Options for expressing the broader relationship among aspects

[rjyounes]

@uscholdm @philblackwood

In the new uom model, aspects are organized into a hierarchy. Options for the linking relationship, including pros and cons:

1. skos:broader

• Note: skos:broader has been proposed, but we actually want the transitive version, skos:broaderTransitive.
• Broadly disliked because it doesn't indicate direction.
• Up to now we have not wanted to import another ontology into gist. gist is an upper ontology - should it therefore be self-contained?

2. New property gist:hasBroader

• Con: Easily confused with skos:broader? - especially if the gist one is transitive.

3. New property with some other name

• Pro: Addresses problems with both 1 and 2.

4. Reuse gist:hasSuperCategory

• Pro: Doesn't involve using SKOS or defining a new predicate.
• Con: Although no domain and range are specified, the name suggests it only applies to categories, and therefore will be confusing to use with aspects.

5. Rename gist:hasSuperCategory (major change)

• Use a name to indicate hierarchical, broader/narrower relationships that does not suggest categories.
• Pro: Gets the advantages of 4 and solves the issues with all those above.
• This may bring us back to gist:hasBroader.
• Con: Changing the name of a long-standing property.

[rjyounes]

Doug:

• We're already using SKOS
• SKOS is the ontology of choice in the taxonomy world

Michael: imports are ignored, so import of SKOS isn't a problem
Dave/Rebecca: but we would want to load the imports into a KG

Dylan: Include a stub definition of skos:broaderTransitive?
Rebecca: but then we don't get transitivity
Michael: do transitivity with SPARQL property paths. Transitivity explodes the KG

Doug: Define our own property with an owl:equivalentProperty to skos:broaderTransitive
Jess: show that we work well with others
Peter: Cleaner to make our own property. Use an additional file with the equivalence for optional import, as we do with SHACL.

Rebecca: gist:hasBroader (transitive), gist:hasDirectBroader (intransitive)
Michael: gist:hasBroader (intransitive) and gist:hasBroaderTransitive
Dylan: hasParent?
Scott: don't worry about the SKOS names since we're not using SKOS. Add annotation pointing to SKOS terms.

Imported file would have properties back to skos:semanticRelation and include skos:Concept.

Whatever we decide we should document it.

DECISION:

1. Define 2 gist properties, transitive and non-transitive, and remove the 2 supercategory properties
2. Names: gist:hasBroader (transitive), gist:hasDirectBroader (intransitive)
3. We will not create a file containing the equivalences of SKOS to gist