Marwane El Kharbili

Jul 7, 2008

ESWC - Part II

I know this is a very strange post here, since it had to be written one month ago now, as I was still at the ESWC conference in teneriffe. But the things being what hey already are, I have decided to still publish it.Here is the content:

"Now a week after I left Teneriffe, here is the rest of my report about the conference. Here I will tell a bit about the interesting talks I have heard, the demos I have seen and the poster sessions where I could ask deeper questions about some tools such as Nepomuk, WSMX and the WSMT.

First of all many presentations were about searches in triple stores. If one knows that the semantic language that is most widely used is still RDF, one can understand that. Ususally, the most simple things are the best accepted by people, and scientists are no exception. So RDF triples are stored in repositories and retrieving and querying information is all about searching these triple stores. So no wonder that there were several tracks precisely tackling this and other related issues. Pure semantic web is not really my area of predilection, even if I always keep an eye on the advances reached in this field. I didn't visit those talks.

What I was very much interested into however, was everything that has something to do with sws (semantic web services) in a first place, and with inference engines in a second place. For inference engines...I'll make it short, I only got one 10 minutes long presentation in the demo session of day two of the conference about some probablistic extension to an existing OWL-DL inference engine....the idea seems nice bt you ain't becoming an expert by listening to such shorties..I guess my journey deep into inference engines is going to take me some time and cost me lots of energy, since I will have to learn it alone (as always). But I would really appreciate any help guys, really :D

I attended several talks about semantic web services which essentially explain the advances made in the main tools currently existing with the WSMO framework. Those tools were also presented during the demo session that took place in the first 2 days of the conference from 19:00 till 20:00. I had the opportunity to talk to many developers on those research tools and I particularly liked the new things done in the new WSMX environment, in the WSMT tool (Web Service Modeling Toolkit) concurrent of the WSMO Studio and the service discovery functionalities of the Maestro tool. In the SUPER research project, we use the WSMO studio as a platform for implementation and also for modeling our ontologies in WSML.

I also attended talks which present new ontologies. The two that I liked most were the business process analysis ontology (by colleagues from the SUPER project) and the a software model that takes an MDA approach to the SDLC by the guys over at the SAP CEC in Dresden. I ask some questions to the host of the first talk and decided that I would read both papers. The reason for this is simple, everybody in research is by nature curious about a lot of things, but nobody has the time to cover even an infinite part of the things that sound interesting. So researchers are particularly selective about what they read. It is publish or perish but the reverse side of the medallion is "Read it in the morning if you''ll need it in the evening". I want to first have an idea about how to write a good paper to present a new ontology (and conferences are particularly selective when it comes to ontologies because there is just an explosion of them) and an unpublished ontology is not recognized by anybody in the community. The second reason is that I have very similar ideas to the ones presented in the second ontology but right now, PLM (Product Lifecycle Management) is not my focus, I am concentrating on BPM. However, the MDA ideas and concepts developed by this community are always similar to those needed in the more software oriented BPM community.

I must also say that I was very much impressed by the demos of the tools coming out of the nepomuk project, which seeks to develop a semantic desktop. the reason for this is that tese guys just decided to illustrate all their concepts by integrating many of them into the newest release of the Gnome system, a desktop layer for Linux distributions. I watched and discussed some of the tools with the guys and I think it is really the bedtest for future development software inresearch on social semantic desktops and interfaces in general. Or you guys will have to do at least as good as the nepomuk guys.

Marwane El Kharbili

No comments: