Re: [RT] rethinking the need for inferencing

From: Leigh Dodds <ldodds_at_ingenta.com>
Date: Tue, 07 Mar 2006 12:20:11 +0000

Danny Ayers wrote:

>> 1) generation (anything to XML)
>> 2) transformation (XML to XML)
>> 3) serialization (XML to anything)
>
> anything to RDF :
> worst case, do anything to XML, apply XSLT
>
> RDF to RDF :
> SPARQL CONSTRUCT
>
> RDF to anything :
> SPARQL SELECT + XSLT on XML result set

Or XSLT+extensions. For heavily nested XML one needs
multiple, embedded SPARQL SELECTS (master-detail queries).

For me RDF rules ARE a transformation mechanism. Whether
one mixes in the inferred triples with the originals
is a configuration item.

L.
Received on Tue Mar 07 2006 - 12:18:38 EST