ISWC 2006

November 8th, 2006 3:13 PM

The International Semantic Web Conference is currently taking place here in Athens, Georgia, and it’s been really great so far.

Photos are online from the two days of workshops and dinners and the first day and poster session.

Chris discussing his poster

Chris and I rolled out a last-minute version of PhotoStuff for the conference which preloads the conference ontologies and a bit of the related FOAF data, and makes use of the ESWC PhotoStuff backend. There were some big differences at ISWC with the conference RDF compared to what we found at ESWC earlier this year, and I think there were some important things learned about producing useful conference RDF data.

I had a great talk with Norman Gray about the current state of the PhotoStuff code base, and the potential for modularizing the code to make it more flexible and extensible. Leo Sauermann asked if I’d be interested in getting PhotoStuff to work with his Gnowsis Semantic Desktop project (which I think would be great).

The Policy workshop on Sunday seemed all over the map, with several of the presenters having conflicting views of what “policy” meant, leading to a somewhat variable set of talks. I did find the invited talk by Grit Denker very interesting. She presented a system of spectrum-agile radios that use an inference engine to work within appropriate broadcasting policies.

Taowei and Chris eating wings

The User Interaction workshop on Monday seemed much better, including: TimBL giving a presentation on the Tabulator, a panel by the Web Science people, Tom giving a good talk on what was learned from the use of semantic technologies at ESWC. On a more humorous note, David Karger had some interesting things to say about the important part of the Semantic Web layer cake: his version compared to TimBL’s version (photos by Paul Shabajee).

Some of the other highlights from the talks so far:

  • Jorge A. Perez gave an excellent talk on the Semantics and Complexity of SPARQL (nominated for the best paper award). I found this particularly interesting having implemented a SPARQL engine, and having run into some of the same issues of formal semantics versus query performance.
  • Eyal’s presentation on a formalized model of faceted browsing, and approaches to automatically optimizing facet selection.
  • The Fresnel presentation on browser-independant RDF presentation rules. (Especially interesting was TimBL’s question regarding cascading Fresnel lenses like CSS stylesheets.)