- From: Tom De Nies <tom.denies@ugent.be>
- Date: Mon, 22 Apr 2013 10:54:09 +0200
- To: Provenance Working Group <public-prov-wg@w3.org>
- Message-ID: <CA+=hbbcZWZQFPW6yjXTDBi_M0=GkUqttYcR53JdsXvUvPhyEKQ@mail.gmail.com>
This issue is now closed 2013/4/17 Provenance Working Group Issue Tracker <sysbot+tracker@w3.org> > PROV-ISSUE-668 (TomDN): Disambiguated DBPedia URIs to QNames in XML > example [PROV-DICTIONARY] > > http://www.w3.org/2011/prov/track/issues/668 > > Raised by: Tom De Nies > On product: PROV-DICTIONARY > > For the XML example of PROV-Dictionary we ran into some trouble when > validating against the schema > Since we use real-world dbpedia URIs as example, there were some that > could not be turned into a valid QName due to its character restrictions. > Therefore, we had to find a workaround. The workaround is to add a valid > NCName (derived by removing the invalid characters) as a fragment, and > split the URI after the # symbol. The first part of the URI is then used to > override the dbpedia: prefix in the element that uses the URI as an > identifier, and the valid NCName is used as the LocalPart of the QName. > This is in conformance with the guidance given in the the RDF/XML Syntax > Specification. > The only thing that raises some doubt is the use of the nonexistent > fragment, but since the URI still resolves, this is worth it in my opinion. > We added a remark explaining this construct to the end of the example. > > https://dvcs.w3.org/hg/prov/raw-file/default/dictionary/releases/NOTE-prov-dictionary-20130430/Overview.html#example-prov-xml > > If anyone has any objections to this method, please let us know by e-mail > or on the call. The alternative to this method is to leave the invalid > characters out, and end up with non-existent URIs, which in my view, is > worse than non-existent fragments. > > ISSUE is marked pending review until a decision is made to accept/reject > this solution during the telecon on April 18th. > > > >
Received on Monday, 22 April 2013 08:54:37 UTC