Unlike the report of last year’s OWLED and DL, this one about OWLED’08 in Karlsruhe (co-located with ISWC 2008) will be a bit shorter, because few papers were online before the workshop so there was little to prepare for it properly and during the first day the internet access was limited. However, all OWLED’08 papers are online now here, freely available to read at your leisure.
There were two user experiences sessions, one on OWL tools, one on reasoners (including our paper on OBDA for a web portal using the Protégé plugin and the DIG-QuOnto reasoner), one on quantities and infrastructure, and one on OWL extensions. In addition, there was a ‘break out session’ and a panel discussion.
The notion of extensions for “database-esque features” featured prominently among the wide range of topics; for instance, the so-called easy keys  that will be added to the OWL 2 spec and dealing with integrity constraints in the light of instance data [2,3]. An orthogonal issue—well, an official Semantic Web challenge—is the “billion triples dataset”, which is alike a very scaled up LUBM benchmark. Other application-motivated issues came afore were about modeling and reasoning with data types, and in particular quantities and units of measurements, probabilistic information, and calculations.
The presentation of the paper about probabilistic modelling with P-  generated quite a bit of discussion, both during the question session as well as the social dinner. The last word has not been said about it yet, perhaps also because of the example with the breast cancer ontology and representing relative risk factors and probabilities for developing cancer in general and that in conjunction with instance data. The paper’s intention is to give a non-technical user-oriented introduction to start experimenting with it.
The 1.5hr panel discussion focused on the question: What are the criteria for determining when OWL has succeeded? Or, with a negative viewpoint: what does/can/could make it to fail? Given the diversity of the panelists and attendees, and of the differences in goals, there were many opinions voiced on both questions. What is yours?
 Bijan Parsia, Ulrike Sattler, and Thomas Schneider. Easy Keys for OWL. Proc. of the Fifth OWL: Experiences and Directions 2008 (OWLED’08), 26-27 October 2008, Karlsruhe, Germany.
 Evren Sirin, Michael Smith and Evan Wallace. Opening, Closing Worlds – On Integrity Constraints. Proc. of the Fifth OWL: Experiences and Directions 2008 (OWLED’08), 26-27 October 2008, Karlsruhe, Germany.
 Jiao Tao, Li Ding, Jie Bao and Deborah McGuinness. Characterizing and Detecting Integrity Issues in OWL Instance Data. Proc. of the Fifth OWL: Experiences and Directions 2008 (OWLED’08), 26-27 October 2008, Karlsruhe, Germany.
 Pavel Klinov and Bijan Parsia. Probabilistic Modeling and OWL: A User Oriented Introduction to P-SHIQ(D). Proc. of the Fifth OWL: Experiences and Directions 2008 (OWLED’08), 26-27 October 2008, Karlsruhe, Germany.