I have just attended a seminar on ‘text technologies and the semantic web’ by Ernesto De Luca. Students were to present a topic and write a short paper on it. It was very insightful and inspiring. As i had understood, the key for a meaningful integration of information across many sources would be an ontology-based approach. I have written a short seminar paper (PDF) about how ontologies might help to overcome semantic heterogeneity across many information sources. It seemed quite right. The computer should be able to understand the semantics of the sources, so that logical queries such as “Who shot JFK?” could be answered by processing source ontologies using the ‘vocabulary’ of the shared ontology. Such a system would be really smart.

But at the end of the presentations covering ontology-based approaches (including mine), a question remained: Who would put such a great effort in creating the shared ontology and especially the local ontologies? I thought that Wikipedia is the best proof that this can only happen voluntarily on a non-profit basis. Yet, WP is not yet semantic. But it could well be extended progressively. After some research, i have found somesimiliar thoughts on WP. This will probably take quite a while until WP is able to accept logical queries — yet, the first steps are in the making. But then, existing sources such as external websites would not be integrated: the articles of WP have to get modified or written anew.

For the integration of many resources (today), there are tag-based social databases (read: folksonomies) such as del.icio.us and flickr. They proof that many normal people are better in organizing and categorizing sources than a few professionals. And it is cheaper, too. Del.icio.us seems to be in the focus of research in that regard. There is a well-written article arguing that the approach of using an “Ontology is Overrated [while arguing in favor of]: Categories, Links, and Tags”. The author shows on one hand that the rigid classification schemes found in libraries are not appropriate anymore and on the other hand that those systems that really work are not run by professional ontologists but by regular people (read: us) putting tags on resources and creating a vast folksonomy.

I think ontology-based infrastructures could be much more superior to tag-based systems, if only many people would be involved in managing such a system. It seems that ontologies will remain an academic topic, until we have found ways for including this social dimension for crafting, updating, and using ontologies for information sharing. Right now tags and links will provide the best bang for the buck — to put it into economists terms. Still, I am very excited about how semantics could be created and used by many in a folksonomist way in the future. Any thoughts on this?

This entry was posted on Wednesday, July 20th, 2005 at 5:54 pm and is filed under english, technology. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.
Statistcally related posts: Semanticizing Wikipedia This semester will be a good one Wrapping up WWW2007