Re: A proposal for entailment tests

> OK, great. Suggestion: how about showing the *proofs* of the
> entailments in some form? For rdf, this could be simply the
> intermediate subgraph of the (merge of the) antecedent(s) which
> generalizes to the consequent graph. For the rdfs cases it could
> include the rules used to generate the relevant part of the rdfs
> closure, with references to the numbers in the MT table (?)
>
> If this would be a lot of work to generalize, then never mind; but I
> think it might make it easier to see what is going on, and it would
> certainly emphasise the difference between (and relationships
> between) rdf  and rdfs entailments.

That is indeed a very good suggestion Pat.
We made an initial attempt to start doing that.
We put an N3 set of rdfsrules at
  http://www.agfa.com/w3c/n3/rdf/rdfsrules.n3
and as a matter of initial test, we ran

////////////////////////////////////////////////////////////////////////////////
C:\>java -classpath /euler -Dhttp_proxy=http://vam619.roam.agfa.be:80 Euler http
://www.w3.org/2000/10/rdf-tests/rdfcore/rdfs-domain-and-range/test001.nt http://
www.w3.org/2000/10/rdf-tests/rdfcore/rdfs-domain-and-range/test002.nt http://www
.w3.org/2000/10/rdf-tests/rdfcore/rdfs-domain-and-range/test003.nt http://www.ag
fa.com/w3c/n3/rdfsrules.n3 http://www.w3.org/2000/10/rdf-tests/rdfcore/rdfs-doma
in-and-range/test004.nt
# Generated with http://www.agfa.com/w3c/euler/#27.079 on Sat Sep 22 15:27:02 CE
ST 2001
# for query http://www.w3.org/2000/10/rdf-tests/rdfcore/rdfs-domain-and-range/te
st004.nt
# given [http://www.w3.org/2000/10/rdf-tests/rdfcore/rdfs-domain-and-range/test0
01.nt, http://www.w3.org/2000/10/rdf-tests/rdfcore/rdfs-domain-and-range/test002
.nt, http://www.w3.org/2000/10/rdf-tests/rdfcore/rdfs-domain-and-range/test003.n
t, http://www.agfa.com/w3c/n3/rdfsrules.n3]

@prefix e: <http://www.agfa.com/w3c/euler#>.
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>.
@prefix log: <http://www.w3.org/2000/10/swap/log#>.
@prefix : <http://www.agfa.com/w3c/n3/rdfsrules#>.
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>.

 {<http://foo/bar> rdfs:domain <http://foo/Domain1>.
  :RULE5 a rdfs:Resource.
  <http://foo/baz1> <http://foo/bar> <http://foo/baz2>} log:implies
{<http://foo/baz1> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://foo
/Domain1>}.
 {<http://foo/bar> rdfs:domain <http://foo/Domain2>.
  :RULE5 a rdfs:Resource.
  <http://foo/baz1> <http://foo/bar> <http://foo/baz2>} log:implies
{<http://foo/baz1> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://foo
/Domain2>}.
 {<http://foo/bar> rdfs:range <http://foo/Range1>.
  :RULE6 a rdfs:Resource.
  <http://foo/baz1> <http://foo/bar> <http://foo/baz2>} log:implies
{<http://foo/baz2> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://foo
/Range1>}.
 {<http://foo/bar> rdfs:range <http://foo/Range2>.
  :RULE6 a rdfs:Resource.
  <http://foo/baz1> <http://foo/bar> <http://foo/baz2>} log:implies
{<http://foo/baz2> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://foo
/Range2>}.

# Proof found for http://www.w3.org/2000/10/rdf-tests/rdfcore/rdfs-domain-and-ra
nge/test004.nt in 268 steps (26797 steps/sec)
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

in this mess of output, I hope you see the :RULE5 and :RULE6
which refer to the numbers in the MT table

--
Jos

ps for the RDF entailment we use the -core option and that produces
  "simply the intermediate subgraph of the (merge of the) antecedent(s) which
   generalizes to the consequent graph"

Received on Saturday, 22 September 2001 09:36:45 UTC