RE: Review of WSDL 2.0 Pt 3 Last Call WD

The Working Group is still deliberating two of your comments, but the
others have been resolved as indicated below.  We'll assume you agree
with our resolutions unless we hear from you by Oct 21st.

> -----Original Message-----
> From: public-ws-desc-comments-request@w3.org [mailto:public-ws-desc-
> comments-request@w3.org] On Behalf Of Marc Hadley
> Sent: Thursday, September 09, 2004 7:19 AM
> To: public-ws-desc-comments@w3.org
> Subject: Fwd: Review of WSDL 2.0 Pt 3 Last Call WD
> 
> 
> Forwarding on behalf of the XML Protocol WG.
> 
> regards,
> Marc.
> 
> Begin forwarded message:
> 
> > Resent-From: w3c-xml-protocol-wg@w3.org
> > From: Marc Hadley <Marc.Hadley@Sun.COM>
> > Date: September 6, 2004 10:30:24 PM EDT
> > To: "'w3c-xml-protocol-wg@w3.org'" <w3c-xml-protocol-wg@w3.org>
> > Subject: Review of WSDL 2.0 Pt 3 Last Call WD
> >
> >
> > I have an action to review WSDL 2.0 Part 3[1]. Here are my comments.
> >
> > Marc.
> >
> > [1] http://www.w3.org/TR/2004/WD-wsdl20-bindings-20040803/
> >
> > - Section 2 WSDL SOAP Binding
> >
> > "Notice that there are no default binding rules defined for
Interface
> > Fault components by this binding. Thus, if a given Interface
component
> > has any Fault components, then such Interface components MUST be
bound
> > via Binding components which indicate a specific interface and
contain
> > as many Binding Fault components as there are Fault components in
the
> > Interface Fault component."
> >
> > Comment: It's unclear why default binding rules are not defined for
> > fault components. The additional information that would be required
is
> > minimal and this feature would be useful.

[See http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC29a]

The rationale for omitting default binding rules for fault components is
that there is no good default value for code and subcode, and thus
defaults have negligible value.  We agree to add this rationale to the
spec.

See the latest editors draft to verify the fix;
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-bindings.
html?content-type=text/html;%20charset=utf-8

> > - Section 2.6 Declaring SOAP Modules
> >
> > Comment: The relationship between SOAP Modules declared in the
binding
> > and features declared in the interface is unclear. From the SOAP 1.2
> > Rec (section 3.3): "A  SOAP module realizes zero or more SOAP
> > features". I would expect a similar relationship in WSDL such that a
> > SOAP module in a binding would reference one or more features in the
> > interface, a module being the binding of those features to the
> > protocol.

[See http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC29b]

We're still working on this one.

> > - Section 3 WSDL HTTP Binding
> >
> > "Notice that there are no default binding rules defined for Fault
> > components by this binding."
> >
> > Comment: Similar to the case for the SOAP binding, it's unclear why
> > the extra step was not taken to define such default binding rules.

[See http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC29c]

Similar to the resolution for your first comment above, there is no good
default value for the HTTP status code.  We resolved it as we did for
the SOAP case, by adding a similar rationale to the HTTP binding.

See the latest editors draft to verify the fix;
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-bindings.
html?content-type=text/html;%20charset=utf-8

> > - Section 3.3 Default Binding Rules
> >
> > "Mechanisms that are outside the scope of this specification MAY
> > modify the serialization format of the instance data corresponding
to
> > the output message. An example of such modification is the
combination
> > of the serialization as application/x-www-form-urlencoded and the
> > SOAP-Response Message Exchange Pattern ([SOAP 1.2 Part 2: Adjuncts],
> > Section 6.3)."
> >
> > Comment: More detail required here, it's not clear what the example
is
> > trying to illustrate.

[See http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC29d]

We're still working on this one.

> > - Section 3.8.1.1 Case of elements cited in whttp:location attribute
> >
> > "When constructing the request URI, each pair of curly braces (and
> > enclosed element name) is replaced by the corresponding content of
the
> > element."
> >
> > Comment: What is the expected behavior when the element is not
present
> > in the instance data or is nil. The rules in section 3.9.1 appear to
> > allow both things to happen.

[See http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC29e]

The Working Group agrees to clarify the behavior in this case.  We will
make it an error for instance documents to have elements with
xsi:nil=true; or for elements to have nil values when trying to
auto-generate parameters.  The spec already says what happens if the
element is not present in the instance data.

> > - Section 3.8.3 Serialization as "multipart/form-data"
> >
> > Comment: What is the expected serialization of an element that is
nil
> > in the instance data ? An empty part, an omitted part ?

[See http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC29f]

We agree to add a similar clarification for this case as well.

See the latest editors draft to verify the fix;
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-bindings.
html?content-type=text/html;%20charset=utf-8

> > - Section 3.9 Operation Styles
> >
> > Comment: The relationship between the operation styles defined in
this
> > section to the serializations defined in section 3.8 would be
> > clarified by listing the permitted serializations for each style.
> > Currently the serializations list the allowed style but not the
other
> > way round. Consider switching the order of 3.8 and 3.9.

[See http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC29g]

We delegated this to the editors as editorial, and they resolved it.
See the latest editors draft to verify the fix;
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-bindings.
html?content-type=text/html;%20charset=utf-8

> > Comment: Should there be a style defined to match the serialization
> > defined in 3.8.2 ?

[See http://www.w3.org/2002/ws/desc/4/lc-issues/issues.html#LC29h]

The Working Group was unable to find a compelling need to define such a
style.  We made no change to the spec in this case.

Thanks again for your comments, and we hope to resolve the remaining two
open issues promptly.

Received on Thursday, 7 October 2004 21:53:27 UTC