How do RDFS inferencers behave in quadstores with respect to graph/context information

  • Do such inferencers only inference within the scope of a context/graph, or do they inference over all contexts/graphs?
  • If inferencing over all contexts/graphs, how do they decide what context/graph the inferred statements belong to?

As recently discussed in the Sesame mailing list, Sesame's 2.x forward-chaining RDF Schema inferencer

… takes as input the statements from all contexts (including the null context) and applies inferencing to the combined graph, adding the inferred statements to the null context.

Other inferencing patterns are being proposed for Sesame 3.x.