W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > July to September 2005

Re: Experience with SPARQL/P/SOAP/Axis

From: Kendall Clark <kendall@monkeyfist.com>
Date: Wed, 27 Jul 2005 14:47:06 -0400
To: DAWG Mailing List <public-rdf-dawg@w3.org>
Message-ID: <20050727184706.GA22385@monkeyfist.com>

On Mon, Jul 18, 2005 at 03:51:46PM +0100, Seaborne, Andy wrote:

> WSDL 1.1. file attached, including a possible SOAP binding.

Hmm. IIRC, WSDL 1.1 was never more than a W3C Note from MS & IBM. I don't
see how we can use it normatively at all. When I go to W3C site, and when I
ask web service experts in my lab, I get pointers to WSDL 2. There aren't
many tools yet, of course.

> Summary:
> + Removing the rdf-dataset level and merging the elements simplifies

Good suggestion, which I've incorporated. The editor's draft of the protocol
spec and the WXS file agree with rq23 about the cardinality of rdf-dataset
elements. 

> + Request for HTTP POST binding

Yep, on my TODO list for a while. I put an initial bit of binding into the
WSDL file for this.

> + Minor problem with defining the encoding for HTTP returns

Yep. I don't know what to do about this. RDF/XML is XML, so I think the
application/xml type isn't wrong, but it's not the most specific type...

> 2/ We should change "default-graph-uri" to be maxOccurs="unbounded" to align
> with the WG description about multiple FROM URIs.

Yep, I did this over a month ago, but forgot to update the public copy.
Oops.

> 3/ I needed to define rdf:RDF, not just import the namespace.  I defined it 
> as
> an element with no further details about it.  JAXB happy (Java-XML binding).

I haven't done this yet and am not convinced it's needed.

> 5/ Please include a whttp:methodDefault="POST" binding with an encoding of
> application/x-www-form-urlencoded (which is not the default). 

Done.

I also added a query-request-refused fault type, which gets serialized in
HTTP as a 500 status code (which is the most general server-side fault code
in HTTP, and the only one generic enough to work here, I think).

> 6/ I was not sure what the {http output serialization} should be because it
> isn't only application/xml for result sets but also application/rdf+xml for 
> a graph.

Yep, I have to reread the WSDL2 HTTP binding stuff to see if there's
anything we can do here. I said "application/xml" for now. 

Do we say anywhere that the RDF returned by CONSTRUCT query has to be
RDF/XML? 

Kendall Clark
Received on Wednesday, 27 July 2005 18:49:07 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:23 GMT