The Semantic Web In One Day

Almost forgot to mention this: last year my collegues from the AIFB and I went to Schloss Dagstuhl, were we got divided in six teams and faced a perilious challenge: build the The Semantic Web In One Day.

"To determine just how far Semantic Web technologies have come, we wanted to create a snapshot of what you could do by applying and assembling existing Semantic Web technologies — in one day. Our experiment’s main aim was to get a feel for the practical applicability of current research by integrating different technologies into something "up and running." As an added benefit, we learned a lot about the areas in which the Semantic Web’s many research directions intersect, such as knowledge representation, natural language processing, and peer-to-peer."

Being in the In-Team was pretty cool, but having just started to comprehend the basics of the whole Semantic Web idea back then - even before this blog started - I wasn't much of a help (well, not totally true, I contributed to some pretty cool slides). But seeing the other presentations and results I was very impressed! Especially the winning team - the One - who made a the MatrixBibster-System, integrating their individual work to create something bigger - and the Team Semantic, who were my personal favourites, and having the most enjoyable presentation by offering a natural language based oracle inspired slightly by Odysseys HAL.

IEEE Intelligent Systems wrote about this as well, take a look if you like.

Abraham Bernstein on users

"The regular user is not able to cope with strict inheritance."

Abraham Bernstein of the University of Zürich was today at the AIFB and gave a talk on SimPack - A Generic Java Library for Similarity Measures in Ontologies. Not being an expert in mapping, alignment and similarity I still saw some of the interesting ideas in it, and I liked the big number of different approaches towards measuring similiarity.

Which struck me much more was the above statement, which is based on his experience with, you know, normal users, who are "not brainwashed with object-oriented paradigms". Another example he gave was his 5 years old kid being perfectly able to cope with default reasoning - the "pinguins are birds, but pinguins can't fly"- thing, and thus do not follow strict inheritance.

This was quite enlightening, and leads to many questions: if the user can't even deal with subsumption, how do we expect him to be able to deal with disjunctions, complements or inverse functional properties?

Abraham's statement is based on experience with the Process Handbook, and not just drawn from thin air. There are a lot of use cases for the Semantic Web that do *not* require the participation of the normal end user, thus there still lie plenty of possibilities for great research. But I still believe that the normal end user has to unlock the Semantic Web in order to really make the whole idea lift off and fly. But in order to achieve that we need to tear down the wall that Abraham describes here.

Any ideas how to do this?

Live from ICAIL

"Your work remindes me a lot of abduction, but I can't find you mention it in the paper..."

"Well, it's actually in the title."

ESWC2005

The ESWC2005 is over and there have been a lot of interesting stuff. Check the proceedings! There were some really nice idea, like RDFSculpt, good work like temporal RDF (Best Paper Award), the OWL-Eu extensions, naturally the Karlsruhe stuff like ontology evolution, many many persons to meet, get to know, many chats, and quite some ideas. Blogging from here is quite a mess, the uplouad rate is catastrophal, so I will keep this short, but I certainly hope to pick up on some of the talks and highlight the more interesting ideas (well, interesting to me, at least). Stay tuned! ;)