Example Data Type Mismatch
awanczowski opened this issue · 11 comments
Looking at the Team Roster the values for some properties do not match with the rdfs:range specified in the Ontology.
- uniformNumber is a String, should be an Integer
- membershipStatus is an IRI, should be xsd:anyURI
- positionRegular is an IRI, should be xsd:anyURI
Hey @awanczowski could you paste examples here? We had trouble spotting it in the meeting today.
Hi @pauljkelly, here are some examples.
Uniform Number should be an integer: https://github.com/iptc/sport-schema/blob/develop/samples/ttl/team-roster.ttl#L44
IRI instead of xs:anyURI: https://github.com/iptc/sport-schema/blob/develop/samples/ttl/team-roster.ttl#L43
If Uniform Number is an integer it will be unable to accommodate those historical usages of "00", and the contemporary use of Usain Bolt wearing "9.58" for Stromsgodset and in Soccer Aid matches.
Reviewing the Ontology, the property for uniformNumber should be xsd:string not xsd:integer.
https://github.com/iptc/sport-schema/blob/main/ontologies/iptc-sport-ontology.ttl#L421
Hi @awanczowski. We're reviewing this issue. We fixed the uniform-number issue. Is there something we wtill need to change regarding the IRIs?
Hi @pauljkelly,
I'm not sure what the intention was for some of these items that point out to external vocabularies.
Ontology is xsd:anyURI - https://github.com/iptc/sport-schema/blob/main/ontologies/iptc-sport-ontology.ttl#L501
Samples are IRIs - https://github.com/iptc/sport-schema/blob/develop/samples/ttl/team-roster.ttl#L43
<http://cv.iptc.org/newscodes/spichposition/center>
vs "http://cv.iptc.org/newscodes/spichposition/center"^^xsd:anyURI
I'll do some more research on the proper way to reference the external items that are not maintained within the same ontology.
OK, I see that.
So this is an IRI? --> http://cv.iptc.org/newscodes/spichposition/center
I'm not sure I know what an IRI is but does it not encompass URIs?
All URIs are valid IRIs according to the specification.
I believe for our use case here, an IRI is should be resolvable to another item in the ontology. If you are unsure if the value would be resolvable it would best be a URI.
@awanczowski you able to attend this Tuesday's meeting to discuss this issue? I'd like to put it to bed before submitting the ontology for ratification this week.
@pauljkelly, Yes I should be able to attend.