On 3/27/17 10:42 AM, Daniel Kinzler wrote:
Am 27.03.2017 um 15:13 schrieb Kingsley Idehen:
I see Wikidata is a collection of reified RDF Statements. I don't see how this model differs from RDF's model. It just so happens (in my eyes) that Wikidata includes description of statements about things which provides rich metadata, in line with the goals of Wikidata.
It's a matter of perspective.
I agree that Wikidata can be *represented* as a collection of reified RDF Statements. That's what we do for the query service. But I do not agree that this is what Wikidata *is*.
My point is that your model boils down to treating statements as "first class citizens" so to speak. If true, then it is as I described i.e., still an RDF model, but with emphasis of statement reification, which is actually a good thing.
RDF and the Wikibase model are quite different conceptually.
RDF doesn't exclude reification. Put differently, using reification doesn't amount to a new model different from that of RDF.
But they are of equal power and thus formally equivalent: one can be represented using the other. Just because a Turing Machine is computationally equivalent to lambda calculus, that does not mean they are the same thing.
I am not implying that.
Understanding one in terms of the other may be helpful in some context, and irrelevant in another.
There is nothing special about the relationship between Wikibase/Wikidata and RDF; Wikibase has an RDF binding, but it is not defined in terms of RDF, its specification does not rely on RDF concepts.
RDF concepts boil down to the use of sentences to describe anything, just as we do everyday in the so-called real world.
The Wikibase model can just as well (or perhaps more easily) be understood and represented in terms of the Topic Maps model (ISO 13250).
Academically, the Wikibase model could perhaps be described as an extended model logic with reasoning rules for provenance. I think W. Stelzner explored related ideas in the 80s. Maybe one day I'll find the time to dig into this some more.
I think we can just agree to disagree for now, since nothing you've stated is fundamentally contrary to my view of RDF -- as a Language for describing anything (including statements) :)