Use of langString

[mereolog]

Why do we need rdf:langString?

If the reason is to have language-specific strings, why not to use rdf:PlainLiteral as the OWL spec recommends?
Then we wouldn't need to define rdf:langString as we currently do in:

fibo/FND/AgentsAndPeople/Agents.rdf

Lines 56 to 61 in a3a23e6

	<rdfs:Datatype rdf:about="&rdf;langString">
	<rdfs:label>langString</rdfs:label>
	<skos:definition>literal with a non-empty language tag that is well-formed according to section 2.2.9 of [BCP47]</skos:definition>
	<fibo-fnd-utl-av:explanatoryNote>This declaration is included in order to support language-tagged strings in FIBO.</fibo-fnd-utl-av:explanatoryNote>
	<fibo-fnd-utl-av:usageNote>The use of custom datatypes is outside the OWL 2 RL profile and so users should consider commenting this and its usage out in applications that are constrained to OWL 2 RL.</fibo-fnd-utl-av:usageNote>
	</rdfs:Datatype>

[ElisaKendall]

@mereolog You will see this in Commons as well. We are not using it explicitly in FIBO, but this is needed for some applications. As it happens, rdf:PlainLiteral was defined in the OWL specification because there was a gap between what was in RDF 1.0 and the second edition of the XML Schema Datatypes. The OWL 2 working group added it to OWL as a stop gap, but no one has really implemented it. Since the OWL 2 specification was revised in 2012, there is a new version of RDF, RDF 1.1, published in 2014. In that version of RDF, they implemented the notion of a language tagged string, or rdf:langString, which is what most triple store and other vendors have implemented. But it's a gap in OWL 2 and so we've included it in this union in the Text datatype for those that need it. That approach, using such a union, was recommended by Evren Sirin at Stardog, in fact.

[oldskeptic]

rdf:langString is the solution that I see used most often with support with SPARQL queries. rdf:PlainLiteral opens up new possibilities but has limited implementation support.