unclarity in adjuncts 5.9.6 - soap header block component fragid

[Jacek Kopecky]

Hi all,

finishing the implementation of the RDF mapping, I uncovered one more
small problem in the spec, this time in the adjuncts.

Section 5.9.6 defines the fragment ID for a SOAP header block component.
It says:

wsdl.extension(http://www.w3.org/ns/wsdl/soap,
        wsoap.header(parent/element declaration))

     1. parent is the pointer part of the {parent} component, as
        specified in WSDL Version 2.0 Part 1: Core Language.
     2. element declaration is the value of the {element declaration}
        property.

First, I expect that the "pointer part of the {parent} component" means
the wsdl.* xpointer part of the fragment identifier of the {parent}
component, without any xmlns parts which are moved before the
wsdl.extension part, right?

Second, point two needs to be clarified to say either that "element
declaration" is the QName that is the value of the {name} property of
the component that is the value of the {element declaration} property,
or that "element declaration" is the wsdl.elementDeclaration part of the
fragment identifier of the component referred to by the {element
declaration} property.

In other words, which is correct? I think A). I hope the formatting
survives in a useful form.

A.
xmlns(ns1=http://othernamespace.example/)
xmlns(ns2=http://schemanamespace.example/)
wsdl.extension(http://www.w3.org/ns/wsdl/soap,
               wsoap.header(wsdl.bindingFault(bindingname/ns1:faultname)
                            /ns2:element))

B.
xmlns(ns1=http://schemanamespace.example/)
wsdl.extension(http://www.w3.org/ns/wsdl/soap,
               wsoap.header(xmlns(ns1=http://othernamespace.example/)
                            wsdl.bindingFault(bindingname/ns1:faultname)
                            /ns1:element))

C.
xmlns(ns1=http://othernamespace.example/)
xmlns(ns2=http://schemanamespace.example/)
wsdl.extension(http://www.w3.org/ns/wsdl/soap,
               wsoap.header(wsdl.bindingFault(bindingname/ns1:faultname)
                            /wsdl.elementDeclaration(ns2:element)))

D.
xmlns(ns1=http://schemanamespace.example/)
wsdl.extension(http://www.w3.org/ns/wsdl/soap,
               wsoap.header(xmlns(ns1=http://othernamespace.example/)
                            wsdl.bindingFault(bindingname/ns1:faultname)
                            /wsdl.elementDeclaration(ns1:element)))


Since I'm asking for clarification, I believe it can be handled as
editorial, or it can be handled as errata, but I'd like to have an
answer before we go to Rec. 8-)

Jacek

[Arthur Ryman-2]

A looks right to be. All xmlns parts go first. The element should be named by its QName.

Arthur Ryman, PhD, AoT, DE
Process and Portfolio Management, Rational Division

phone: [phone], TL 969-3077
assistant: [phone] (T/L: 318-8867)
fax: [phone], TL 969-4920
mobile: [phone], text: 4169395063@...

Jacek Kopecky <jacek.kopecky@...>
Sent by: www-ws-desc-request@...

05/02/2007 02:11 PM

To	WS-Description WG <www-ws-desc@...>
cc
Subject	unclarity in adjuncts 5.9.6 - soap header block component fragid

Hi all,

finishing the implementation of the RDF mapping, I uncovered one more
small problem in the spec, this time in the adjuncts.

Section 5.9.6 defines the fragment ID for a SOAP header block component.
It says:

wsdl.extension(http://www.w3.org/ns/wsdl/soap,
       wsoap.header(parent/element declaration))

    1. parent is the pointer part of the {parent} component, as
       specified in WSDL Version 2.0 Part 1: Core Language.
    2. element declaration is the value of the {element declaration}
       property.

First, I expect that the "pointer part of the {parent} component" means
the wsdl.* xpointer part of the fragment identifier of the {parent}
component, without any xmlns parts which are moved before the
wsdl.extension part, right?

Second, point two needs to be clarified to say either that "element
declaration" is the QName that is the value of the {name} property of
the component that is the value of the {element declaration} property,
or that "element declaration" is the wsdl.elementDeclaration part of the
fragment identifier of the component referred to by the {element
declaration} property.

In other words, which is correct? I think A). I hope the formatting
survives in a useful form.

A.
xmlns(ns1=http://othernamespace.example/)
xmlns(ns2=http://schemanamespace.example/)
wsdl.extension(http://www.w3.org/ns/wsdl/soap,
              wsoap.header(wsdl.bindingFault(bindingname/ns1:faultname)
                           /ns2:element))

B.
xmlns(ns1=http://schemanamespace.example/)
wsdl.extension(http://www.w3.org/ns/wsdl/soap,
              wsoap.header(xmlns(ns1=http://othernamespace.example/)
                           wsdl.bindingFault(bindingname/ns1:faultname)
                           /ns1:element))

C.
xmlns(ns1=http://othernamespace.example/)
xmlns(ns2=http://schemanamespace.example/)
wsdl.extension(http://www.w3.org/ns/wsdl/soap,
              wsoap.header(wsdl.bindingFault(bindingname/ns1:faultname)
                           /wsdl.elementDeclaration(ns2:element)))

D.
xmlns(ns1=http://schemanamespace.example/)
wsdl.extension(http://www.w3.org/ns/wsdl/soap,
              wsoap.header(xmlns(ns1=http://othernamespace.example/)
                           wsdl.bindingFault(bindingname/ns1:faultname)
                           /wsdl.elementDeclaration(ns1:element)))


Since I'm asking for clarification, I believe it can be handled as
editorial, or it can be handled as errata, but I'd like to have an
answer before we go to Rec. 8-)

Jacek

[Jacek Kopecky]

Hi,

first, it's funny how the editors' drafts say Proposed Recommendation 21
March 2007, all the jazz with status of the doc etc.

I don't see a bug in our Bugzilla or that the spec would be clarified on
an editorial issue I sent earlier, so I'm resending it. It seems that
Arthur agrees the intention is A), so here's a proposed text for the
spec:

adjuncts, 5.9.6:
update the bullets like this:

     1. parent is the "wsdl.*" pointer part of the {parent} component,
        as specified in WSDL Version 2.0 Part 1: Core Language, i.e.
        without the xmlns() pointer parts.
     2. element declaration is the value of the {name} property of the
        Element Declaration component that is referred to by the
        {element declaration} property of the SOAP Header Block
        component.

I believe this will clarify the ambiguity.

Jacek

On Wed, 2007-05-02 at 20:11 +0200, Jacek Kopecky wrote:

> Hi all,
>
> finishing the implementation of the RDF mapping, I uncovered one more
> small problem in the spec, this time in the adjuncts.
>
> Section 5.9.6 defines the fragment ID for a SOAP header block component.
> It says:
>
> wsdl.extension(http://www.w3.org/ns/wsdl/soap,
>         wsoap.header(parent/element declaration))
>
>      1. parent is the pointer part of the {parent} component, as
>         specified in WSDL Version 2.0 Part 1: Core Language.
>      2. element declaration is the value of the {element declaration}
>         property.
>
> First, I expect that the "pointer part of the {parent} component" means
> the wsdl.* xpointer part of the fragment identifier of the {parent}
> component, without any xmlns parts which are moved before the
> wsdl.extension part, right?
>
> Second, point two needs to be clarified to say either that "element
> declaration" is the QName that is the value of the {name} property of
> the component that is the value of the {element declaration} property,
> or that "element declaration" is the wsdl.elementDeclaration part of the
> fragment identifier of the component referred to by the {element
> declaration} property.
>
> In other words, which is correct? I think A). I hope the formatting
> survives in a useful form.
>
> A.
> xmlns(ns1=http://othernamespace.example/)
> xmlns(ns2=http://schemanamespace.example/)
> wsdl.extension(http://www.w3.org/ns/wsdl/soap,
>                wsoap.header(wsdl.bindingFault(bindingname/ns1:faultname)
>                             /ns2:element))
>
> B.
> xmlns(ns1=http://schemanamespace.example/)
> wsdl.extension(http://www.w3.org/ns/wsdl/soap,
>                wsoap.header(xmlns(ns1=http://othernamespace.example/)
>                             wsdl.bindingFault(bindingname/ns1:faultname)
>                             /ns1:element))
>
> C.
> xmlns(ns1=http://othernamespace.example/)
> xmlns(ns2=http://schemanamespace.example/)
> wsdl.extension(http://www.w3.org/ns/wsdl/soap,
>                wsoap.header(wsdl.bindingFault(bindingname/ns1:faultname)
>                             /wsdl.elementDeclaration(ns2:element)))
>
> D.
> xmlns(ns1=http://schemanamespace.example/)
> wsdl.extension(http://www.w3.org/ns/wsdl/soap,
>                wsoap.header(xmlns(ns1=http://othernamespace.example/)
>                             wsdl.bindingFault(bindingname/ns1:faultname)
>                             /wsdl.elementDeclaration(ns1:element)))
>
>
> Since I'm asking for clarification, I believe it can be handled as
> editorial, or it can be handled as errata, but I'd like to have an
> answer before we go to Rec. 8-)
>
> Jacek

[Philippe Le Hegaret]

Jacek,

per Jonathan's message at
 http://lists.w3.org/Archives/Public/www-ws-desc/2007Jun/0001.html

I changed the draft as you requested.

Check:

http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-adjuncts.html?content-type=text/html;%20charset=utf-8#soap-headers-decl-fragid



Philippe



On Sat, 2007-05-19 at 18:58 +0200, Jacek Kopecky wrote:
> Hi,
>
> first, it's funny how the editors' drafts say Proposed Recommendation 21
> March 2007, all the jazz with status of the doc etc.

fixed :)

> I don't see a bug in our Bugzilla or that the spec would be clarified on
> an editorial issue I sent earlier, so I'm resending it. It seems that
> Arthur agrees the intention is A), so here's a proposed text for the
> spec:
>
> adjuncts, 5.9.6:
> update the bullets like this:
>
>      1. parent is the "wsdl.*" pointer part of the {parent} component,
>         as specified in WSDL Version 2.0 Part 1: Core Language, i.e.
>         without the xmlns() pointer parts.
>      2. element declaration is the value of the {name} property of the
>         Element Declaration component that is referred to by the
>         {element declaration} property of the SOAP Header Block
>         component.
>
> I believe this will clarify the ambiguity.
>
> Jacek
>
> On Wed, 2007-05-02 at 20:11 +0200, Jacek Kopecky wrote:
> > Hi all,
> >
> > finishing the implementation of the RDF mapping, I uncovered one more
> > small problem in the spec, this time in the adjuncts.
> >
> > Section 5.9.6 defines the fragment ID for a SOAP header block component.
> > It says:
> >
> > wsdl.extension(http://www.w3.org/ns/wsdl/soap,
> >         wsoap.header(parent/element declaration))
> >
> >      1. parent is the pointer part of the {parent} component, as
> >         specified in WSDL Version 2.0 Part 1: Core Language.
> >      2. element declaration is the value of the {element declaration}
> >         property.
> >
> > First, I expect that the "pointer part of the {parent} component" means
> > the wsdl.* xpointer part of the fragment identifier of the {parent}
> > component, without any xmlns parts which are moved before the
> > wsdl.extension part, right?
> >
> > Second, point two needs to be clarified to say either that "element
> > declaration" is the QName that is the value of the {name} property of
> > the component that is the value of the {element declaration} property,
> > or that "element declaration" is the wsdl.elementDeclaration part of the
> > fragment identifier of the component referred to by the {element
> > declaration} property.
> >
> > In other words, which is correct? I think A). I hope the formatting
> > survives in a useful form.
> >
> > A.
> > xmlns(ns1=http://othernamespace.example/)
> > xmlns(ns2=http://schemanamespace.example/)
> > wsdl.extension(http://www.w3.org/ns/wsdl/soap,
> >                wsoap.header(wsdl.bindingFault(bindingname/ns1:faultname)
> >                             /ns2:element))
> >
> > B.
> > xmlns(ns1=http://schemanamespace.example/)
> > wsdl.extension(http://www.w3.org/ns/wsdl/soap,
> >                wsoap.header(xmlns(ns1=http://othernamespace.example/)
> >                             wsdl.bindingFault(bindingname/ns1:faultname)
> >                             /ns1:element))
> >
> > C.
> > xmlns(ns1=http://othernamespace.example/)
> > xmlns(ns2=http://schemanamespace.example/)
> > wsdl.extension(http://www.w3.org/ns/wsdl/soap,
> >                wsoap.header(wsdl.bindingFault(bindingname/ns1:faultname)
> >                             /wsdl.elementDeclaration(ns2:element)))
> >
> > D.
> > xmlns(ns1=http://schemanamespace.example/)
> > wsdl.extension(http://www.w3.org/ns/wsdl/soap,
> >                wsoap.header(xmlns(ns1=http://othernamespace.example/)
> >                             wsdl.bindingFault(bindingname/ns1:faultname)
> >                             /wsdl.elementDeclaration(ns1:element)))
> >
> >
> > Since I'm asking for clarification, I believe it can be handled as
> > editorial, or it can be handled as errata, but I'd like to have an
> > answer before we go to Rec. 8-)
> >
> > Jacek
>

[Jacek Kopecky]

Checked, looks good. Thanks! 8-)
Jacek

On Fri, 2007-06-22 at 21:50 +0000, Philippe Le Hegaret wrote:

> Jacek,
>
> per Jonathan's message at
>  http://lists.w3.org/Archives/Public/www-ws-desc/2007Jun/0001.html
>
> I changed the draft as you requested.
>
> Check:
>
> http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-adjuncts.html?content-type=text/html;%20charset=utf-8#soap-headers-decl-fragid
>
>
>
> Philippe