- From: Ossi Nykänen <ossi.nykanen@tut.fi>
- Date: Tue, 08 Feb 2011 16:45:23 +0200
- To: www-rdf-validator@w3.org
Dear Receiver, The "Submit problem report" button in W3C RDF Validator page (http://www.w3.org/RDF/Validator/ARPServlet) does not seem to work (or then it just does not give feedback of working...), so I'm using my regular email client. Anyway, parsing... <rdf:RDF xmlns:n0="http://example.org/packages/vocab#" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:Description rdf:about="http://example.org/packages/X11"><n0:DistributionSite rdf:nodeID="GukCeVwQ2"/></rdf:Description><rdf:Description rdf:nodeID="GukCeVwQ2"><rdf:li rdf:resource="ftp://ftp1.example.org"/><rdf:_1 rdf:resource="ftp://ftp.example.org"/><rdf:type rdf:resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#Alt"/><rdf:li rdf:resource="ftp://ftp2.example.org"/></rdf:Description></rdf:RDF> gives triples... Number Subject Predicate Object 1 http://example.org/packages/X11 http://example.org/packages/vocab#DistributionSite genid:UGukCeVwQ2 2 genid:UGukCeVwQ2 http://www.w3.org/1999/02/22-rdf-syntax-ns#_1 ftp://ftp1.example.org 3 genid:UGukCeVwQ2 http://www.w3.org/1999/02/22-rdf-syntax-ns#_1 ftp://ftp.example.org 4 genid:UGukCeVwQ2 http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://www.w3.org/1999/02/22-rdf-syntax-ns#Alt 5 genid:UGukCeVwQ2 http://www.w3.org/1999/02/22-rdf-syntax-ns#_2 ftp://ftp2.example.org i.e. where the rdf:_1 membership property appears twice. I suspect this (Triples output) might be considered as an error. The informative rdf:Alt definition in the spec says that rdf:_1 is intended to be considered as the default or preferred value (see, e.g., the RDF Primer and RDF/XML syntax spec). In this case, the parse result invalidates this default (unique rdf:_1) information, even if no actual entailment is required. A possible correction would be to check whether rdf:_1 is explicitly defined in the model, and if so, start enumerating rdf:li's from 2 (seed value of the "li-counter"). Another (a better?) correction would be giving triples with a single rdf:_1, encoding others as rdf:li, i.e. not aiming to translate li's into numbers in the first place. However, the latter might be considered a violation against RDF Semantics since rdf:li is not recognised in the RDF Container Vocabulary (which is a bit weird, btw.). Of course, this might land somewhere in the grey area for RDF (and I do get similar parse output from, e.g., using RDFLib 3.0.0). To my understanding, however, RDF specs to not forbid mixing rdf:li and rdf:_1 properties in serialisations, which raises this issue. Since the W3C validator might be considered "normative", I simply thought reporting this might make sense. (The specs do not seem to define this.) If I am mistaken, I would not mind getting a brief note in email, thanks. Best regards, --Ossi -- Ossi Nykänen, Adjunct Professor, Dr.Tech. Tel +358 40 849 0730 Tampere University of Technology Fax +358 3 3115 3549 Institute of Mathematics, HLab http://matriisi.ee.tut.fi/hypermedia/en/ ossi.nykanen@tut.fi P.O. Box 553, FIN-33101 Tampere, Finland W3C Finnish Office, http://www.w3c.tut.fi ossi@w3.org -- ossi.nykanen@iki.fi, http://www.iki.fi/ossi.nykanen/
Received on Tuesday, 8 February 2011 14:46:33 UTC