- From: Jeremy Carroll <jeremy@topquadrant.com>
- Date: Tue, 23 Aug 2011 16:11:35 -0700
- To: public-rdf-wg@w3.org
On 8/23/2011 10:29 AM, Alan Wu wrote: > Hi Gavin, > > If that file is claimed to be N-TRIPLE. I would say it is a bug ;) > Zhe, I agree there is a bug, but not the one that you think. The RDF Test Cases document explicitly does not *recommend* ntriples, it explains the format used to write the test cases. Ntriples is only recommended for the less than a 1000 files published as part of the test cases. So the use of the file ending .nt is merely a conventional setting, as a non-standard communication about the file format, and the file is served saying it is in UTF-8 not ASCII, and several clients understand it. Since there is no standard, there is no conformance bug. An Ntriples reader is compliant if it correctly reads the published RDF test cases. Thus a reader that had a simple look up from the individual file names to the correct RDF graph would comply. Equally a reader that always uses UTF-8 is compliant. An ntriples document, other than one of the RDF test cases published by the W3C is not compliant, since Ntriples is explicitly not recommended. So the bug is publishing anything in NTriples, except for the handful of documents created by the RDF Core WG Jeremy
Received on Tuesday, 23 August 2011 23:13:34 UTC