In a recent posting titled, "The two ways of web 2.0", Lorcan Dempsey explores an interesting view of the "2.0" discussion, introducing the notions of concentration and diffusion.
Dempsey applies the term diffusion as a label for the communication, social networking, data syndication aspects of Web 2.0. Whereas concentration is essentially the opposite: harvesting, combining and reusing data that has been "diffused" out onto the web. The two aspects are obviously complementary and, in truth, like much of Web 2.0 aren't new. As techniques for sharing information these are well-trodden paths. Think "Broadcast" and "Aggregation". But many concepts get a new lease of life when combined with the great levels of interactivity and socialization that the web now offers.
I've made several attempts myself to tease apart Web 2.0 into more manageable chunks. Most recently in a paper in Serials called "The Threads of Web 2.0" in which I tried to decompose the concept into several buzzword free trends. Speaking to the same notions of data flow, albeit with a slightly more technical angle, I've also explored the ideas of Streams, Pools and Reservoirs as a model for data publication and aggregation on the web.
I've seen a few discussions lately about whether there is a continuing role for aggregators on the web in these days of near ubiquitous search. I think Dempsey's notion of concentration addresses that point directly: there is a definitely a role for aggregators, but that role is changing from one of simply compiling large volumes of material, towards compilation of relevant subject-specific collections for specific communities. This is where in my posting I differentiated between Pools (simple aggregations) and Reservoirs (pools that support a community).
Dempsey observes that librarians need to begin thinking about concentrations of data and how this might benefit their mission. I think publishers would do well to do the same. It strikes me that societies and other member organizations are particularly well suited to creating and driving these new aggregation models.