- From: Dimitris Kontokostas <kontokostas@informatik.uni-leipzig.de>
- Date: Fri, 30 Oct 2015 10:03:15 +0200
- To: Holger Knublauch <holger@topquadrant.com>
- Cc: public-data-shapes-wg <public-data-shapes-wg@w3.org>
- Message-ID: <CA+u4+a3mPM05PL-RomPmbFwd4EisXM8Z-K+VfcVzn+FCKuxXyA@mail.gmail.com>
Arguments and annotations have different scopes but if aligning is necessary I would go this way: We can drop sh:annotationVarName and use a similar NCName mechanism in sh:annotationValue. The value of sh:annotationValue can be an IRI or a literal. If the value is an IRI and the NCName of the IRI matches a SPARQL variable (or a JS variable if supported later) the values for this result annotation are taken from the SPARQL Query. If the NCName of the IRI does not match a SPARQL variable or the value is a literal then the value is taken as constant and copied on the results. In RDFUnit I have only rut:annotationValue but prepend '?' in front of a sparql variable to make the parsing easier without the need of the sparql query. For example - static IRI: https://github.com/AKSW/RDFUnit/blob/master/rdfunit-core/src/main/resources/org/aksw/rdfunit/configuration/autoGeneratorsOWL.ttl#L221-L224 - SPARQL Var: https://github.com/AKSW/RDFUnit/blob/master/rdfunit-core/src/main/resources/org/aksw/rdfunit/configuration/autoGeneratorsOWL.ttl#L221-L224 - Literals: https://github.com/AKSW/RDFUnit/blob/master/rdfunit-core/src/main/resources/org/aksw/rdfunit/configuration/patterns.ttl#L917-L921 (in this case the value is a literal when the pattern bindings are applied) On Fri, Oct 30, 2015 at 7:22 AM, Holger Knublauch <holger@topquadrant.com> wrote: > One way of resolving this would be to drop sh:annotationVarName and rely > on sh:annotationProperty only, using the same NCName mechanism as used > elsewhere. This would slightly limit the use cases, because variable and > property names would have to be always aligned. > > I have no strong opinion, and this feature was mainly driven by Dimitris, > so I would like him to make a proposal. > > Holger > > > > On 10/30/2015 6:51, Arthur Ryman wrote: > >> +1 >> >> Also, the definition of the local name of results in an NCName. This >> is safe as a SPARQL variable. I assume it is safe enough for >> Javascript. >> >> -- Arthur >> >> On Wed, Oct 28, 2015 at 10:58 PM, RDF Data Shapes Working Group Issue >> Tracker <sysbot+tracker@w3.org> wrote: >> >>> shapes-ISSUE-107 (annotations v. arguments): annotations and arguments >>> use different mechanisms for specifying the SPARQL variable name [SHACL >>> Spec] >>> >>> http://www.w3.org/2014/data-shapes/track/issues/107 >>> >>> Raised by: Peter Patel-Schneider >>> On product: SHACL Spec >>> >>> Annotation properties use sh:annotationVarName to provide the SPARQL >>> variable name to use. Arguments use the local part of their IRI. It would >>> be better to have one mechanism. >>> >>> >>> >>> > > -- Dimitris Kontokostas Department of Computer Science, University of Leipzig & DBpedia Association Events: http://wiki.dbpedia.org/meetings/California2015 (Nov 5th) Projects: http://dbpedia.org, http://rdfunit.aksw.org, http:// http://aligned-project.eu Homepage:http://aksw.org/DimitrisKontokostas Research Group: http://aksw.org
Received on Friday, 30 October 2015 08:04:11 UTC