SLDS ScDU CEDS JSON-LD Workgroup Proposing Human Readable versus Opaque IRIs

[AEMDuaneBrown]

The SLDS Scalable Data Use Workgroup developing CEDS JSON is proposing a change to the Ontology to allow for the JSON-LD to be more human readable. This involves changing the Ontology IRI's from ID's to the CEDS technical names. Examples below of how this would impact the Ontology are provided. Note that the ID's will still be provided as metadata for the Classes, Properties, and Individuals, but will exist as dc:identifiers only. Everything else remains the same.

Current Ontology Naming:

<!-- Class: OrganizationName -->
**<rdfs:Class rdf:about="http://ceds.ed.gov/terms#C200246">**
	<rdfs:subClassOf rdf:resource="http://ceds.ed.gov/terms#BaseCEDSResource"/>
	<rdfs:label>Organization Name</rdfs:label>
	<skos:notation>OrganizationName</skos:notation>
	<dc:creator>Common Education Data Standards</dc:creator>
	<dc:identifier rdf:datatype="http://www.w3.org/2001/XMLSchema#token">C200246</dc:identifier>
    <rdfs:comment>Information about an organization.</rdfs:comment>
</rdfs:Class>

Proposed Ontology Naming:

<!-- Class: OrganizationName -->
**<rdfs:Class rdf:about="http://ceds.ed.gov/terms#OrganizationName">**
	<rdfs:subClassOf rdf:resource="http://ceds.ed.gov/terms#BaseCEDSResource"/>
	<rdfs:label>Organization Name</rdfs:label>
	<skos:notation>OrganizationName</skos:notation>
	<dc:creator>Common Education Data Standards</dc:creator>
	<dc:identifier rdf:datatype="http://www.w3.org/2001/XMLSchema#token">C200246</dc:identifier>
    <rdfs:comment>Information about an organization.</rdfs:comment>
</rdfs:Class>

Current Ontology Naming:

<!-- Data Property Relationship: OrganizationName -->
**<rdf:Property rdf:about="http://ceds.ed.gov/terms#P000009">**
<rdfs:label>Activity Title</rdfs:label>
	<skos:notation>ActivityTitle</skos:notation>
	**<schema:domainIncludes rdf:resource="http://ceds.ed.gov/terms#C200246"/>**
	<schema:rangeIncludes rdf:resource="http://www.w3.org/2001/XMLSchema#string"/>
	<dc:creator>Common Education Data Standards</dc:creator>
<dc:identifier rdf:datatype="http://www.w3.org/2001/XMLSchema#token">P000009</dc:identifier>
</rdf:Property>

Proposed Ontology Naming:

<!-- Data Property Relationship: OrganizationName -->
**<rdf:Property rdf:about="http://ceds.ed.gov/terms#ActivityTitle">**
<rdfs:label>Activity Title</rdfs:label>
	<skos:notation>ActivityTitle</skos:notation>
	**<schema:domainIncludes rdf:resource="http://ceds.ed.gov/terms#OrganizationName"/>**
	<schema:rangeIncludes rdf:resource="http://www.w3.org/2001/XMLSchema#string"/>
	<dc:creator>Common Education Data Standards</dc:creator>
<dc:identifier rdf:datatype="http://www.w3.org/2001/XMLSchema#token">P000009</dc:identifier>
</rdf:Property>

[AEMDuaneBrown]

Note where there are double asterisks in the code above. That is where the original and changes are being highlighted.

[philbarker]

Well, that's annoying. I spent I do not how long changing the PESC JSON-LD transcript from using technical names to numbers (it put that work back about month because I don't have much time per week to spend on it). With this I would have to change them back again.

I do understand the human factors in working with opaque ids. When I suggested using opaque ids, because they are more maintainable (i.e. you can change labels without creating inconsistencies) and more international, I also said that you need some technology to support human users. One would be a lookup service so that when the technical name is entered the numeric id is provided.

Is the CEDS JSON group using JSON LD? If not then they don't need to use the RDF ids at all, but could produce a mapping from JSON to RDF in RML to get RDF when it is needed. If they are, then they can use technical names as keys in the JSON and give these the RDF @id in the @context.