Andy thanks, I have been thinking over your idea and I think it is brilliant. Instead of using the graphs, I will reify my mappings using one URL per mapping resolving to the reified statement and the provenance information.
Many thanks Alexandra P.S. If you are interested, I can post the example RDF as well. On Sat, Jun 16, 2018 at 5:26 PM, Andy Seaborne <a...@apache.org> wrote: > > > On 14/06/18 13:02, Alexandra Kokkinaki wrote: > >> Thanks all for your prompt responses, >> >> TriX does exactly what I would like to do, but Martynas suggestion is what >> I ideally need. >> I will describe what we have to do, and any suggestions or ideas are more >> than welcome: >> >> We currently host 250 vocabularies with more than 60.000 terms used by >> many >> users all over the world. (so we have to cater for backwards >> compatibility, >> as many software apps are based on our vocabularies). >> The vocabularies are exposed as Linked Data, via sparql endpoint (Jena >> Fuseki) and REST interface (RDF/XML sadly no other serialization) using >> mostly SKOS. >> Users asked more information about the provenance of term mappings >> (triples) and this is where the named graph comes in. >> How do I display the provenance information in RDF/XML? >> > > One thought - publish the provenance in a separate graph to the > vocabularies. The provenance graph can refer to the other graphs. > > If it is the provenance of specific triples within graphs may come down to > using reification. Quads/named graphs can be used (the whole vocabulary > being the union of the named graphs and the named graphs being the > fragments from different sources. > > Andy > > > >> >> Many thanks, >> Alexandra >> >> >> >> >> >> >> >> On Wed, Jun 13, 2018 at 10:37 PM, Andy Seaborne <a...@apache.org> wrote: >> >> Alexandra, >>> >>> I was wondering what use such a format would have for you - if it is a >>> non-standard format, no-one else can read it. There are non-XML formats >>> that capture named graphs that are standard - JSON-LD, TriG, and N-Quads. >>> >>> Andy >>> >>> >>> On 13/06/18 13:38, Alexandra Kokkinaki wrote: >>> >>> Dear all, >>>> >>>> I also want to use named graphs to capture provenance information on >>>> triples but in RDF/XML serialization. >>>> Iam replying Martynas email, as he first asked for that, wondering if >>>> anything happened since then, and I haven't found it? >>>> >>>> Many thanks >>>> Alexandra >>>> >>>> On Thu, Jun 9, 2016 at 4:58 PM, Martynas Jusevičius < >>>> marty...@atomgraph.com> >>>> wrote: >>>> >>>> I found the "abandoned" discussion on RDF 1.1 WG wiki: >>>> >>>>> https://www.w3.org/2011/rdf-wg/wiki/TF-RDF-XML#Change_8:_ >>>>> named_graph_support_in_RDF-XML >>>>> >>>>> On Thu, Jun 9, 2016 at 5:26 PM, Martynas Jusevičius >>>>> <marty...@atomgraph.com> wrote: >>>>> >>>>> It seems that someone has thought about this before: >>>>>> https://www.w3.org/Submission/rdfsource/ >>>>>> >>>>>> TriX is just not a natural structure for XSLT transformations. >>>>>> >>>>>> On Thu, Jun 9, 2016 at 4:34 PM, Andy Seaborne <a...@apache.org> >>>>>> wrote: >>>>>> >>>>>> >>>>>>> >>>>>>> On 09/06/16 14:47, Martynas Jusevičius wrote: >>>>>>> >>>>>>> >>>>>>>> Good points. Yes TriG-like structure makes more sense -- but then it >>>>>>>> is clearly non-standard. >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> That's a good thing - no risk of wrong data or missing data. >>>>>>> >>>>>>> Using an attribute, rdfx:graph - won't it be a property if it is not >>>>>>> understood as additional syntax attribute? >>>>>>> >>>>>>> >>>>>>> Isn't this a gap in RDF standardization -- an XML format for quads? >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> IIRC When it came down to it, no one was interested in spending time >>>>>>> on >>>>>>> >>>>>>> it. >>>>>> >>>>> >>>>> >>>>>> There are (probably) some notes in the RDF 1.1 WG wiki. >>>>>>> >>>>>>> TriX is a de facto standard. >>>>>>> >>>>>>> Andy >>>>>>> >>>>>>> >>>>>>> >>>>>>> On Thu, Jun 9, 2016 at 1:51 PM, Andy Seaborne <a...@apache.org> >>>>>>>> wrote: >>>>>>>> >>>>>>>> >>>>>>>>> On 08/06/16 15:22, Martynas Jusevičius wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>>> >>>>>>>>>> Hey, >>>>>>>>>> >>>>>>>>>> would it be possible to adopt RDF/XML writer for quads (Dataset)? >>>>>>>>>> >>>>>>>>>> What >>>>>>>>> >>>>>>>> >>>>> would that take? >>>>>> >>>>>>> >>>>>>>>>> I know it would involve a non-standard syntax, but if we used >>>>>>>>>> namespaced attributes, XML-compatible tools shouldn't break. >>>>>>>>>> >>>>>>>>>> I am thinking it should be possible to add an attribute (e.g. >>>>>>>>>> rdfx:graph) with graph name on each of the property elements, >>>>>>>>>> something like this: >>>>>>>>>> >>>>>>>>>> <rdf:Description >>>>>>>>>> rdf:about="https://www.w3.org/People/Berners-Lee/card#i"> >>>>>>>>>> <foaf:givenName >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> rdfx:graph="https://www.w3.org/People/Berners-Lee/card"> >>>>>>>>>> >>>>>>>>>> Tim</foaf:givenName> >>>>>>>>> >>>>>>>> >>>>> <foaf:familyName >>>>>> >>>>>>> >>>>>>>>>> >>>>>>>>>> rdfx:graph=http://data.semanticweb.org/person/tim-berners-lee/rdf >>>>>>>>>> >>>>>>>>>> "">Berners-Lee</foaf:familyName> >>>>>>>>> >>>>>>>> >>>>> </rdf:Description> >>>>>> >>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> And if a triple is in 2 graphs? The default graph? >>>>>>>>> >>>>>>>>> >>>>>>>>> What do you think? Would someone else be interested in such >>>>>>>>>> serialization? I know there is TriX, but it is not convenient for >>>>>>>>>> >>>>>>>>>> XSLT >>>>>>>>> >>>>>>>> >>>>> transformation. >>>>>> >>>>>>> >>>>>>>>>> Martynas >>>>>>>>>> atomgraph.com >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> An alternative is more TriG like : >>>>>>>>> >>>>>>>>> >>>>>>>>> <rdfx:Graph rdfx:name="..."> >>>>>>>>> .... RDF/XML here ... >>>>>>>>> </rdfx:Graph> >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> >>>>>>> >>>>> >>>> >>