From: Mike Dean (mdean@bbn.com)
Date: 06/25/03
Following up on yesterday's discussion, I developed some possible test cases [1] for DAML/OWL Rules combining ontology and rules features in ways that I expect most users will want to do. [2] is one of the examples we discussed yesterday. I expect the most common uses will be like: http://www.daml.org/2003/06/ruletests/uncle-1.n3 (property chaining) http://www.daml.org/2003/06/ruletests/translation-2.n3 (translation) Some tests come from my daily work: http://www.daml.org/2003/06/ruletests/translation-1.n3 http://www.daml.org/2003/06/ruletests/domesticflight-1.n3 The test cases are written in N3 [3], because it's relatively easy to read and an engine (cwm) exists that can already handle that syntax. cwm successfully processes many of the current tests, and could probably handle all of them if taught a little more about OWL (I think that the version I'm using only knows about RDF(S) and DAML+OIL). Ascending test numbers in a sequence generally represent more complicated tests. I'm soliciting the following feedback: 1) Is this a useful technique for grounding our discussions? 2) Are there any of these tests that we wouldn't expect DAML/OWL Rules to handle? 3) More tests, e.g. a realistic case that goes beyond Description Logic Programs. 4) Any errors in the tests :-( I haven't yet gone back and looked at the tests for RDF(S), OWL, or cwm. If we decide that this approach is useful, we should probably leverage this previous experience with automated testing, reporting, etc. Thanks! Mike [1] http://www.daml.org/2003/06/ruletests/ [2] http://www.daml.org/2003/06/ruletests/equivalence-3.n3 [3] http://www.w3.org/2000/10/swap/doc/
This archive was generated by hypermail 2.1.4 : 06/25/03 EST