Open danbri opened 8 years ago
This only really makes sense after we get the ability to write schemas in JSON-LD. see also schemaorg/schemaorg#178 (do we have an issue for actually using rdflib to permit json-ld schemas? can't find one...).
I don't like the Role-based design sketched here, but this is a common situation. Parties using schema.org are naturally inclined towards re-use, but the more they re-use rather than create new terms, the less they're provided with a surface for documenting and defining. There is some connection to the notion of shapes here - see https://github.com/schemaorg/schemaorg/issues/1715
Tagging this for continued discussion in the suggestions-questions-brainstorming repo
See issue #7 for the context of the move from the main Schema.org issue tracker to this repository.
We very often have the following workflow:
Generally re-use is good, but it has meant some of our highly re-used properties have definitions that have tended either towards being very bland ("the of the item"), or else being lists of types ("the of the Event, Order, ScholarlyArticle, Person or MedicalCondition").
We should consider allowing per-type documentation strings to be recorded, and to be reflected into the site. I suggest we only do this if/when we integrate a JSON-LD parser into the site tooling (via rdflib), because the markup would be rather complex in RDFa.
Here's an example:
Triples (from json-ld playground parser):
Of course this would only be worth doing if we integrated the text back into our site somewhere.
Related issues: JSON-LD support; multi-lingual labels.