Fixing the definition of instance of a graph to consider triple terms

[franconi]

The definition of instance of a graph should be fixed to include triple terms as follows:

Suppose that M is a functional mapping from a set of blank nodes to some set of RDF terms. Any graph obtained from a graph G by replacing some or all of the blank nodes N in G by M(N) is an instance of G.

[pfps]

What is the definition of "in" here?

[franconi]

N in G if N appears in the parse tree of G.
I guess that the same sort of ambiguity happens also in the RDF 1.1 spec.

[pfps]

And what is the "parse tree"?

[franconi]

Ok, so your issue is not about the change introduced by the new formulation, but about part of the formulation which was present before already; fair enough.

We could either mention the grammar (e.g., it appears as a RDF term in the graph) or explicitly write the inductive definition of "appears in", which is used elsewhere in the semantics document. Maybe this latter approach would be better.

[franconi]

Definition of "in" (sometimes written as "appears in"), with A, B triples or terms and G graph:

A is in a graph G if A is in a triple B of G.
A is in B if either A is equal to B or A is in the subject or predicate or object of B if B is a triple or triple term.

[franconi]

Using the above definition, we should replace the text "if the triple structure appears in S" (used to define entailment patterns) with "if the triple or the triple term is in S".

[doerthe]

I would be OK with keeping the "in" but I think an issue is that this notion got more complicated because we now have nested terms which were not present in RDF 1.1. I am not happy with the definition though, mainly because you use "in" to define "in".

Definition of "in" (sometimes written as "appears in"), with A, B triples or terms and G graph:

A is in a graph G if A is in a triple B of G. A is in B if either A is equal to B or A is in the subject or predicate or object of B if B is a triple or triple term.

Maybe:

Given a graph G, we define the set T(G) of terms of G recursively as follows:
for each triple (s p o)\in G, s,p,o\in T(G)
for each (s,p,o) \in T(G), s,p,o\in T(G)

We say that A is in G if A\in T(G)

Of course we need to make it more "RDF spec style".

[franconi]

Given a graph G, we define the set T(G) of terms of G recursively as follows:
for each triple (s p o)\in G, s,p,o\in T(G)
for each triple term (s,p,o) \in T(G), s,p,o\in T(G)

[franconi]

Added to PR #74.