Closed pfps closed 1 year ago
It might be good to have a way term to distinguish language-tagged strings from other literals. I've seen the term "datatyped literal" used to distinguish between the two. Previously, there would have been three types of literals: "plain", "datatyped", and "language tagged".
If a term is removed, we should probably try to keep some anchor (e.g., <span id="dfn-..."></span>
) so that references to the old term continue to resolve.
In RDF 1.1 language-tagged strings have datatypes so using datatyped literals to refer to the other literals is, in my opinion, confusing.
I don't believe that my changes remove any dfn spans.
It might be good to have a way term to distinguish language-tagged strings from other literals.
We have a way: rdf:langString vs xsd:string.
language-tagged string
describes a literal with a language-tag
and datatype rdf:langString
. Other literals have a datatype, but no language-tag, and the datatype must not be rdf:langString
. There is a component which is a string, but I would not say that string
describes literals that are not language-tagged.
@gkellogg Is your request to have a term that describes all literals except language-tagged strings? The place to introduce that is almost certainly in RDF Concepts, where the other syntactic categories are defined, not in RDF Semantics.
The RDF Semantics document uses syntax to refer to the abstract syntax of RDF, where every literal does indeed have a datatype IRI. But it would be better to just say "RDF" instead of "RDF literal syntax", which I have done.
RDF 1.0 had untyped and typed literals. RDF 1.1 has only typed literals but some occurrences of ''typed literal" and "datatyped literal" remained in RDF semantics. They are mostly harmless so it should be considered an editorial change to remove the unnecessary adjectives.
Preview | Diff