On 16 June 2017 at 13:43, Svensson, Lars <L.Svensson@dnb.de> wrote: > On Friday, June 16, 2017 1:50 PM, Dan Brickley [mailto:danbri@google.com] > wrote: > > > On the Google side regarding dataset discovery and > > https://developers.google.com/search/docs/data-types/datasets we're > currently > > consuming basic DCAT (if expressed in json-ld, rdfa, or microdata) as > alongside > > Schema.org dataset descriptions in any of those. My sense is that much > of the > > vocabulary needed to improve the state of things is already around > (CSVW, Data Cube, > > DC, DCAT, Schema.org, SKOS, SOSA, QUDT, etc.), but there's a need to > figure out > > ways of combining them. So the likes of https://www.w3.org/TR/shacl/ > may prove > > useful for capturing such patterns (as well as simple examples / > documentation of > > course). There is also btw some potential via JSON-LD context files to > hide some multi- > > namespace complexity via indirection, since the @context can map a > seemingly flat set > > of terms into several different "behind the scenes" namespaces. > > Sounds like a use case for profiles to me... > Yes, same ideas, and some common history eg see http://dcpapers.dublincore.org/pubs/article/download/866/862 and https://www.w3.org/2001/sw/wiki/images/0/00/SimpleApplication-SpecificConstraintsforRDFModels.pdf although I suspect the JSON-LD @context angle has been under-explored so far. JSON-LD is a little less appealing for IoT scenarios than for dataset discovery because it makes (graph-extraction) parsing dependent on the contents of a remote file, and IoT often needs to operate offline and without leaking signals to the outside world. But that's a discussion for elsewhere. cheers, Dan
Received on Friday, 16 June 2017 12:52:29 UTC