- From: Luc Moreau <l.moreau@ecs.soton.ac.uk>
- Date: Fri, 01 Mar 2013 09:08:16 +0000
- To: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>
- CC: Stephan Zednik <zednis@rpi.edu>, W3C provenance WG <public-prov-wg@w3.org>
Hi Stian, On 03/01/2013 02:16 AM, Stian Soiland-Reyes wrote: > On Thu, Feb 28, 2013 at 5:26 PM, Luc Moreau <L.Moreau@ecs.soton.ac.uk> wrote: >> Thanks Stian. >> My preference would be to avoid these since in provdm the prefix should denote a uri. > I would still allow qnames without prefix as xmlns="somethinguseful" > is very commonly used in XML. QNames with unknown prefixes is > (already) not allowed. You suggest we should also require the XML > namespace for identifier qnames to be non-empty - is that true for any > prefix then? Yes absent prefix, but default namespace is also fine. I didn't state the rule in full. Also concatenating the uri denoted by prefix and local name should give a uri. See http://www.w3.org/TR/prov-dm/#term-qualified-name Luc > > In PROV-N, this is allowed, thanks to IRI_REF allowing empty strings > and relative IRI references. (Perhaps this was not intended?) > > document > default <> > prefix ex1 <> > prefix ex2 <../soup#> > entity(fred) > entity(ex1:fred) > entity(ex2:soup) > endDocument > > But with your suggestion I can only express ex2:soup in PROV-XML. > > > As a side note, PROV-N does not define how to resolve relative IRI > references according to the base IRI of the PROV-N. This might be > important to clarify inside bundles. (I think they should be resolved > according to the document resource's requested/known IRI, not the > bundle's identifier) > -- Professor Luc Moreau Electronics and Computer Science tel: +44 23 8059 4487 University of Southampton fax: +44 23 8059 2865 Southampton SO17 1BJ email: l.moreau@ecs.soton.ac.uk United Kingdom http://www.ecs.soton.ac.uk/~lavm
Received on Friday, 1 March 2013 09:08:53 UTC