Reification Alternatives

From OWL
Jump to: navigation, search

This page is designed to discharge ACTION-129, which is to explicate the space of alternative representations of negative property assertions ISSUE-81 and axiom annotations ISSUE-67.

Space of Alternatives

Technique Example Issues
Reification Current solution; people don't like reification
Reification shadow vocab Doesn't really improve things over base reification
Named Graphs Not a standard feature.
Encoding using equivalent forms (like nominals) Doesn't work for axiom annotations; hurts roundtripping/transparency/editing; increases the complexity of the base logic
Using Literals Some parsers can't handle that.
Some other syntactic form Loss of symmetry with regular expressions.
Data schemed uris In email Compact and clean, at least in some ways. But it requires special casing those URIs. You might have a different syntax/mimetype in the data uri than in the parent document. RFC 2397 Wikipedia article

A problem with all triple based reification is that nested reification *really* sucks. See http://cs-www.cs.yale.edu/homes/dvm/papers/McDermottDou02.pdf

Proposals

  • Leave things as they are, i.e., use the RDF reification vocabulary for negative property assertions and most annotated axioms, possibly (but probably not) with the change that the base triple is added for annotated axioms that do do reification. pfps 13:09, 7 May 2008 (EDT)