Resent-From: sw-team@w3.org From: "Ralph R. Swick" Date: Thu Jun 19, 2003 10:49:22 US/Eastern To: sw-team@w3.org Subject: Re: ?bug in cwm's relative URI handling At 09:07 AM 6/3/2003 -0400, Ralph R. Swick wrote: http://lists.w3.org/Archives/Team/sw-team/2003Jun/0005.html I believe that cwm's handling of relative URIs is not correctly preserving base document URIs in its result. Here is another case, of a slightly different nature: Given a document named cwm_baseURI_test2.rdf with content [[ Ralph Also Ralph ]] The command % cwm --rdf cwm_baseURI_test2.rdf --n3 produces: [[ #Processed by Id: cwm.py,v 1.129 2003/04/08 16:12:43 timbl Exp # using base file:/tux/u2/swick/Projects/zakim/cwm_baseURI_test2.rdf # Notation3 generation by # notation3.py,v 1.138 2003/04/25 19:12:46 sandro Exp # Base was: file:/tux/u2/swick/Projects/zakim/cwm_baseURI_test2.rdf @prefix : <#> . @prefix dc: . @prefix rdf: . :obj1 dc:creator "Ralph" . :obj2 dc:creator "Also Ralph" . #ENDS ]] I get the same output if I use --n3=r for the output and also if I give a full file: URI for the input (matching what cwm reported as the base URI in the comment in its output), whereas I would like to see full URIs for file:/tux/u2/swick/Projects/zakim/cwm_baseURI_test2.rdf#obj1 and file:/tux/u2/swick/Projects/zakim/cwm_baseURI_test2.rdf#obj2 Furthermore, given the input [[ Ralph Also Ralph ]] the command % cwm --rdf cwm_baseURI_test2a.rdf --n3 produces [[ #Processed by Id: cwm.py,v 1.129 2003/04/08 16:12:43 timbl Exp # using base file:/tux/u2/swick/Projects/zakim/cwm_baseURI_test2a.rdf # Notation3 generation by # notation3.py,v 1.138 2003/04/25 19:12:46 sandro Exp # Base was: file:/tux/u2/swick/Projects/zakim/cwm_baseURI_test2a.rdf @prefix : . @prefix dc: . @prefix rdf: . :obj2 dc:creator "Also Ralph" . <#obj1> dc:creator "Ralph" . #ENDS ]] Which does provide the (equivalent of) the full URI for obj2 but something that seems entirely wrong for the name of obj1. I believe that these two inputs are equivalent (modulo the filename modified for test purposes) and that the two statements in test2a are describing objects whose names are identical but for the final character. Am I misunderstanding something, or is cwm? My need is to preserve the full uri of things identified using the RDF/XML 'about="#foo"' syntax in cwm's results.