rdf accommodation

I'm late to the game, so excuse me if this topic has already been 
disposed of.

Did the WG consider including some standard step definitions for RDF 
processing?  At minimum, an optional p:sparql analog to p:xslt seems 
reasonable.  Beyond that, some convenience definitions for simple graph 
modifications would enable RDF processing comparable to what is now 
provided for XML processing.

RDF processing steps can easily be defined and implemented by 3rd 
parties, using the existing framework.  But I expect there will be a 
proliferation of variant definitions that will impede portability of 
xproc scripts for RDF processing.

Perhaps the step definitions for RDF processing could be specified in a 
profile or module so that existing implementations will not be affected, 
and future implementations can choose whether to support them.

--Paul Tyson