Hi Pat, I put together an example a year or so where I executed a SPARQL query, and got a result, and then added a triple to the graph, and ran the same query, and one of the binding sets in the original result was not present in the new result. That sure sounds non-monotonic to me. Now it may be that the triple store that I ran on had an incorrect implementation of SPARQL, but if so that vendor was unaware of the fact. Are you claiming that the behavior that I saw would be an indication of an incorrectly-implemented SPARQL on whatever RDF store I observed it?
- Bob
