DocRulesClosure

From W3C Wiki

The DocRulesClosure of an RDF graph is another RDF graph which is a particular subgraph of the (enormous) deductive WebClosure of the first graph. It is a subgraph which is hoped to be useful in size. This is an experimental design for FollowLinksForMoreInformation, based on the doc:rules predicate.

The DocRulesClosure of an RDF graph serialized as document D1 is defined roughly like this:


   create an empty formula R	
   for each URI U in D1 (starting document):
     fetch the document (D2) at address U
     for each triple `U doc:rules ?U3` in D2:
        fetch the document (D3) at address U3
        add all contents of D3 to R
   Compute the deductive closure of D1 with respect to R.


Should we do this kind of processing on each URI in D2 and D3???

Test: includes triples from D1 (yes)

Test: includes triples from (each) D2 (no)

Test: includes triples from (each) D3 (yes)

(Also test whether such triples are used in rule application; should be same.)

Test: do we use rules in D1 (no, that's a separate thing)

Test: do we use rules in D2 (no)

Test: do we use rules in D3 (yes)

Test: various forms for D2

Test: various forms for D3

Test: various indirections and redirections on U and U2

Test: multiple D3's for a single D2