photo by Ed Schipul
The first keynote on Thursday night was by Marc Smith of Microsoft Research, who shared some of his insights based on the analysis of newsgroups. What I found most striking was the fact that only 2% of that community supplies all the answers (what he calls the "answer persons") and 66% only posts once. So a very small group of people actually defines such a community.
The keynote on Friday was a hilarious presentation by Rob Malda and Jeff Bates, the founders of Slashdot. The quote I take away from their story is: give people a number, and they will try to maximize it. In their case, this relates to a reputation score, but this principle was a recurring theme in many of the other presentations: a lot of things can (and will) be treated as a game.
Another thread that ran through the conference was that of social network analysis. One impressive visualization followed the other, often involving enormous data sets. A much needed word of caution was presented by Marleen Huysman and her colleagues. We have left a period behind us where we thought we could store and transfer knowledge by means of information technology. The danger is that we will now use that same instrumental approach for social networks: managers trying to create and optimize networks for knowledge sharing that are not properly embedded in the work practices.
The keynote this morning was by Judith Donath of MIT Media Lab, who gave an interesting talk about signaling. In biology, signals are a way to indicate qualities that are not directly observable. However, humans are able to manipulate these signals. I can rent an expensive sports car for a day to signal that I am rich, even though in reality I am not. In a computer-mediated context, this becomes an even bigger issue. I can create a very desirable avatar in Second Life that has very little to do with what I really look like. One of the keys here seems to be cost: an avatar in Second Life takes maybe fifteen minutes to create, a level-70 character in World of Warcraft takes months. The latter signal is therefore much more reliable.
The Communities & Technologies conference ended today, so it is time to share some of my personal highlights with you. It was certainly a very worthwhile conference with Charles (Chip) Steinfeld (pictured above) as an excellent host.
The first keynote on Thursday night was by Marc Smith of Microsoft Research, who shared some of his insights based on the analysis of newsgroups. What I found most striking was the fact that only 2% of that community supplies all the answers (what he calls the "answer persons") and 66% only posts once. So a very small group of people actually defines such a community.
The keynote on Friday was a hilarious presentation by Rob Malda and Jeff Bates, the founders of Slashdot. The quote I take away from their story is: give people a number, and they will try to maximize it. In their case, this relates to a reputation score, but this principle was a recurring theme in many of the other presentations: a lot of things can (and will) be treated as a game.
Another thread that ran through the conference was that of social network analysis. One impressive visualization followed the other, often involving enormous data sets. A much needed word of caution was presented by Marleen Huysman and her colleagues. We have left a period behind us where we thought we could store and transfer knowledge by means of information technology. The danger is that we will now use that same instrumental approach for social networks: managers trying to create and optimize networks for knowledge sharing that are not properly embedded in the work practices.
The keynote this morning was by Judith Donath of MIT Media Lab, who gave an interesting talk about signaling. In biology, signals are a way to indicate qualities that are not directly observable. However, humans are able to manipulate these signals. I can rent an expensive sports car for a day to signal that I am rich, even though in reality I am not. In a computer-mediated context, this becomes an even bigger issue. I can create a very desirable avatar in Second Life that has very little to do with what I really look like. One of the keys here seems to be cost: an avatar in Second Life takes maybe fifteen minutes to create, a level-70 character in World of Warcraft takes months. The latter signal is therefore much more reliable.
No comments:
Post a Comment