Let’s see…

• Whilst pointing out a few cool things Edd mentions Planet Gnome and Monologue. Both of these are community sites formed from aggregating GNOME and Mono developers blogs;
• James Strachan announces the new hausmates feeds. Java Blogs we already know about, and similarly the blogs on Java.net;
• Simon Willison notes a move to build a PHP community site and ponders whether such as site might be similarly beneficial to Python folk;
• And then there goes Sam Ruby pointing XML USA attendees at a blogroll page

What’s the connection between all these things?

Easy it’s the people: People have blogs. People have interests. People attend conferences. If you describe these people using an extensible format including event data then much of the above can be automatically generated. There’s even some existing code in the form of Matt Biddulph’s Chumpologica.
The only piece that I believe is missing is a service that runs a Scutter across the available data and builds appropriate views.
The general idea of easy group forming isn’t new, of course.
The interesting thing to me is that most (all?) of the efforts in the list above involve a process in which the community has to become large enough to recognise it has an identity (“Java Bloggers”, “XML Conference Attendees”, “PHP Users”). Recognising it has identity it then seeks a place to congregate; somewhere to be seen. This then requires a sign-up/opt-in/barn-raising effort to build that space.
Using FOAF and other semantic web tools, that space can build itself. Right alongside the community.