update to graph isomorphism

[pfps]

As far as I can tell, graph isomorphism needs the following extra clause

• M(tt) is the triple term ( M(s), M(p), M(o) ) for tt a triple term of the form (s, p, o)

This is slightly clunky but doesn't use any concrete syntax.

[hartig]

I will take care of this one.

[hartig]

I was just about to implement the change as proposed in the issue description above. Yet, it turns out that it may not be exactly that simple, for the following reasons: By the current definition, the domain of the bijection_M_ is the set of all nodes of the graph G. Yet, it is not guaranteed that any of the three elements of the triple term tt = (s, p, o) is such a node. Therefore, M(s) may be undefined, and so may M(p) and M(o).

The fix to this problem may be to change the domain and the co-domain of the bijection M to both be the set of all RDF terms. With this change, the extra clause mention in the issue description above can then be added without problems. So, to be explicit, in addition to adding this extra clause we need to change the sentence before the bullet points in the section. I propose to change that sentence as follows:

Two RDF graphs G and G' are isomorphic (that is, they have an identical form) if there is a bijection M from the set of all RDF terms into that same set, such that all of the following properties hold:

I will implement this change once PR #137 is merged.

[pfps]

Good catch

[pchampin]

Also, note that M(p) is not necessary in the recursion (p is always an IRI, so M(p) is always equal to p).