Re: voting against use of rdf:ReificationProperty

On 1 Sep 2024, at 15:28, William Van Woensel <william.vanwoensel@gmail.com> wrote:

I think an example using foaf:name is obvious that there’s a modeling issue. I wonder what you think of my example about trying to mix in use of RDF 1.0 reification, though? If you end up with two statements like:

_:a rdf:object <<( :s2 :p2 :o2 )>> .
_:b rdf:object “William” .

Would you agree that we can’t really say anything about whether there’s a modeling issue in the larger dataset?

I would say that it still depends on the dataset author (and, anyone using the dataset and its inferences). They will see "rdf:object" appearing as a ReificationProperty, and may have a problem with that, or not. If they happen to work with RDFS 1.1, then _:a and _:b would be flagged<https://www.w3.org/TR/rdf11-mt/#rdfs-interpretations> as as rdf:Statements, i.e., using the "old" reification properties. This may help with pointing out a problem, or not; it would certainly be against the properties' intended usage, and for me it would certainly point out a modeling issue. Btw, if it breaks the system they are using (i.e., if that system also supports the old way of reification), then they will more likely have a problem with it; and may want to rethink their approach (or, turn off backwards compatibility in the system :-).

I fully agree with William.
Obviously the authors of the graph decided to use a standardised property such as "rdf:object” — which has an impact on entailment, namely that _:a and _:b are a rdf:statement — in a different standardised context, which entails that "rdf:object” is a rdf:reificationProperty. They have to live with the consequences, unless they realise by looking at the entailments that there is a mistake in their model: as William says, both cases are fine, in the sense that it is up to the authors to think about which one makes sense in their context.
—e.