Re: Local names may now start with a digit?

* Ivan Herman <ivan@w3.org> [2012-06-28 17:52+0200]
> Note that RDFa has the same feature. In the case of RDFa, the request came (a fairly long time ago, actually) from, eg, IPTC[1] that has internal data structures that denote individual items by numbers. This is the same with, say, geonames URI-s, if my memory serves me well, but also in a number of library cataloguing services. In all those cases the 'natural' prefix:localname' structure would lead to a numeric-only local name.
> 
> Ivan
> 
> [1] http://www.iptc.org/site/Home/
> On Jun 28, 2012, at 17:48 , Gavin Carothers wrote:
> 
> > On Thu, Jun 28, 2012 at 8:41 AM, David Booth <david@dbooth.org> wrote:
> >> I notice that the Turtle grammar now permits a local name to begin with
> >> a digit, such as foo:1234 or _:5678 .
> >> http://www.w3.org/TR/turtle/#term-turtle2-PN_LOCAL
> >> As far as I remember, this is a divergence from RDF/XML, because XML
> >> local names are not permitted to start with a digit.  When I check the
> >> RDF WG list of resolutions I see no mention of this change, so
> >> 
> >> I am wondering about the rationale for this change, but I see no mention
> >> of it in the RDF WG list of resolutions:
> >> http://demo.3roundstones.net/rdf/2012/rdfwg/resolutions.xhtml?view
> >> 
> >> Was this divergence discussed?  And if so, can someone please supply a
> >> pointer to the rationale?
> > 
> > http://www.w3.org/2011/rdf-wg/track/issues/1
> > 
> > Part of the alignment of Turtle with SPARQL 1.0 and SPARQL 1.1, both
> > of which allow for local names that start with digits. The RDF data
> > model has always allowed for any IRI to be used, this in fact meant
> > that there is valid RDF that cannot be expressed in RDF/XML.

Noting here for future reference, RDF/XML could potentially be extended to address this by inventing e.g. rdf:PropertyLabel:
  <http://www.w3.org/TR/REC-rdf-syntax/#section-grammar-productions>

6.1.2 Element Event
  + URI-property-label
    If the rdf:resource attribute is present, the value of that attribute,
    otherwise, URI-string-value of the element information item.

7.2.7 Production propertyAttributeURIs:
  s/rdf:Description/rdf:Description | rdf:LabeledElement | rdf:propertyLabel/

in each of (resourcePropertyElt, literalPropertyElt, parseTypeLiteralPropertyElt, parseTypeResourcePropertyElt, parseTypeCollectionPropertyElt, parseTypeOtherPropertyElt, emptyPropertyElt):

  s/e.URI-string-value/e.URI-property-label/


This would look like
    <rdf:Description rdf:ID="foo">
     <rdf:LabeledElement rdf:propertyLabel="http://bio2rdf.org/taxonomy:9606" rdf:resource="bar"/>
    </rdf:Description>
This says { <foo> http://bio2rdf.org/taxonomy:9606 <bar> }, which doesn't make much sense. I haven't come up with a good example 'cause the use cases for SPARQL allowing leading digits were all around subjects and objects of assertions, e.g. pubmed IDs, etc. The extension I propose here would likely not get used much.


It would be nice to s/URI/IRI/g.


> > Thank you for your comment. I hope that clears up your question.
> > 
> > Cheers,
> > Gavin
> >> 
> >> Thanks!
> >> 
> >> 
> >> --
> >> David Booth, Ph.D.
> >> http://dbooth.org/
> >> 
> >> Opinions expressed herein are those of the author and do not necessarily
> >> reflect those of his employer.
> >> 
> >> 
> > 
> 
> 
> ----
> Ivan Herman, W3C Semantic Web Activity Lead
> Home: http://www.w3.org/People/Ivan/
> mobile: +31-641044153
> FOAF: http://www.ivan-herman.net/foaf.rdf
> 
> 
> 
> 
> 
> 

-- 
-ericP

Received on Thursday, 28 June 2012 17:36:37 UTC