Taking the Semantic Web for a Walk

One of the joys of working in a shared office containing half a dozen people who all work part-time on different projects and on unpredictable schedules is never knowing when a chat will be productive and inspiring.

Last week I had found a paper and website about a project called DBpedia Mobile .
In this app you can walk round a city and on your mobile phone you can see a map which is centred on you and full of little flags which point out interesting places near you. Touch one of the flags and you’ll see more information about that place…. and then you can browse the web to find any other connected information. If one of the flags indicates that you are close to the birthplace of a writer you might navigate as far as Project Guthenberg and read the author’s works. DBpedia Mobile doesn’t just use the DBpedia information that’s derived from Wikipedia. It relies on Linked Open Data, the ever-increasing web of publicly available information which is linked in a meaningful way using Semantic Web ontologies, and it uses semantic search to pick out the right information. So this an application which combines a physical exploration of a city with a virtual exploration of the semantic web.

There’s a resemblance between moving round a city and navigating the World Wide Web. DBPedia mobile organises information around physical location, the place where you are now. But physical location is only one way to organise connections. And just as it’s interesting to tour a city, it could be interesting to tour the semantic web. Physical location doesn’t have to matter. I (as a user) would supply some minimal information about my interests and the system would first search to find web pages relevant to my interests and then follow links from those pages, possibly using my stated interests to direct which links to follow, possibly not. Wesbites I encountered in this way might be more or less interesting, more or less related to my interests, more or less serendipitous.

I was discussing some of these ideas with my colleague and occasional office-mate Alison Williams (of the SPIRES project, and whose own research codifies the impact of physical space on creativity) and she starting telling me about Debord’s Theory of La Derive (the drift) :

“In a dérive one or more persons during a certain period drop their usual motives for movement and action, their relations, their work and leisure activities, and let themselves be drawn by the attractions of the terrain and the encounters they find there….”

Even in a large city each resident may only ever visit a very few places, and the locations that different residents use may not overlap. Locations can be physically close to each other but remote because people don’t move between them. Psychogeography offers a way of understanding what constrains the way people move through cities and which areas they visit, while La Derive is a way of exploring cities that crosses over the usual boundaries, overcoming some of those constraints to make new discoveries and explore parts of the city that you normally wouldn’t.

She described an activity she had undertaken as a student where she and a group of students navigated the city using “drift”. The students shared a start point and end point, but walked separately. They each had to invent a rule for choosing where to go next and a way to record where they had been. So one student decided to follow the colour yellow; another decided to go from cafe to cafe, and photographed menus; another wrote a song describing the places she visited. The important thing is to invent and follow a rule of one’s own; not to follow habits or other peoples’ conventions.

Search engines and recommender systems offer us pre-set paths through the World Wide Web, linking us to the resources we want. Each one tries in different ways to offer us the shortest path to the best resources. The rules they use are predetermined and opaque to us when we use them. There are places on the web we visit often and places where we never go, as well as places where we might go if we’d thought of them. The semantic web links all kinds of web resources through a network of meaningful labels, and we might choose to walk that network in all sorts of ways. Using semantic search we could create rules for walking the network that keep us in familiar places close to our own current interests, or we could create rules that break down the barriers and explore parts of the web where we might not think to go, but where we might make interesting discoveries.

Diana Bental