter.