Re: Support for reified sets



> On 17 Apr 2024, at 22:52, Luke VanderHart <lvanderhart@modern.energy> wrote:
> 
> If the group determines that reifiers can be card-many, then reifiers
> will become, de facto, a mechanism for creating and describing graphs
> (which is, by definition, a set of triples.)

I have read and listened to this argument several times. I have said several times the many reasons why it is not sound.
Today I am adding a new fancy reason using your very same argument in the other direction: a triple term “creates and describes“ a class (“which by definition is a set of” resources).
So, triple terms “will become, de facto, a mechanism for creating and describing” classes.
This would be true much more in general for any many-to-* predicates: they always induce the existence of a set of resources. But the resource on the other side of the predicate is never meant to create and describe a class, with the exception of the case when it is meant to: rdf:type.
The predicate rdf:reifies is meant to talk about reifications, not about naming  graphs. So, yes, it induces a set of the triple terms reified by the reifier: so what?
By the way, this set of triple terms is not a graph for several reasons I already mentioned several times. One is that a RDF graph defines the boundary of the scope of bnodes, while here bnodes are bound outside the set of triple terms (because they are transparent).
—e.

Received on Wednesday, 17 April 2024 22:01:02 UTC