Egon pointed me to an interesting blog post on RIF and OWL. The post highlights a situation very relevant for my project: The apparent existence of two main approaches in the field of "knowledge representation and reasoning" (maybe it should be called simply handling?) on the Semantic Web.
One more focused on strict (aka description logic) "knowledge representation" and "ontologies", and the other more focused on "rules" and "logic programming" (while there is still a LOT of overlapping here though!).
Of two technologies proposed for W3C, OWL (Ontology Web Language), would fall more under the first of these while RIF (Rules Interchange Format) would fall more under the other. For my project, Pellet might be more related to the first and BLIPKIT/BioProlog to the other.
As said, there are really no clear lines here though, as seen for example by the fact that BLIPKIT/BioProlog is compatible with OWL via the thea library, and even with Pellet, but it is interesting to try to figure out what is unique in these two approaches, why they are designed like that, and if they neccessarily have to be like that.
In my thinking there should in principble be possible to capture both worlds in a single framework, provided you can manage to design it generically enough. At least that's the pattern I've seen in the world of object oriented design, more specifically with the Bold/ECO framework (These guys know what generic design means!), and what becomes possible using them.