Social theory is making many of our intuitions concrete. We've long had hypotheses about 'tribe size', but Roland Dunbar has given us some good reasons to believe that humans work best in groups of 150 (a finding briefly glossed here.) We've always known that "It's not what you know, its who you know", but recent books by Duncan Watts (Six Degrees) and Albert-Laszlo Barabasi (Linked) give us mathematical ways of understanding how and why.

Now a conference in Helsinki this fall is asking: How does (or should) this work on social network analysis affect the way we build software? and in particular, software for computer-supported cooperative work? The conference is in September, so there's not much information up yet, but the Call for Papers is up online.