Re: Comment on POWDER WDR, 2008-06-30: should use rdf:ID

Ivan Herman wrote:
> Comment on http://www.w3.org/TR/2008/WD-powder-dr-20080630/

Thank you for taking the time to do this Ivan.

> 
> This may be a bug in the GRDDL or simply in the document.

There is no GRDDL yet so it's a mistake in the doc ;-) (the GRDDL will 
do what the doc says. Work has begun on it but all the examples in all 
our docs were created by hand)

> 
> In example 2.11 powder2.rdf cannot refer to a blank node defined by 
> powder1.rdf, which it currently does by
> 
> <rdfs:subClassOf rdf:resource="powder1.rdf#ks_1"
> 
> where ks_1 is an rdf:nodeID value in powser1.rdf.
> 
> I suspect the intention is that all xml:id values are translated into 
> rdf:ID, which makes them referenceable. Ie, in the powder1.rdf file:
> 
> 
> - rdf:nodeID="ks_1" should be rdf:ID="ks_1"
> - rdf:nodeID="ks_2" should be rdf:ID="ks_2"

True. Yes, the idea is that if the POWDER doc includes an xml:ID then 
this is translated in POWDER-S as an rdf:ID.
> 
> Note, however, that if a mechanical xml:id -> rdf:ID mechanism is used, 
> that will affect example 2-14 (as well as the argument stated somewhere 
> at the beginning that the usage of blank node means that nobody from the 
> 'outside' can massage with those). It is up to the WG to decide whether 
> this is an issue or not. Maybe, for the purpose of 2-14, there should be 
> two different identification mechanisms: xml:id for a universally 
> visible ID and something like 'id' for internal purposes only (ie, 
> introducing blank nodes without talking about them:-)

Nice idea. We will cogitate on this.

> 
> Ivan
> 
> P.S. Editorial comment: it is very handy that you had line numbers added 
> to all previous examples. Why don't you do that for all?

Ah well, I guess that comes down to sheer idleness. We did it where the 
text needed to refer to specific lines but not where it didn't. maybe we 
should add them for the next version.

Phil.

Received on Thursday, 3 July 2008 12:56:22 UTC