Closed siuc-nate closed 1 year ago
On the surface, I don't see any problem with this change.
I thought the purpose of sourceDocumentation was for when something like an instruction manual had been used to derive competences. In this case the source document would be the job ad from which the competence was inferred. I think EMSI use wikipedia as a reference for what has been derived from other docs, so not a source more of a target.
I'd suggest subjectWebpage as a better match for including a URL for a reference giving more information about a competence. [Better still, at least for machine readability, sameAs pointing to the dbpedia or wikidata id, but that would take some computation to work out.]
They may pick which words to use from job ads, but they pull from wikipedia to actually define those words. For non-EMSI cases, I would think sourceDocumentation would be a little more straightforward. I'm not opposed to using subjectWebpage either though, if that covers all cases.
They may pick which words to use from job ads,
yes, I think that defines "source" of sourceDocumentation
but they pull from wikipedia to actually define those words.
and I think that is describing the entity, which is what a subjectWebpage does
I think this works for named entity extraction and disambiguation (NEE/NED) in general.
From Oct 18 meeting:
For the EMSI use case, ceterms:subjectWebpage
works.
For other potential use cases:
Add skos:Concept, skos:ConceptScheme, ctdlasn:CompetencyFramework to domain of https://credreg.net/ctdlasn/terms/sourceDocumentation
Update the definition from: "Human-readable information resource other than a competency framework from which this competency was generated or derived by humans or machines." to: "Human-readable information resource other than a competency framework from which this resource was generated or identified by humans or machines." comment: "The intended use case for this is when documentation is analyzed in order to find and describe entities such as competencies and concepts."
Proposal:
Add:
Subject: ceterms:subjectWebpage Predicate: schema:domainIncludes Object: skos:Concept
Subject: ceasn:sourceDocumentation Predicate: schema:domainIncludes Object: skos:Concept, skos:ConceptScheme, ctdlasn:CompetencyFramework
Update:
Subject: ceasn:sourceDocumentation Predicate: rdfs:comment Remove Object: "Human-readable information resource other than a competency framework from which this competency was generated or derived by humans or machines." Add Object: "Human-readable information resource other than a competency framework from which this resource was generated or identified by humans or machines."
Add:
Subject: ceasn:sourceDocumentation Predicate: dct:description Object: "The intended use case for this is when documentation is analyzed in order to find and describe entities such as competencies and concepts."
@philbarker Is subjectWebpage meant to be added to any other classes?
The above changes have been made in pending CTDL/CTDL-ASN and noted in the history tracking.
I will close this thread pending @philbarker's approval.
That looks good to me.
The property https://credreg.net/ctdlasn/terms/sourceDocumentation could serve the same purpose for Concepts with the following modifications:
The motivating factor for this, currently, is to enable including the wikipedia URL in the EMSI data (if we document them as concepts).