Joris van Hoboken

This week and th enest two weeks I am attending the Summer Doctoral Programme of oii and Berkman Center at Harvard University in Boston. This afternoon Michael Zimmer gave a presentation about his dissertation I was especially looking forward to. It was called “the quest for the perfect search engine”.

Central in Michael Zimmer’s work is the trade-off between search engine user’s privacy and perfect recall. The idea is that in order to serve us better as tools for finding information search engines need to understand a lot about our backgrounds and interests, so that they can better interpret the particular search query the user enters. Examples of problems relating to queries would be apple (the computer, the type of fruit or the city?) and Paris Hilton (hotel or someone that seems to get a lot of attention in this country). Future examples are us asking the search engine where to go on holiday, what to vote(why vote at all, they could do that for us) and whether or not to like a certain movie.

I thought his presentation was great. It also made me think about my own research a lot. One of the more problematic points in his presentation (I think he probably goes into that in his dissertation) was that search engines might not have the incentive to become perfect in the sense of perfect recall. The reason for that might be that search engines should be seen more as advertisement companies than as information location tools. Thus data aggregation would not be so much a tradeoff with perfect search but instead with perfect advertisement. This makes the ‘faustian tradeoff’ quite different. (We are not getting perfect search with advertisent, but fairly good search with perfect advertisement).

Surely of equal value to me was the way Jonathan Zittrain and later Urs Gasser commented on the presentation and explained what could have made the presentation better and what could make a presentation perfect. I am going to see at least 15 other student presentations and I am sure this type of interaction and feed-back from our tutors will help me to make some giant leaps in understanding what it is to do a good presentation. Such an understanding is definitely something that is missing in most of Dutch academia, so I have a lot to learn. I am scheduled in the end of next week so I can already test some of the ideas I get.

This entry was posted on Tuesday, July 17th, 2007 at 10:40 pm and is filed under commercial communication, literature, personalization, privacy, sdp2007, search engines, user data.