ACTION: Ralph ask Eric Prud'hommeaux if there's a clever way to write a variant of tests 34-36 without @rel [recorded in http://www.w3.org/2007/08/30-rdfa-minutes.html#action09] The objective is to test that the implementation does *not* produce triples (e.g. in the case of test 34') relating <http://sw-app.org/mic.xhtml#i> to <http://sw-app.org/img/mic_2007_01.jpg> when the @rel attribute is omitted. Eric suggests we could use ASK WHERE { <http://sw-app.org/mic.xhtml#i> ?p <http://sw-app.org/img/mic_2007_01.jpg> . } and check that the answer is 0. (Eric actually suggested the more general 'ASK WHERE { ?s ?p ?o }', but I pointed out that we'd decided it was OK if an implementation produced more triples than the RDFa specification demands, though I don't think we've characterized expected limitations on any such 'extra' triples.) Eric's suggestion would require us to add to our test manifest a property for each test indicating whether the 'pass' result is 1 (as all the current tests are) or 0.Received on Thursday, 30 August 2007 19:20:42 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:15:09 GMT