Office conference wrapup

The Bill Gates keynote held no surprises, and in any case will likely be transcribed and published, but during the Q and A several points came out that I found noteworthy and may not be transcribed. First, a question about unified storage, from someone whose name I didn't catch:

Attendee: Those of us who saw the webstore in Sharepoint 2001 thought, wow, that's sort of what Cairo was going to be. And now looking at WinFS, it has vague echoes of what webstore was going to be. Is SQL really the underlying storage that's going to be in Sharepoint in the future?

Gates: Yeah, what's happened is that there's this dream of unified storage, which is the world of files, mail, records, all these things coming together in a very rich store. That's a dream we've been investing in for a long long time. It's about taking a very advanced version of SQL that can deal with XML, and can deal with streams, and putting a very high-level data model -- you could almost call it an entity-relationship data model -- on top of SQL, so it can deal with all these things. And that's the path we're going down. WinFS is merely the client implementation of that strategy. And so what Sharepoint is going to sit on top of is a database engine, and WinFS is just a framework on top of this database engine, those are one and the same thing. It's the next big iteration of SQL that gives us all of those powers. It's perfectly symmetric -- client to server, WinFS to Sharepoint. It's not even clear that WinFS and Sharepoint are necessary because what you're going to see is they're exactly the same thing. Sharepoint just evolves up on the server, WinFS evolves down on the client.

The epochal industry-wide significance of SQL/XML hybridization, and what Quentin Clark has called the object/relational/XML trinity, is something I've thought and written a lot about. What I hadn't considered is the notion that Sharepoint's real mission is to encapsulate the "next big iteration of SQL" on the server in the same way that WinFS does on client. That's a bit of a rhetorical stretch at the moment, but if Sharepoint is now the label for that initiative, then OK, it clarifies the goal.

Second, a question about electronic forms from Michael Herman of Parallelspace:

Herman: It's nice to see Microsoft consolidating around a smaller set of core technologies, but when it comes to electronic forms, Word and Excel have their own point solutions, Outlook has its own point solution, InfoPath has its own point solution, Access has its own point solution. In the developer platform you have ASP.NET and WebForms. We're constantly in the situation where we're trying to guess which ones are strategic. Can you give us some insight?

Audience: Laughter, applause.

Steve Sinofsky: I think in that short period of time you've summarized the last ten or twelve 101s one-on-ones that we've had.

Gates: He didn't mention Avalon, though.

Audience: Laughter.

Steve Sinofsky: I think that's a good summary, and I honestly think its a fair assessment and a fair critique. For me, the slide about the rich client and the thin client, and the tradeoff you have to make, for us in the forms space, we're constantly battling these kinds of tradeoffs. Do we want the calculations behind Excel, do we want the rich editing in Word, do we want the thin deployment of ASP.NET, all of these tradeoffs, and we just struggle with trying to have the one that solves all of the problems. What I can say for sure is that all of the ones you mentioned, we're going to continue to support for a very long time. That said, where we are now in the Office system, you're seeing us on the Web-based side focusing on Sharepoint and on the client side focusing on InfoPath. For me, InfoPath pulls together all of the elements Bill talked about today: it has a rich client interface, it has improvements in deployability and management, it's XML-based, it has all the connectivity to Web services. Have many you looked at InfoPath and thought about it at all? Those of you that have are probably asking us where's the thin-client element of it. We kind of hit that right away, so that's clearly some feedback we've got, and some things we're working on.

Gates: Today we have HTML and then our rich forms, where InfoPath is the one that's definitely rising. Where we want to get to is that InfoPath's the high level, then we have all these rich controls we can use, and underneath we have the Avalon runtime. And so we have a roadmap for InfoPath that gets richer and richer, embraces our controls, and sits on the latest presentation system.

This exchange struck me as slightly surreal. InfoPath is the best example of what Microsoft's thin-client strategy could have been had its browser continued to evolve. The product is built on the IE engine, XSLT, JavaScript, and CSS. It was well-positioned for thin-client deployment. Technology didn't prevent that from happening, business decisions did.

According to Micah Dubinko, Mozilla's beta XForms implementation is now available as a click-to-install XPI file. I haven't had a chance to try it, but it's yet another reminder that the standards-based browser isn't the dead end Microsoft wants it to be.

Former URL: