Yes I mean, first off I'd like to say that we're not really saying that all Web applications ought to be Rich Internet Applications, that form-based applications should all now have video, or audio, or MP3 streaming or all the sort of rich media, or the rich user behaviours. HTML is still a great thing and it's not going anywhere. But there are specific cases where data really benefits from a richer user experience and from richer involvement with media. One of my favourite batch of use cases comes from the bucket of occasionally connected applications. An example would be the salesman or a business person who gets on an airline and he wants to continue to interact with his enterprise's data while he's in flight and there is no network, he's not connected to the network, but he still wants to be able to interact with it. And then when he's online again, the application should instantly know and message data asynchronously back to his data store, if that's what the application requires. But the functionality of an application shouldn't entirely be defined by things in the client that always insist on the presence of a network. Occasionally connected clients also include things like intermittent connectivity. I guess an analogy would be someone on a cell phone who's travelling through a subway tunnel. We ideally would want to keep the appearance of connectivity even when the network is down, and the moment the network is available again, the Rich Internet Application would be able to take care in the background of asynchronously synchronizing data exchange. So the occasionally connected case is an interesting one. There are also cases that are sort of classic to peer-to-peer. And now peer-to-peer has in a lot of cases a negative connotation, because people think of MP3 file sharing, Kazaa, and Napster and that sort of thing. But it actually is relevant in the enterprise too to the extent that we want to be able to support peer-to-peer collaboration on business documents like spreadsheets and presentations and other materials in a collaborative setting that don't involve a central server.

There's also a batch of use cases related to context. Usually a Web browser doesn't know a whole lot about the user who's using it. There are some environmental variables and some cookies and some basic data that a browser has but other than that, those clients don't know very much about the user. So it doesn't know for instance if the user is in a meeting giving presentations and doesn't know if the user is engaged in an IM session, doesn't know if the user is reading a blog, doesn't know what the user is doing from moment to moment. So rich internet applications are a little bit smarter about what the user is doing and know whether he's sitting on his couch watching his TiVo or whether he's giving a presentation and can adjust accordingly. So those are several cases that I think are pretty interesting in the RIA (Rich Internet Applications) case that aren't usually dealt with in the Web browser, thin client case.