- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Fri, 26 Mar 2004 13:57:28 +0000
- To: James Carlyle <james.carlyle@takepart.com>
- Cc: www-rdf-interest@w3.org
Hmmm ... James Carlyle wrote: > Hi, > > I want to rework my RDF XML-->N-Triples XSLT stylesheet for the XML > syntax recommendation [1]. Previously I tested the output against the > test suite N-Triples by loading both sets into a relational database and > running intersection queries, but now I want to consider a different > approach. > > 1) Convert the test suite N-Triples to Trix documents e.g. one trix doc containing all the tests > 2) Convert the XSLT stylesheet output N-Triples to Trix documents > 3) Write another XSLT stylesheet to compare the test suite and generated > Trix sets. > > So my questions are: > 1) Can you see a flaw in this approach? Yes, see 3) below > 2) Has anyone written a tool to generate Trix from N-Triples? Not me. > 3) Has anyone written a Trix-Trix comparison tool or stylesheet? This would be hard, since Trix-Trix comparison is not yet defined. This is one of the things currently being discussed on www-archive as some of us try and improve the formal model underneath Trix. More positively, jena.rdfcompare [1] can read two N-triples files and tell you whether they are the same or not. It implements graph equivalence as defined by RDF Concepts. I would be surprised if other toolkits do not provide similar functionality. Hence I would suggest running the test suite thus: XSLT from RDF/XML => N-triples use jena.rdfcompare to ompare the answer with the correct result. Jeremy [1] http://cvs.sourceforge.net/viewcvs.py/jena/jena2/src/jena/rdfcompare.java
Received on Friday, 26 March 2004 09:01:53 UTC