Hello Peter, You're right, the RDF/XML can be pretty-printed like turtle. However N3 is human-friendly so there's a reason to improve the result when it does not affect speed whereas RDF/XML is not for human eyes anyway so I did not care about it.
If needed then it can be improved. However, big RDF/XML documents are supposed to be compressed for transfer and the length of the compressed document will stay basically the same, because pretty-printing has as much repeating strings as current output (even if they'll be longer). Best Regards, Ivan Mikhailov, OpenLink Software http://virtuoso.openlinksw.com On Tue, 2008-07-15 at 20:11 -0700, Peter Ansell wrote: > I can kind of understand why Virtuoso does a simple line by line dump from > SQL row results to RDF/XML files for CONSTRUCT queries. I don't however > understand why it would be less efficient for the same algorithm that > currently generates N3 Turtle results can't be ported almost directly over to > have a simpler RDF/XML result format, ie, start with rdf:Description > rdf:about="" the same way that N3 does, and then nest one level in with the > rest of the results. and start another rdf:Description if a different subject > was encountered.
