Opera at SXSW 2008

Conleth O'Connell of Vignette Corporation began the conversation by presenting the problem of information overload. Now that iPods can hold 80g of music, the question of how users can find what they're looking for becomes a serious one, one that the iPod interface is particularly adept at, especially in cases like driving. At some point, he argues, mobile search will have to morph so that we are not entering a query and searching through results, but so we will have a search that is connected to our personal interests and behavior and the contexts in which we work. In other words, what is important in the mobile space is not just what you are saying, but what is happening when you are saying it.

Len Hause supported this assertion, arguing that moving through different contexts—for instance, listening to an iPod while driving—has illustrated the limits of text's usability. He went on to claim that one way in which the mobile space is changing is that we are moving from a content-based system to a system based on contexts or what he calls "personas." One example he gives is business cards. It used to be the case that on a business card everything in our home and work life was assigned a category based on network access—we had separate home and work phone numbers and addresses. Now we have a persona model, where the network connections are being effaced and we are providing a connection through a single persona—a cell phone or email, for example. Similar to Tim Berners-Lee's graph, the network is being hidden to emphasize our connections between each other and the environment. According to Hause, it is important to bring this persona model to mobile search.

One final point of interest: after Mike Svatek of Baynote introduced the idea of metadata, Hause argued that textual metadata will eventually be replaced by an object-oriented model where a persona object—consisting of a photograph and a name, for instance—would become metadata which could be used across platforms for multimedia search like facial recognition.

This was another good panel; when I originally read the title, I was expecting the discussion to be more concerned with the details of UI construction than with search theory, but since some of my research has focused on search and the relationship between metadata and data, I still found the ideas presented by the panelists quite interesting.