Use dct:source to indicate relationship with toco_ontology #5
Description
Ok one more because now that it's in turtle it's easier for me to read your ontology:
Try using something like https://www.dublincore.org/specifications/dublin-core/dcmi-terms/#http://purl.org/dc/terms/source to link between your ontology terms and the ones you re-used from toco. I see definitions are copied. Wanting to re-use a term from another term leaves you with three options generally:
- Reference their IRI in your ontology, either by "taking ownership" of that IRI (you define their IRI using their defintion in your own ontology). The risk is that they change their definition, and you will not be "aligned" with the parent ontology when a user visits the IRI you re-used and dereferences it to find different information from what you expect. (e.g. you reuse the property "hasName" which is defined as "full name" in your ontology/front-end, but then if a user were to browse to "hasName" of the source ontology, may find that it has been specified down to "just last name".
- Create your own IRI under your namespace, and refer to the orginal IRI using dct:source or something (prov-o probably has a similar term). Added benefit is that you can further specify the definition to match your specific use-case. Alternatively, subclass it (though you will keep a bit of the problem from point 1 in such a case).
- Do as you did, and copy and paste the term from the parent ontology, without providing any relations to it. This is the least preferred option in my opinion, as it essentially negates any benefit of using RDF - the connectivity aspect. If instance data exists according to your ontology and according to toco ontology, you would not be able to unify them easily, as neither refers to the other. Whoever tries to use this data in the same context would have to create a mapping first. Imagine everyone were to create entire new ontologies containing everytime the same information just under different namespaces - that would be a bit of a waste.
Just some thoughts. I'm curious about what instance data looks like according to your ontology. Do you have any real life applications of this ontology yet that I can check out?