opening this group

Dear all,

Thank you for your interest in this community group. The topic for this group is: SKOS and OWL for interoperability. This topic was discussed during the #SDSVoc bar camp in Amsterdam last December, and we felt that we needed more time and more opinions to come up with best practices and design patterns for using SKOS versus OWL.

Some background: in the wild there are many examples of SKOS implementations. People seem to find it a useful, easy standard to “capture things”. On the other hand, people are very creative in extending the basic SKOS scheme, and by doing so move towards a more OWLy version of the original SKOS scheme. (Some examples here.)

Statement: To support interoperability we should agree on using standard design patterns and move away from the tempation to make the perfect semantic model.

This would perhaps imply that it is best to use the SKOS scheme in a clean way and not play with it too much, otherwise the user of your data has to “do something extra”, such as run a reasoner before (s)he can use. And if you want to add complexity, maybe just create an OWL class, and not subclass from skos:Concept.

What is your opinion?

Lieke