Closed JachymHercher closed 5 years ago
I would vote for an attribute as well as this would help us to keep a simpler structure in my view.
A property is the lowest granularity of a concept. Semantic linking cannot be done at property level it is done at class level. Predicates in the ontology are from class to class.
Lots have there own properties and are semantically linked to other classes. The advantages of having an ontology would be reduced if we did not create a lot class, as open data evolves this could become a limiting factor.
I agree with @muricna.
I think implementation is getting mixed up with modelling in the earlier comments. When it comes to implementing the ontology in a real system, an attribute can be used on whatever concrete models are implemented in that system (in many systems, the only way to link classes is via foreign key attributes). The ontology is not a prescription that, if something is a class in the ontology, then it must be defined as a class in e.g. the object-oriented programming language used for implementation.
@muricna I'm not sure I understand, sorry. For example, many classes have the property "description". "Description" is not the lowest granularity of, for example, "Options", no?
You are right that we lose the predicates. However, for lots, aren't all predicates just "applies to" (or something very similar)?
Having Lot as a class is both a business and a design decision:
The Working Group endorses this opinion as per today, 9th July 2019. So we close this issue.
In the ontology, lot is a class, not a property. Why?
Instead of having a class and linking (almost) everything to that class, (almost) everything could have a lot property. That would make it clear which lots criteria, estimated values, submission deadlines, etc. apply to.
I would assume either approach should be possible, I'm wondering which one is simpler and more readable.