emmo-repo / domain-datamodel Goto Github PK

Even though that the ontology is fixed, there are still some choices for how entities and instances can be serialised in RDF, namely

• whether to serialise dimensions and properties as classes or individuals
• whether to shape is explicit serialised for instances

These alternatives are illustrated in this figure:

It is important that we choose an alternative. The choice should be documented. The above figure could be used for that.

Any opinions?

I think I like alternative 2 best (serialisation as individuals and without explicit of serialisation of shape for individuals) because of its simplicity.

Hi @jesper-friis ,

I am currenlty including a specification for the installation of this ontology within osp-core. The result of my implementation will make it pico-installable via cmdline (e.g. pico install datamodel or pico install datamodel.v0xyz ).

Would it be possible to add an inferred serialization of the datamodel ontology, hence I can add it as rawgithub to the yml-specification.

Best wishes
Matthias

We should use implementation-independe names of the subontologies.

Both relations range a xsd:string and have the example (m/s); which should be used when ?
Also in EMMO are classes not literals and EMMO does also include the hasReferenceUnit object property. Do I correctly assume, that hasUnit and hasUnitSymbol are used here as a simplification, because it's difficult to map the correct unit class ?