Before I embark on a multi-day exploration of best ways to address this, I thought I’d ask what is possibly a FAQ (but not popping up on google).
The scenario: I have a TDB store containing statements that reference concepts from a variety of ontologies, including some that are in flux. I have been initializing my TDB instance by invoking methods to load all of the needed ontologies, and then using the store with the ontologies in our experiments. At present most of the data are expendable - I don’t mind purging the contents and starting afresh. But when the data are important and need to be retained, I’m wondering how best to 1) incorporate the ontologies, and 2) update the ontologies as the need arises. I’m aware of the broader problem with alignment of the data as the ontologies change, and for now choose to be an ostrich about that. My questions: 1) is there a practical difference between incorporating the ontologies via the Fuseki config and just loading the ontologies from within my code? 2) Does the config approach mean that the ontologies are referenced in at each Fuseki server startup or merely loaded as content to the store? Do changes in the ontologies between server restarts have any effect? 3) Is there a way to replace a given ontology once it’s been loaded into the store? Or are the new/revised elements pulled in alongside of the existing statements? 4) What else should I know about this? Thanks, Mark
