I've learned a lot from developing garbagescout.
Since I first let my friends start using it 3 months ago, the main comment I've received is "nice idea, but why is it only in New York?"
Because my goal was to overcome the burnout that I usually get on complex projects, and just get this one done, I coded the app in a very sloppy way. Basically, New York City was hard-coded in to the three major components of the application: the email parser, the address-to-latitude-and-longitude converter (geocoder), and the web front-end. I spent a lot of time generalizing the code so that it could be used anywhere.
I'm glad I did that, because the new code is much better, but the results weren't a complete success. The lesson here is to create a site that works well with a very small contributor to user ratio. This is valid advice for all sorts of sites that rely on user-generated content.
I should've known, because WebJay, a site I was active in the creation of, has a ratio of about 100 consumers to one producer. Garbagescout has a serious technological barrier to use: you need a camera phone, and you need to know how to use it.
Webjay works well for consumers who are not producers. Even if there are only a few playlists there, that's several hours of listening.
Another lesson, related to the first, is that you'll have to rely on your friends or a few enthusiasts to get the site going. In this vein, I wonder about the beginnings of Wikipedia. It started in January of 2001. By September 2002, it had 48152 articles. What happened inbetween? Wikipedia satisfied an urge in the producer. The producers of content became a sub-community of the larger Wikipedia user community.
Obvious, maybe, but important for me to remember as I think about my next project.
Update: More notes on garbagescout.
on developing garbagescout is filed under programming (8) users (3) garbagescout (5) garbage (7) .