Sorry about a blank subject before...
Came across a good CalTech presentation called Relevance of Topic Maps
at
http://www.cacr.caltech.edu/semast/files/mahabal.pdf
The presentation has the infamous Cregan datamodel on page 5. The question
is why not use this already-certified ISO datamodel being mapped to RDF
somewhere -- why not IT be the foundational conceptual model of Wikidata?
Topic Map API could be built into the WOM. It's a good initial architecture
for a cognitive model to interact at the human level and the data storage
model at the triples level.
I like SKOS and Dublin Core as core ontologies for Wikidata. SKOS is a
subject index and Dublin Core is the index card. SMW subobjects like
fullpagename#fullpagename expressions can be (have) a (the) Dublin Core set
of properties; this design satisfies directly the needs of RSS folks an
importatn requirement.
Maybe melding Topic Maps, RDF and Wikis might boost web semantics by
creating an orientation towards understanding pithy lexical language
expressions common in subject headings. Doing so would be a major practical
milestone to achieve alongside semantic infoboxes -- that might create a bit
more buzz than "formalizing the adoption of database technology" would, in
its own right (I say this with reference to Ward Cunningam's latest
interests). Essentially it'd be great to integrate wikidata topic maps with
(group) mind mapping contexts - another natural fit that so could rock.