comments on wsdl1.2 WD

Some comments on the WSDL1.2 WD at [6].

Issue:

Use of the RFC2119 term OPTIONAL to designate cardinality of an element 
should be
avoided. Specifically, in [1] the wsdl:documentation element is listed as 
OPTIONAL.
Is it the intent of the WSDWG that this EII be truly OPTIONAL (to 
implement) or is
it the WG's intent that a wsdl:definitions EII have zero or one (or more) 
wsdl:documentation
EIIs in its [children] property? I believe that it is the latter. Suggest 
that the term
OPTIONAL be replaced with 'zero or one' (but see issue below w/r/t i18n).

Issue:

Use of the RFC2119 term OPTIONAL to designate cardinality of an element 
should be
avoided. Specifically, in [3] the wsdl:documentation element is listed as 
OPTIONAL.
Is it the intent of the WSDWG that this EII be truly OPTIONAL (to 
implement) or is
it the WG's intent that a wsdl:types EII have zero or one (or more) 
wsdl:documentation
EIIs in its [children] property? I believe that it is the latter. Suggest 
that the term
OPTIONAL be replaced with 'zero or one' (but see issue below w/r/t i18n).

Issue:

While wsdl:documentation EII permits zero or more AIIs in its [attributes] 
property,
it would seem wise to specifically include the xml:lang AII in the 
definition of the
defined AIIs for this EII. It would also seem wise to provide for multiple 
instances
of the wsdl:documentation EII in the [children] property of each of the 
WSDL
EIIs declared to support it,, each with a distinct value for the xml:lang 
AII so as to
provide for i18n of the documentation of a WSDL definition. I would also 
suggest that 
the xml:lang attribute be REQUIRED.

Issue:

There is no mention in [3] of wsdl:documentation as being a valid EII 
child
of the wsdl:message EII, yet the WSDL1.2 schema has the tMessage type
extending tExtensibleDocumented.

Issue:

There is no mention in [7] of wsdl:documentation as being a valid EII 
child
of the wsdl:portType EII, yet the WSDL1.2 schema has the tPortType type
extending tExtensibleDocumented.

***Actually, there is no mention of the wsdl:documentation EII as being a 
valid
child EII of any of the WSDL defined EIIs except the wsdl:definitions and 
wsdl:types
EIIs.***

Issue:

Use of the RFC2119 term OPTIONAL to designate presence or absence of an 
AII
should be avoided. Specifically, in [2] the type and element AIIs are 
specified as
OPTIONAL when I believe that the intent is that one of the two AIIs is in 
fact REQUIRED
as is further qualified in the prose. In order to avoid possible confusion 
and interop
issues related to these AIIs, suggest that the prose be changed to read:

A REQUIRED attribute information item which has:
A [local name] of type OR element
A [namespace name] which has no value.

Issue:

In [3], the wsdl:types EII is defined to allow for zero or more AIIs in 
the [attributes]
property without further qualification. Yet, in [4], the spec provides for 
zero or more
AIIs in the [attributes] property with the caveat that they have a 
[namespace name]
property OTHER than the WSDL namespace URI. I think that the statement in 
[3] is therefore
redunant and self contradictory and should be removed in deference to [4].

Issue:

In [1], the wsdl:documentation EII is listed before the wsdl:import EII 
yet in the
example WSDL document in [5], the wsdl:documentation element succeeds the
wsdl:import element. Suggest that all examples be schema validated before
they are included. In this case, the example in [5] is incorrect, the 
wsdl:documentation
element should precede the wsdl:import element.

Cheers,

Christopher Ferris
Architect, Emerging e-business Industry Architecture
email: chrisfer@us.ibm.com
phone: +1 508 234 3624

[1] http://www.w3.org/TR/wsdl12/#eii-definitions
[2] http://www.w3.org/TR/wsdl12/#eii-message
[3] http://www.w3.org/TR/wsdl12/#eii-types
[4] http://www.w3.org/TR/wsdl12/#language-extensibility
[5] http://www.w3.org/TR/wsdl12/#doc-structure
[6] http://www.w3.org/TR/2002/WD-wsdl12-20020709/
[7] http://www.w3.org/TR/2002/WD-wsdl12-20020709/#eii-portType