Yesterday was the first full day of the AAAI Symposium on Computational Approaches to Analyzing Weblogs. The room was pretty much packed (I'm guessing somewhere between 40 and 60 people, which is not bad for this type of affair). Talks ranged from observational/field studies of blog data (e.g. describing the distribution of ages, posting times, geography and the correlations between them), sentiment analysis (for politics in English and products in Chinese), models of attention, a comparative study of where news breaks (in blogs or in MSM?) and some interesting discussion on the application of multilingual text analytics to blog data.
In addition, there was a panel including Chris Redlits (Feedster), Michael Sippey (Six Apart), Howard Kaushansky (Umbria), Tony Perkins (AlwaysOn), Andrew Bernstein (Cymfony), Cameron Marlow (Yahoo) and Carrie Grimes (Google). Mark Liberman (of Language Log) moderated the panel as they discussed the analysis of blogs and some forward thinking in the space. Panels are always a tricky business as they can appear like talk shows (with scripted questions and answers) or they can lack cohesion. This group was diverse enough to provide some interesting discussion.
Of particular note was the discussion on privacy and the idea that social media will separate into the public space (where content creators use the web to create visibility) and private space (where the web is used as a mechanism to host and archive data, but where access to that data is restricted to specific users). Perkins of AlwaysOn believes that what he calls the IM generation doesn't care about privacy (I disagree with this - I think they don't have a good model of it). Sippey of SixApart believes strongly that the issue is going to result in different tools sets to interact with and manage data in the two different spaces.
I have to agree with you on the IM-generation & privacy issue. As a member of the IM generation (at least, at the leading edge of it), I think many young people on the web want privacy (at least from people outside their affinity groups), but assume that obscurity will be sufficient. Most don't realize how findable they are.
Posted by: ryan king | March 28, 2006 at 06:22 PM