Re: GRDDL and named graphs

On 09/11/2014 07:37 PM, Richard Smith wrote:
>
> Is it still considered good practice to GRDDL as a way of extracting RDF
> from custom XML formats?  I notice the linked data platform (LDP)
> candidate recommendation doesn't makes no mention of GRDDL.

GRDDL does not actually do the transformation from XML to RDF.  It 
merely specifies a standard way to *associate* a transformation with XML 
documents, so that a GRDDL-aware processor can find that transformation 
when processing the XML.  But you still have to write your own 
transformation somehow, such as with XSLT or something else.

>
> My XML files logically map to an RDF dataset with multiple named
> graphs.  Obviously I can't express that in RDF/XML, so I'd need to use
> something like N-Quad or TriG to serialise the RDF corresponding to my
> XML.  However generating these formats is difficult in XSLT because of
> the need to escape characters in literals and URIs.  Is there a better
> way of achieving this?

We currently have a gap in the RDF-related standards, in that there is 
no standard way to transform XML to RDF.  It would be really helpful if 
there were a standard mapping analogous to the Direct Mapping for 
mapping relational data to RDF:
http://www.w3.org/TR/rdb-direct-mapping/
Similarly, the "CSV on the Web" working group is standardizing ways to 
convert CSV (or TSV) files to RDF, and hopefully they will also define 
something analogous to the Direct Mapping for relational data.

Some things you may want to look at:

  - Gloze, by Steve Battle, and part of the Jena tool set, can transform 
between XML and RDF given an XML Schema:
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.88.8929

  - XSPARQL
http://xsparql.deri.org/

  - TopQuadrant's tools can import XML into RDF.
http://www.topquadrant.com/

This is not an exhaustive list.  These are just the three that first 
came to my mind.

Other readers: please list other suggestions.

David

Received on Friday, 12 September 2014 00:28:07 UTC