RE: simple case of IRIs for Components in WSDL 2.0

Dan,

1. Yes, you constructed the URI correctly.

2. The parentheses are a result of our decision to be compliant with the 
XPointer Framework. We thought that was the W3C endorsed way to contruct 
fragment identifiers. It also solved our problem of how to support QNames 
since the xmlns scheme was already defined.

3. The wsdl. prefix was added in response to a previous comment. The 
commenter thought we were grabbing too many schemes so we added the 
prefix.

4. The two URI cases support different use cases. The one you are 
interested in is what we refer to as "component designators" which are 
just identifers for components. The other use case is for pointing to 
parts of a document. For example, if a Web browser supported our IANA 
application/wsdl+xml media type registration then you could view parts of 
a WSDL 2.0 document via our fragment identifier syntax. The two use cases 
collapse into one if you adopt the practice of locating your WSDL 2.0 
document at its own namespace (i.e. as opposed to putting a RDDL document 
there).

Arthur Ryman,
IBM Software Group, Rational Division

blog: http://ryman.eclipsedevelopersjournal.com/
phone: +1-905-413-3077, TL 969-3077
assistant: +1-905-413-2411, TL 969-2411
fax: +1-905-413-4920, TL 969-4920
mobile: +1-416-939-5063, text: 4169395063@fido.ca



Dan Connolly <connolly@w3.org> 
Sent by: public-ws-desc-comments-request@w3.org
09/27/2005 01:18 PM

To
Jonathan Marsh <jmarsh@microsoft.com>
cc
David Orchard <dorchard@bea.com>, Arthur Ryman/Toronto/IBM@IBMCA, Bijan 
Parsia <bparsia@isr.umd.edu>, "Henry S. Thompson" <ht@inf.ed.ac.uk>, 
public-ws-desc-comments@w3.org
Subject
RE: simple case of IRIs for Components in WSDL 2.0







On further review of the WSDL 2.0 spec, I found
a form of identifier that's much closer to what I want:

| http://example.org/TicketAgent.wsdl20#wsdl.interface(TicketAgent)

 -- http://www.w3.org/TR/2005/WD-wsdl20-20050803/#wsdl-iri-references

That seems to be just a function of the target namespace,
the interface name, a qualifier "wsdl.interface" and some
punctuation.

The other examples, along with the inclusion of
the ".wsdl20" extension in the target namespace URI
led me to believe that I needed to include the address
of a WSDL document in the IRI, not just the target
namespace name.

The () punctuation means that such IRIs cannot be abbreviated
with QNames in RDF/XML syntax. That's a royal pain, so I
hope you'll re-consider it. But it might be acceptable;
IRIs can be written out long-hand in RDF syntaxes.

And the "wsdl.interface" qualifier is clearly redundant
in the case of the SparqlQuery interface, so I hope you'll
consider making it optional too. But it might be
acceptable.

Can you confirm that this URI...
http://www.w3.org/2005/08/sparql-protocol-query/#wsdl.interface(SparqlQuery)


refers to the interface described by the following?

<description xmlns="http://www.w3.org/2005/08/wsdl"
  ...
  xmlns:tns="http://www.w3.org/2005/08/sparql-protocol-query/#"
  targetNamespace="http://www.w3.org/2005/08/sparql-protocol-query/#">

  <documentation>
    This document describes the SPARQL Protocol for RDF as a web
    service with one interface, SparqlQuery, containing one operation,
    query; as welll as HTTP and SOAP bindings of that interface. See
    http://www.w3.org/2001/sw/DataAccess/proto-wd/ for the SPARQL
    Protocol for RDF specification.
  </documentation>

  <interface name="SparqlQuery"
styleDefault="http://www.w3.org/2005/08/wsdl/style/iri">

  ...
  </interface>
</description>


By the way, the WSDL 2 spec says

| There are two main cases for WSDL 2.0 IRIs:
|
| * the IRI of a WSDL 2.0 document
|
| * the IRI of a WSDL 2.0 namespace

but that doesn't appeal to me at all. The main case for a WSDL IRI
is to refer to things described in WSDL, i.e. interfaces and such.

Also, regarding...

| The scheme names all begin with the prefix "wsdl." to avoid name
| conflicts with other schemes.

that seems odd. The risk of XPointer scheme collision is managed
by a registry, no? You might change that to say that they're
prefixed with wsdl. for mnemonic reasons. Or you might just
get rid of the wsdl. prefix.


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E