Back to Technical Modeling Framework

Back to start

This week we looked a new suggestion of using Named Graphs as a mechanism for the Shared Semantics work.

See next week 10 November for detailed treatment of this.

The following areas of the model were set up ahead of today's session for review today:

New Profile packages for Annotation Metadata

• DCT
• Provenance
• SKOS

This is the new diagram containing (very tentative) implementation of the proposed FIBO extensions for the DC Source relationship, namely the terms for the existing textual elements for Term Origin and Definition Origin.

• These have been tentatively shown as extending the DC Source relationship.

• Term Origin is further sub-classed into Term Originator (which would be an entity) and Term Original Name (the actual name of the term in the original source), since these elements are both present in the textual “Term Origin” entries in the current model.

• Definition Origin has been split into two separate but similar terms (each of them a sub-relationship of DCT:Source). These are “Definition Origin” and “Definition Modified From”.

Note that the naming of these should probably be lowerCamelCase since these are not intended for natural language reading, they will be stereotypes which can be used for modeling; aliasing will be required in reports to give the natural language variant of these as per the natural language naming table in the RFC.

Suggestion: Can we create named graphs into which we import the sub-sets of the OWL ontologies we are interested in using SPARQL construct clauses.

Lets us manage the lifecycle for a particular namespace.

Use named graphs to keep everything straight in a single repository. Have different versions of an ontology in the same repository. Treatment Would not need to create two of everything, but would have the Named Graph in the FIBO section itself e.g. the Geo section.

• We can add our own axioms
• We can add the archetype

Action: Detailed proposals to be presented for the 10 November session.

Diagram Note: Target:

Not a literal but an OWL Individual

Type: URI or Bibliographic Reference

Discussion: MB had tentatively put this relationship Dublin Core Source, as having a target (range) of RDFS Literal.

Review consensus is that this is incorrect: the target should be an OWL Individual.

Action: Added “Individual” in this place in the model (10 November)