[Matthijs] Decker says: "A domain model, defining objects and relationships of a domain of interest can be represented naturally in RDF, so translation steps, as required when using XML, are not necessary. [..] Of course, this does not solve the finding the general operability problem, that is, finding semantic-preserving mappings between objects." Is RDF a significant help in solving this problem? If not, isn't RDF overrated?
[Matthijs] Does the RDF structure have advantages above a class with properties as we know it from object oriented programming?
[Matthijs] Doesn't cause multiple inheritance problems when a class extends two classes with conflicting properties?
[Gwyneth] Could we model the real world in OWL? What are the restrictions? Is that a technical issue or a philosophical issue?
[Gwyneth] We want to define as much as possible for the semantic web. To do so we need to use OWL full, because this is the most powerful. It seems therefore also hard to reason in OWL full. Is an ontology defined in OWL full useful?
[Wilco] Let's have last week's discussion in mind when discussing this week's technique decriptions. OWL is a highly structured we to model content. How is that compared to folksonomies? Are we heading towards a semantic web for users or for machines?
[Wilco] For interaction between programs, structure is needed. this causes the need for a single standard. In this light we can see the developing of RDF, OIL and OWL as needed on the road to such a standard. Is there a counterargument against this reasoning? --> [Thomas] XML offers enough structure for programs to communicate, it's just that for XML you need to decide on a specific XML schema, dtd etc. for enabling automatic communication. Do OWL, RDF etc offer enough structure on the knowledge level to overcome the problem of having to build a common XML schema used by all programs (in a given domain)?
[Thomas] It's only possible to extend an existing OWL ontology, but shouldn't it be possible to remove existing concepts as well? By only allowing extensions OWL has basically become monotomic. (all statements true in a subset are also true in the superset). While nonmonotomicity is more natural in the Semantic Web context...
[Thomas] Wouldn't it be better to have the unique name assumption in OWL?
[Arno] The highly decentralized philosophy of RDF is a major advantage. Or is it a weakness?
[Arno] OWL will be used along with RDF. OWL is too complex for some people as compared to RDF, which will be used for more simple modeling and peer-to-peer sharing of information between individuals or small groups.
[Bart-Jan] RDF is a primitive modelling language.
[Bart-Jan] The author says that machines can not derive semantics from RDF. I agree with that because the machine will never ‘really’ understand what it is talking about. However the author also states that ontologies do capture the required semantics. I agree that there is more semantics in an ontology but still the machine will not ‘really’ understand it.
[Bart-Jan] The internet is a medium for the masses and a medium by the masses. Although languages like RDF and OWL are important steps towards the Semantic Web, the technologies are not yet mature and user-friendly enough to empower the less skilled web designers.
[Bart-Jan] The benefits of adding metadata to information on the web is not yet clear enough for web designers. There should be more short-term benefits for innovative web designers, which will lead to the realization of the Semantic Web vision on the long-term.
[Melanie] Which version of OWL should we use? OWL Full seems the most powerful, so should we accept its undecidability? Or should we use a less powerfull version that is decidable?