Re: clearing up CONSTRUCT's capabilities

Kendall Clark wrote:
> Dogs,
> 
> I know I should be working on the protocol stuff (and I am), but I want to
> suggest some small, clarificatory changes to the language on CONSTRUCT.  The
> way the language reads right now
> 
>   The CONSTRUCT result form returns a single RDF graph specified by a graph
>   template. The result is an RDF graph formed by taking each query solution
>   in the solution sequence, substituting for the variables into the graph
>   template and combining the triples into a single RDF graph by set union.
> 
> This is admirably clear, and I have no problem with it. However, it doesn't
> say anything about ground triples in the graph template and, thus, in the
> result graph created by CONSTRUCT. I've spoken with Andy about this, and we
> agree that the spec doesn't say anything to contract ground triples in the
> graph template.
> 
> However, it doesn't say anything explicitly about this capability, which I
> think is pretty useful. I'd like an example that focuses on this, so I'm
> contributing the following text that I'd like to see put into rq23:
> 
>   The graph template may also contain ground or explicit triples, that is,
>   triples with no variables, and these also appear in the RDF graph returned
>   by the CONSTRUCT query form.

Thanks for the text - I belive the definition of template covers this because it 
refers to triple patterns which do not need to have variables but if adding the 
text makes it cleaer, I can do.

>   
> I think it would fit nicely as the last sentence of this paragraph (from 10.3):
> 
>   If any such instantiation produces a triple containing an unbound
>   variable, or an illegal RDF construct (such as a literal in subject or
>   predicate position), then that triple is not included in the RDF graph,
>   and a warning may be generated.
> 
> The first example in 10.3 can be modified thus to illustrate the ground
> triple capability:
> 
>   @prefix  foaf:  <http://xmlns.com/foaf/0.1/> .
>   
>   _:a    foaf:name   "Alice" .
>   _:a    foaf:mbox   <mailto:alice@example.org> .
>   
>   PREFIX foaf:    <http://xmlns.com/foaf/0.1/>
>   PREFIX vcard:   <http://www.w3.org/2001/vcard-rdf/3.0#>
>   CONSTRUCT   { <http://example.org/person#Alice> vcard:FN ?name;
>                 <http://exampl.org/persona#Ted> vcard:FN "Ted"^^en-us}
>   WHERE       { ?x foaf:name ?name }
>   
>   creates vcard properties from the FOAF information:
> 
>   @prefix vcard: <http://www.w3.org/2001/vcard-rdf/3.0#> .
>   <http://example.org/person#Alice> vcard:FN "Alice" .
>   <http://example.org/person#Ted> vcard:FN "Ted"^^en-us .

Not convinced by this example : firstly, the query pattern is about finding 
names and adding a fixed one is a bit confusing.  Secondly, first examplkes 
shouldn't be too overloaded.

I choose vcard out because, for a simple first example, I didn't want any blank 
node on output issues.  If vcard had an RDFS class, it could introduce that but 
there isn't one in the vocabulary.

How about writing this up as a test case instead?  It could be a "real worlkd" 
example based on your full RSS usage and the needs of a simple first example 
don't apply.

	Andy

> 
> One of my use cases for CONSTRUCT is to be able to create URIs which are
> standing queries, the output of which is an RSS 1.0 channel, or some other
> well-known RDF file, say, FOAF. Having ground triples in the graph template
> is crucial. I'd just like the spec to be more explicit about this capability.
> 
> Cheers, 
> Kendall
>

Received on Friday, 5 August 2005 15:46:47 UTC