My first attempt with Arachnolingua was building the knowledge base directly in OWL using Protege. Protege is a very good ontology editor and because the OWL support in Protege 4+ is based on the OWLAPI, which I have and continue to use (one of my main reasons for continuing to use Java), it should play well with owlbuilder which uses the OWLAPI to store and generate the ontology files. Unfortunately Protege isn’t very good with doi’s. It seemed to do better with http: style doi’s (e.g. http://dx.doi.org/10…. vs. doi:10…), at least for naming the individuals for each publication (and declaring them to be publications). However, when I attempt to refer to them in an assertion that a statement about behavior is part_of a containing publication, Protege complained on loading that the URI had no recognizable prefix. Defining doi: as a prefix didn’t seem to resolve the complaint, so I went back and generated doi: prefixed URIs. It seemed to accept this, but display of the individuals within Protege as <doi:10…> suggests it isn’t entirely happy with this. I’m not sure if I have everything correctly (un)escaped in these new doi’s either. Not the highest priority in the big picture, but I’ll try to resolve it before adding more logic (taxonomy will be next).