I kicked my SongBird installation back into gear today. I have a few takeaways:

1. The Add-Ons community is buzzing. SongBird itself has a few interesting features, but the support through Add-Ons for Wikipedia, Hype Machine, iPod, and Scrobbling (and other cool hacks) is excellent. I’m impressed that this young project has already generated such a rich community of Add-Ons. I suspect the fact that they used the Mozilla Add-On implementation (XUL) was helpful because developers didn’t have to learn some new proprietary system in order to start scratching their own itches.
2. Maybe this is already obvious to people, but the internet is quickly becoming just a pile of data. It is the “Model” in the MVC architecture. SongBird is a new “View” on the Model. Instead of viewing websites as rendered by Mozilla Gecko in Firefox, SongBird has additional View modes such as Playlist View (which is a list of all MP3 linked from the rendered page) and Download View (where you can grab songs off a site to keep). This smells like the promise of the Semantic Web. The web is a database (Model), and you can unlock new value from an existing Model by designing new Views of the Model.
3. SongBird’s UI make me think that my own music library should be the least common listening experience. 90% of the left sidebar is used by bookmarks for various blogs, music stores, sites, and search engines. The remaining 10% is associated with my local music (iPod, local library, download history). That’s a radical shift from every other stand-alone music client, where 90% of the experience and UI revolves around managing your local files. SongBird wants to sing songs from The Cloud, and I dig it. It’s really expanding my listening habits.

Looking forward to watching SongBird develop.