A focus on reification

Hi all,

I've focused on reification practices based on RDF 1.1 and OWL in a
document at [1]. It is "free" from RDF-star, focusing only on what
reification means in terms of its "challenging" flexibilities.

OWL is used to deal with this flexibility. Particularly through
definitions of "statement types" (being a subclass of rdf:Statement
and an OWL intersection of the restrictions for the particular triple
constituent predicates and values), and a "fact type" (simply using
the triple properties as a key).

Of course we cannot base RDF 1.2 upon OWL. The purpose is only to show
what kinds of concepts, interpretations, restrictions and entailments
are possible upon "mere" reification. And to support considering if
the consequence of basing RDF-star on basic reification is acceptable,
by showing some possible advanced usages building on such a fundament.

This includes:
* talking about named occurrences, being the actual reified
rdf:Statement instances;
* allowing even a many-to-many relationship between a statement and triples;
* still favour keeping the reification as one simple statement
(representable by one triple);
* describing the fact itself (represented by the asserted triple in the graph);
* in any case keeping statement integrity, if so desired, through a
backing type.

That should span the range of use cases from simple ones such as
suggested "motivating examples" [2], via explicit descriptions of
chosen facts and lexically precise descriptions, to overclaims for the
pursuer of "truth makers".

The results are comparable to those achieved by either
(non-functionally restricted) named triples, or virtual occurrenceOf
relationships to occurrence sets. Such designs could be used in
implementation for optimization, but would not need to be part of the
concepts, abstract syntax or semantics.

If that is acceptable, [3] could be revised to be about reification
(promoting it from the informative section [4]), and (as has been
suggested occasionally, and more so recently) have RDF-star simply be
syntactic sugar for that.

Best regards,
Niklas

[1]: <https://hackmd.io/@niklasl/rkpp-eXFp>
[2]: <https://hackmd.io/@niklasl/HJ3IudCdp>
[3]: <https://www.w3.org/TR/rdf12-concepts/#section-quoted-triples>
[4]: <https://www.w3.org/TR/rdf12-semantics/#non_semantics>

Received on Tuesday, 16 January 2024 16:51:14 UTC