Last month I blogged about Snowl, an experiment with messaging in the browser.  Two weeks ago I presented it at the Firefox+ Summit.  And last Wednesday I released an initial prototype.

I’ve been getting tons of feedback since then from a variety of sources, including discussion at the summit, comments on the announcement, AMO reviews, discussion in the forum, news and commentary (TechCrunch, ReadWriteWeb, ZDNet, Computerworld, Ars Technica, Slashdot, Digg, etc.), blog posts, and bug reports.

Some opinions I’ve seen expressed more than once:

• the river view feels like it fits into Firefox
• the list view doesn’t mix well with general web browsing
• “I’d rather use a standalone application.”
• “I’d like even more integration with Firefox.”
• many of the prototype’s features should be part of Thunderbird
• it lacks a number of useful features (f.e. sharing)
• it contains a number of bugs
• it has lots of potential

Overall, there’s a lot of interest in Snowl, in browser/messaging integration, and in messaging innovation generally.  There’s plenty more to explore here, and I’d like to see Snowl experiment with the following features in particular:

• simple web search engine-like attribute-based search
• a message view optimized for short, simple messages (Twitter, IM, SMS)
• an experimental message view that uses color, size, graphics, word clouds, animation, or other techniques to help users make sense of their flood of messages
• support for sending messages, including responding to tweets and posting comments to blogs
• easy access to feeds from frecent (frequent+recent) sites, sites you are currently browsing, live bookmarks, and feed items you’ve commented on
• conversations (i.e. threads) as first-class objects
• better support for people as first-class objects
• email, IM, and other message sources
• user-initiated and automated message grouping (Home/Work, Important/Unimportant, Today/Yesterday/Last Week, etc.)

Of those, I’ll be focusing on a short message view, message sending, and conversations as first-class objects for the next iteration.

If you have a burning desire to see something else on this list (or something else entirely) happen sooner than later, your contributions are welcome!  Discuss ideas and post mockups to the discussion forum, pick up the source code from the code repository, and submit patches in bug reports.