ACTION-109 completed: Review for Service Description

Review of http://kasei.us/2009/09/sparql/sd.txt

If properly formatted and the usual boilerplate texts are added, no
objection to publish as FPWD. Some more details below.

==========================================

1) suggested rewording:
"but may also be provided in HTML or other representations by using
content negotiation, RDFa, or other approach."

-> "but may also be provided embedded in HTML by RDFa, or other
    RDF representations by using content negotiation."

(I think it should be made clear that we mean RDF representations only)

2) The SPARQL Service Description namespace URI is: @@ "URI" or "IRI"?

@@ http://www.w3.org/2009/sparql/service-description#

guess you should also mention

"The prefix used in this document for this namespace is sd:"

3) I'd suggest to add a property

   sd:endpointURL

along the lines we discussed
  http://lists.w3.org/Archives/Public/public-rdf-dawg/2009JulSep/0214.html
which could point to the "standard endpoint URL" for the service at
  hand (to have it available, even if accessed e.g. by a redirect, in
  which case <.> would be a different URL).

4) Extensibility/linking to other vocabularies for service/dataset
descriptions should be mentioned (e.g. http://rdfs.org/ns/void/html
could be used as an example.). We shall make clear the purpose of this
vocabulary to be minimal.

5) For formatting, to make your life easier, you could use something
like neologism [1] to generate a skeleton of the vocabulary  
description of the
sd: base vocab, should be fairly straightforward.

6) *This last point is optional/not related to FPWD but rather an  
issue for
further discussion*:

"sd:datasetDescription
	domain: sd:Service

Relates a service to a description of its available dataset. Depending  
on the implementation, this may describe the default dataset available  
when no explicit dataset is specified, or may describe graphs availble  
for explicit inclusion in the dataset via protocol or query level  
features."

In this context, as you point that out, I didn't see a feature in the
existing vocabularies to decribe which named graphs are in the  
endpoint, but it might make sense to have a properties in
this description to point to the standard default graph and to the
list of the named graphs in the dataset, something like...

   sd:hasNamedGraph
      range: rdfs:Resource  (uri/name of a named graph in the default  
dataset)

   maybe analogously, a property

   sd:hasDefaultGraph (likewise  range: rdfs:Resource)  would make  
sense that points to a name of the default graph, but that's probably  
a different issue.
?


1. http://neologism.deri.ie/

-- 
Dr. Axel Polleres
Digital Enterprise Research Institute, National University of Ireland,  
Galway
email: axel.polleres@deri.org  url: http://www.polleres.net/

Received on Sunday, 4 October 2009 19:51:46 UTC