Fork me on GitHub
#rdf
<
2020-09-18
>
apbleonard11:09:40

Is there an RDF vocabulary out there that matches (or is a superset of) the datomic data model? https://docs.datomic.com/cloud/schema/schema-reference.html

apbleonard11:09:38

I looked for an equivalent of :db/cardinality in RDFS and OWL and now I have a sore head.

rickmoynihan14:09:42

@apbleonard: I doubt it. They have very different backgrounds. There is of course owl:maxCardinality but OWL is a logic language for inferencing; not constraints… So owl:maxCardinality 1 doesn’t so much enforce that a predicate has only one value, it just draws an additional conclusion from your data via the non-unique names assumption that all the values for that predicate are actually the same instance.

rickmoynihan14:09:17

If you want closed world constraints, you probably want to enforce them via something like SHACL.

apbleonard15:09:32

@rickmoynihan Thanks. Interesting day reading up on these topics 🙂 Realising that ontologies need not focus on constraints (like cardinality) and still provide value....

rickmoynihan16:09:40

It really depends what you’re doing, and what camp you’re in. If you’re in the linked data camp; then you mainly just use ontologies as a set of global identifiers for things, for basic modelling. If you’re in the semweb camp; then you use ontologies as a logic to conclude more things about the world via reasoning. You can have a foot in both camps… but the distinction is pretty much real. Many vocabularies purposefully avoid having to lean on owl too much e.g. skos… as many people find OWL to be somewhat arcane and cumbersome and it comes with a load of caveats/nuance etc.