rdf processing steps

My last note [1] was overly pessimistic.  While I do believe it is 
overly restrictive to require XML to flow between processing steps, this 
does not prevent a useful amount of rdf processing to be specified.

I submit the following steps for the WG to consider including as 
optional steps in the XProc 1.0 specification.  I think you should seek 
technical assistance from the Semantic Web Deployment WG to design the 
signatures for these steps.

<p:rdf-rewrite>
<!-- rewrite RDF input in specified output serialization syntax -->
<!-- allow input from external source -->
<!-- output to port as RDF/XML, or to external sink in some other
      notation -->
<!-- could be static error if implementation doesn't handle specified
      notation -->
</p:rdf-rewrite>

<p:sparql>
<!-- run sparql query against zero or more input RDF graphs -->
<!-- allow in-line definition of sparql query -->
<!-- output stream to be RDF/XML or sparql results XML -->
<!-- specify sparql version -->
</p:sparql>

<p:rdfa>
<!-- emit RDF/XML from RDFa markup in XML input -->
<!-- specify version of RDFa to glean -->
</p:rdfa>

<p:grddl>
<!-- apply GRDDL transformation to input, emit RDF/XML -->
<!-- allow xproc to specify grddl:transformation list to append to
      or replace transformations specified in source -->
<!-- specify version of GRDDL to use -->
</p:grddl>

<p:rdf-parse>
<!-- parse RDF input and emit error stream and status -->
</p:rdf-parse>

rdf-parse might be dispensable, but it could be a useful guard step.

Later,
--Paul

[1] 
http://lists.w3.org/Archives/Public/public-xml-processing-model-comments/2008Aug/0021.html

Received on Tuesday, 26 August 2008 11:28:28 UTC