Relating the ISO 11179 Metadata Registry (MDR) Standard to Semantic Web Standards

This is a brief note on my perception of the ISO 11179 Metadata Registry (MDR) standard, including my perception of its relation to W3C Semantic Web standards. Examples are taken from this Wikipedia article.

In ISO 11179, there are concepts. There are relations of concepts to wider and more narrow concepts. There are also relations between a concept and one or more representations of the concept (i.e. different terms/synonyms are used, but the concept is the same).

There are object classes. An example is a “person”. Combining an object class with a concept such as “income” forms a more specific data element concept, e.g. a person’s income. Different representations of a data element concept are data elements. Differences in these representations may be due to the use of different terms/synonyms, and/or due to different value domains.

An example of a value domain for the data element concept “sex of person” is “M=Male, F=Female, U=Unknown”. No correctness is implied by selection of a value domain – it simply is in use or is not in use by a given data element. {M,F,U} are permitted values for “sex of person” in a particular data set where “sex of person” is the data element concept and e.g. “sex of individual where the value domain is {M,F,U}” is a data element for that concept, where the specialized representation in this case comprises both a synonym term (“individual” for the object class “person”) and a value domain. Data sets are many-to-many wrt data elements.

Relation to Semantic Web standards

A concept seems like a skos:Concept entity, which relates to e.g. skos:broader and skos:narrower concepts. Other SKOS relations can apply, e.g. skos:related, skos:closeMatch, skos:exactMatch, etc. An object class seems like a rdfs:Class / owl:Class.

A data element concept, an association of an object class with a conceptual characteristic, seems like e.g.

IF   { :p a rdf:Property .
       :p rdfs:domain :MyObjectClass ;
          rdfs:range :MyCharacteristicConcept . }
THEN { :p a :MyDataElementConcept . }

and thus an rdf:Property can be a bridge for multiple perceived data element concepts. Or, you can use one property name for each distinct data element concept.

A data element seems like something that can subclass an owl:Restriction entity, extend its rdfs:domain, or both, e.g.

  a :MyDataElementConcept ;
  rdfs:subClassOf [ a owl:Restriction ;
                    owl:onProperty :p ;
                    owl:allValuesFrom :SomeSet ;
                    rdfs:domain :MySynonymOfObjectClass ] .

I haven’t seem rdfs:domain used inside a owl:Restriction class, but owl:onProperty is singular and required, so the semantics of rdfs:domain entailment via an owl:Restriction seems clear to me. There is likely a simpler way, but this was merely an exercise in mental mapping.