Duplicated or similar terms

[gkellogg]

Terms created via <dfn become exported and can be automatically reconciled by other specifications. Between RDF Concepts and Semantics a number of terms are created, and sometimes overlap.

• It's possible to "define" a term but reference a term from another spec using data-cite. For example, <dfn data-cite="RDF12-CONCEPTS">RDF graph</dfn>, which styles the term in the defining spec, and allows simple references such as <a>RDF graph</a>.
• Terms intended to be used locally can be declared to not be exported, for example <dfn class="no-export lint-ignore" data-dfn-for="rdf">invalid</dfn> (which comes from Collisions with css-syntax's [=invalid=] rdf-semantics#13).
• Former terms can be demoted, but should have an anchor so that any existing references will continue to resolve. For example <dfn class="export">invalid</dfn><span id="dfn-invalid.x"><!-- obsolete term --></span>.

Given the potential range of updates, we should consider what to do about the following terms:

• Denote – defined in RDF12-CONCEPTS and in RDF12-Semantics. Also Denotation in Semantics.
• Entailment – defined in RDF12-CONCEPTS and as Entail in Semantics
• Equivalent – Defined as Equivalence in Concepts and Equivalent in Semantics.
• Referent – defined in RDF12-CONCEPTS and in RDF12-SEMANTICS

Many other terms in Semantics are exported but not used, at least in RDF Semantics (e.g., rdfs13. We might consider just making these "no-export" and just preserve the anchor.

Thus far, SPARQL doesn't export any terms, but probably will at some point.

• RDF Term
• Simple Literal – Obsolete now, but matches simple literal in RDF Concepts.
• RDF Dataset

[afs]

RDF Term and RDF dataset are now defined in RDF 1.1 Concepts. (RDF Dataset is a different definition).

Such terminology wasn't available for SPARQL 1.0 or SPARQL 1.1.

[hartig]

Indeed, the definition of 'RDF dataset' in RDF 1.1 and in SPARQL 1.1 is slightly different. The difference is in terms of what can be a name of the named graphs that are contained in such a dataset. For the notion of RDF dataset in RDF 1.1, the graph names can be IRIs or blank nodes, whereas for the notion of RDF dataset in SPARQL 1.1, the graph names can only be IRIs. I guess that this difference is also something we want to get rid of.

[afs]

Recorded as : w3c/sparql-query#28