I had a bizarre dream last night. I'm sitting with some friends back in NYC at a bar and I started telling them excitedly about a new feature that Google had introduced to Gmail. This was the gist of it:
Gmail had added a link to the top of the page called "Friendsterizations". When I had clicked on that link (it's all fuzzy because, after all, it was just a dream) there was a way in which you could tie your Gmail account to your Friendster account. This opened up a whole new range of search possibilities. You could easily identify all emails from your friends. You could easily extend that group of emails to include friends of friends...
To be honest, I'm not sure how much really appeared in the dream. But now that I'm thinking about it, you could have a very robust mailing-list or bulletin board functionality if a social network was integrated with an email application. (Think of Friendster's message or bulletin board system -- now think of it working instead of breaking half of the time.)
And then there was the potential for spam reduction. You could set it such that you only could receive emails from your first, second, or third degree friends. And to prevent spammers from adding fake friends to the network, friends could be voted off of the network by their peers.
Of course, Google would use Orkut instead of Friendster due to the Google/Orkut affiliation. (As an aside, if Orkut is still online and you're a member, please send me an invite as I'd like to check it out.) Or Google could simply buy Friendster, which has a huge user-base, but very mediocre technology. Friendster's move to PHP seems to have helped, but I'm suspecting this is because it was a second-generation rewrite -- there's absolutely no reason you couldn't write a scalable social networking app in Java. My guess is that they just switched to PHP to get a clean slate.
If you could get a critical mass of people in a social network/email application I would definitely use it. In fact, you don't even need that much of a profile per use -- simply a friend-of-a-friend network of email addresses would offer a tremendous amount of value. I'd write something like that myself, but the email infrastructure and webapp are prohibitively hard to build and scale as well as Google did with Gmail.
[Update: Couldn't Gmail do this in a matter of hours just by adding a flag (similar to the "star") that identified a contact as a "friend" in the address book?]