WWW2006 - Experiences and Episodes: Day 3(Through Taowei’s Looking Glass)
by Taowei Wang
Today’s opening panel had Tim, Jim, and two industry representatives. The focus is the next wave of the semantic web. Questions for the panelists range from the perspectives of the last few years to looking ahead of the future. The surprising thing of importance, according to Jim, is the speed of growth of the semantic web. As such, the scalability of the tools for the semantic web is an important challenge. Some other questions have concern with “Web 2.0″, which, depending on who is talking, mean a little different things. The folksonomy community (flickr, del.i.cious, etc.) mean the world of (collaborative) tagging. On the other hand, the web application people mean Ajax. And there is a third group, perhaps overlapping witht the previous two, that say today’s Web is about commerce, while Web 2.0 is about community… social Web. From these trends, we can see that there is a movment to provide a richer, more interactive personal experience (ajax) and perhaps seemless interaction with desktop applications, and to provide a medium for interpersonal, or community-based collaboration/interaction. Along the lines of cell-phone enabled games, how about cell-phone-enabled online-communities — that translates to real communities? Imagine users use cellp-phone to log in to read their favorite blog, and finds an interesting restaurant, cell-phone shows the map where it is. Once the user walks into the restaurant, web online community finds that the user and one of his online-friends (or collaborators, or someone who has shared lots of online discussions, or same professional interest), and both persons’ cell-phone tell them so… maybe even automatically exchange (digital) business cards. Can we use mobile devices to automatically mine information around us? Let it be the ‘6th’ sense (move aside, esp!), that collects information around us in a world where information does not just sit staticall on the Web, but move around with people. Wouldn’t that be cool? How would SW technology help the exchange of information? I have more questions than answers. There was an ontology session of refereeed papers.
The first paper by Peter Patel-Schneider deals with Datalog and classical logical systems, and attempts to reconcile to increase the utility of OWL. Fist, he compared/contrasted the two systems — unique name assumption, open/closed world assumption, restriction/integrity constraints (domain, ranges). Then he demonstrates that OWL is mostly a classical system (which isn’t surprising), however, there is utility for some of the Datalog properties. He proposes “local UNA, local CWA, and local autoepistemic properties being part of OWL for those who wish use the logic this way. Interesting proposal. Not sure how it works with imports, though. Nicely presented. The second paper was also well presented. This paper, though, needs it’s own section title:
“Segmentation Fault”
The presenter presents a way to ‘carve’ up a slice from a large ontology. The technique uses a traversal of a selected concept from the told class hierarchy, added recursive walks up and possibly across links, and added limits to the depth of the recursion and property filters. They call the process ’segmentation’. The result is a smaller ontology, with (unknown amount of) information/semantics lost. No inferences from the original ontology can be guaranteed from the new one. The only utility is to have a smaller ontology, but isn’t the point of an ontology to have certain well-defined semantics? In that sense, traversal segmentation is no better than random-chopping :/ Core dumped.
Yahoo! paper on Flickr tag vis was very cool. Visually addictive, and algorithmically stimulating. Hmmm… gonna go read the paper in detail and savor it! Saw a demo by some Nokia people.
They had successfully created a tiniest web browser for their s60 cell-phones. The browser displays web pages like a browswer on computers, complete with a cool history view and an overview view. Neat.
