W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2009

Re: Review of Service Description document

From: Gregory Williams <greg@evilfunhouse.com>
Date: Wed, 23 Dec 2009 14:53:55 -0800
Cc: W3C SPARQL Working Group <public-rdf-dawg@w3.org>
Message-Id: <947EEB96-6200-4AE9-B7BE-93E0967AA373@evilfunhouse.com>
To: Alexandre Passant <Alexandre.Passant@deri.org>
On Dec 22, 2009, at 2:54 AM, Alexandre Passant wrote:

> * Title is "Service Description" but the document alternates "Service Description" and "Service Descriptions", i.e. singular and plural forms. That should be fixed before publication
> 
> * B1 must be removed from the T.O.C.

OK

> * I'd like the vocab to be online at http://www.w3.org/ns/sparql-service-description# before publishing the document (OK to postpone if it takes too long to get that URI available)

OK, I'll clean up the RDF version and look into what it will take to get it published.

> * Section 2: 
> "This service description should be made available in an RDF serialization, but may also be provided embedded in HTML by RDFa, or other RDF representations by using content negotiation."
> Since RDFa is a RDF serialization, I'd avoid using "but" and replace by (however, not being a native speaker, I may be wrong here)
> "This service description should be made available in an RDF serialization, and may be provided embedded in HTML by RDFa, or other RDF representations by using content negotiation"

Good point.

> * Section 3.2.3:
> More details are welcome (but OK to leave as is for the current publishing)

Agreed. I'll try to add some descriptive text.

> * Section 3.3:
> Some instances names start with a lowercase, should be better to use uppercase here (as done in the entailment URIs)

Is this common for instances? I'm familiar with common practice regarding classes and properties, but not with similar practices for instances. I actually prefer the lowercase, but will change it if there's agreement on this point.

> * Section 3.3.3:
> "sd:dereferenceURLs" shouldn't it be "sd:dereferenceURIs" (URL / URI) ?

Won't they necessarily be URLs if they can be dereferenced?

> * 3.4.8 - sd:supportedLanguage
> Current range is sd:Language which means that people can use any sd:Language instance here, while there are some provided by the vocabulary to comply with the current SPARQL spec., i.e. sd:SPARQLQuery and sd:SPARQLUpdate. While this will be more restrictive, it could be better to have the range of sd:supportedLanguage limited to these two instances. (OK to postpone if it requires discussions that cannot be addressed quickly)

Not sure about this. Somebody (ericP?) had an example at some point of defining a custom language as comprising the base sparql query language with a specific set of extensions and functions. Let's discuss this on a future call.

> * Section 3.4:
> For consistency, as for each class (in 3.3), you mention they are instances of rdfs:Class, property descriptions should mention that they are instances of DatatypeProperty / ObjectProperty.

I'm hesitant to add lots of OWL terms, but could probably be convinced. I notice that there's already one OWL statement in declaring sd:url an IFP.

> BTW, as a matter of personal taste, instead having these as sentences, esp. for long ones as for "sd:defaultEntailmentRegime is an rdfs:subPropertyOf sd:feature. The rdfs:domain of sd:defaultEntailmentRegime is sd:Service. The rdfs:range of sd:defaultEntailmentRegime is sd:EntailmentRegime.", it might be better to have RDF code for each of these descriptions.

OK. Are there any specs that do this that I could draw on for styling? Inline turtle?

> * Section 4: 
> www.example -> www.example.org
> http://example/ -> http://www.example.org/ (appear 3 times)
> at the URL http://www.example/sparql/ -> at the URL http://www.example.org/sparql/

I used "www.example" to be consistent with the examples used in the SPARQL (1.0) Protocol for RDF document. I can change this if you think it's important, but if it's just a preference I might keep it the way it is for the sake of consistency.

> http://www.example/named-graph/
> It's imo a bit weird to have a graph ending with a / but I guess there is no constraints about that ?

I'll change that. I have no preference here.

> * References:
> Ref to voiD should also be added.

OK. I'll add it and mention it in describing the example.

thanks,
.greg
Received on Wednesday, 23 December 2009 22:54:24 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:40 GMT