Facebook applications are cool and popular right now. Everyone wants one. Indeed, I found it very interesting that BJ Fogg is teaching a class on Facebook Apps. This class seems relevant and interesting. I'd take it!
However, I'm actually not a fan of (using) Facebook applications. I'd rather give applications a whole Web page to work with. I believe that Facebook applications are written because they can draw many users through viral marketing, not because they are generally better or easier to deploy than stand-alone apps.
The benefit that I see from a user's perspective is that it becomes much easier to engage in social applications. Our research lab has been playing Scrabulous on Facebook for a few weeks now. Would anyone have played had they not already been part of the same social network? In general, I would like to see data comparing the number of games played on Facebook with the number played on scrabulous.com.
Compare Facebook-style social applications with MovieLens buddies, an old-style social feature that allows users to share ratings and receive movie recommendations together. Nobody uses it. Now, if MovieLens had a Facebook application, would we also benefit from the viral effect? Is a Facebook application the path to (research) riches and glory?
Max
Most web apps are used
Most web apps are used casually, and this is where the Facebook platform shines. Instead of having your app in the user's bookmarks (everyone has too many), or in her memory (you have to have a REALLY GOOD app to stay there), you can get your app on a page the user visits frequently.
As a developer it's sometimes difficult to come to terms with the fact that the application is not as important for the users as it is for the developer of the app. I think it's best to offer the application in a setting where the user is most likely to be willing to and have the time to use the app.