Re: [sca-bindings] ISSUE 11: "Formal" WSDL generation is unclear,ambiguous, and incomplete

[Eric Johnson <eric@tibco.com>]

Logged as: http://www.osoa.org/jira/browse/BINDINGS-11

Eric Johnson wrote:
> TARGET: Web Services Binding, draft 20070925, section 4.5
>
> DESCRIPTION:
> The WSDL generation as described in section 4.5 is incomplete, ambiguous and unclear.
>
> Examples of unclear meaning:
>  * Use of italics throughout the "formal" definition means different things in different cases.  In some cases, it is meant as a "human" variable substitution, and in some cases it appears to be an actual XPath expression.
>
> Examples of ambiguous meaning:
>  * Generated WSDL "import" statement doesn't specify what "SCA service interface location" actually means.  Is it a relative URL, an absolute URL, or how do you choose?
>  * Use of {(what to do ) ... } construct for repetition or conditionals is a convenient short-hand, but isn't explained in this specification, or by reference to any other specification that might have defined such a use before.
>
> Examples of incompleteness:
>  * The interface for the service may be described using WSDL 2.0.  No accommodation can be found here for such a scenario.
>  * If the component is constrained by a constraining type, the WSDL portType associated with the SCA service is not actually the one that should be published for the WSDL service binding.
>
> PROPOSAL:
>
> (This proposal is incomplete, as the issues are numerous)
>
> Complete redo the formal definition, using an actual language, such as XSLT 1.0, XSLT 2.0 or XQuery 1.0.  Start with an assumption that the component in question has been fully "normalized", that is all interfaces and services are declared on a component, not just implied by the component.
>
> (To be determined - how to deal with the incomplete issues - perhaps they should be broken out into a separate issue?)
>
>
>