Re: Graphs and Being and Time

On 23/02/11 22:38, Nathan wrote:
> Hi Pat,
>
> All I can say is that I believe you've hit on the distinction we can
> make between a Named Graph and a Graph Literal.
>
> Where a Named Graph correlates to a "Graph Token" as you term it, a
> container for triples which can be given a name and accessed, the
> contents of the container may change over time, but the container
> remains the same, with the same name.
>
> Where Graph Literal is an abstract graph, a set of triples, that which
> is given by a named graph when poked.
>
> The "value" of a "named graph" at time t is a "graph literal".
>
> That "value" can be serialized in to a lexical form and passed in a
> representation (rdfa, rdf/xml json/rdf, a representation etc), held in
> memory or in a quad store or whatever.

FWIW:

SPARQL 1.1 Update defines a graph store:

http://www.w3.org/TR/sparql11-update/#def_graphstore
"""
A GraphStore GS is a mutable container of RDF graphs. It has one unnamed 
(default) slot and zero or more named slots identified by an IRI <i>. 
Each slot holds an RDF graph.
"""

This is an attempt to allow graph-as-value (immutable) and capture 
change due to operations.  The contents of a slot (named or the unnamed 
slot) can change value.  "RDF graph" is used to mean the abstract value 
(?? "RDF Abstract graph"), not any representation of the value.

It was a conscious choice to have udpates affect a graph store, rather 
than an "RDF dataset", which is the thing that query acts over. "RDF 
dataset"

SPARQL 1.0:
http://www.w3.org/TR/rdf-sparql-query/#defn_RDFDataset

"""
An RDF dataset is a set:
      { G, (<u1>, G1), (<u2>, G2), . . . (<un>, Gn) }
where G and each Gi are graphs, and each <ui> is an IRI. Each <ui> is 
distinct.
"""
so it's not a mutable container.

	Andy

Received on Sunday, 27 February 2011 22:03:23 UTC