Re: [wsbpel] Issue - 289 - draft text to describe the relationshipXML Schema and Spec

[Alex Yiu <alex.yiu@oracle.com>]

Hi Thomas,

See inline ...

Thomas Schulze wrote:

Hi Alex,

in the second paragraph, maybe we can add a reference to the static analysis table? In the appendix for the static analysis table, there's a reference to the schema appendix, too.
  

[AYIU]:
I am open to the idea of having a reference to SA table.

For the first paragraph, regarding extensibility, there are forward references in the spec to the schemas regarding the details, see chapter
5.3:

"WS-BPEL supports extensibility by allowing namespace-qualified attributes to appear on any WS-BPEL element and by allowing elements from other namespaces to appear within WS-BPEL defined elements. This is allowed in the XML Schema specifications for WS-BPEL."

Means, the schema is more specific regarding the modeled extension points than the spec. One example is: Where in the sequence is the extension element allowed to occur? The spec is silent about that, the schemas are clarifying that. Additionally there's the blockDefault="#all". 

[AYIU]:
Well ... that was the reason all along that I was voicing the concern against adding blockDefault="#all". The spec on this topic is silent, while XSD are not. This small scale asymmetry does not look that good. ... :-) ... Well ... water is under the bridge ...

If one wants to be the spec text very formal about XML syntax, we should use a bunch of XML Infoset term to describe the syntax (see WSDL 2.0 spec text or majority of WS-* spec). Oblivously, we are not using those terms unfortunately.

Maybe we can say something like "Regarding extensibility the schemas are more detailed than the spec. In this case the schemas will take precedence over the spec."?
  

[AYIU]:
IMHO, that is a bad practice to say one exceptional case where the schemas take precedence over the spec, while all other cases spec take precedences over schemas. Very confusing and difficult to enforce.

In JCP world, a javax package can provide extra normative details in additional to the spec, as long as those details does not contradict the spec text.

IMHO, I guess one of acceptable ways to add extra wordings to address your concern is something like:

XML Schema offers supplementary normative XML syntax details, such as where extension points are allowed in a WS-BPEL process definition, as long as those XML syntax details do not violate explicit normative descriptive text.

What do you think?


BTW, one thing to point out: XML Schema are not 100% portable enough across different XML Schema Processor implementation, especially when it comes to advanced part of the spec. In some cases, even Xerces-Java and Xerces-C++ (both are hosted under Apache and donated by the same company) behave differently. ;-) ... Using XML Schema is THE authority for XML syntax is not that bullet-proof either. In fact, I have personally faced an Xerces-Java (of certain vrsion) implementation issue about <xsd:any> handling.


Thanks!


Regards,
Alex Yiu

Best regards/Mit freundlichen Grüßen,

       Thomas Schulze



                                                                           
             Alex Yiu                                                      
             <alex.yiu@oracle.                                             
             com>                                                       To 
                                       Alex Yiu <alex.yiu@oracle.com>      
             09.08.2006 19:04                                           cc 
                                       wsbpeltc                            
                                       <wsbpel@lists.oasis-open.org>,      
                                       Martin Chapman                      
                                       <martin.chapman@oracle.com>         
                                                                   Subject 
                                       Re: [wsbpel] Issue - 289 - draft    
                                       text to describe the relationship   
                                       XML Schema and Spec                 
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           





Hi all,

Here is an updated proposed text (with more input from Martin)
---------------------
Where there is disagreement between the separate XML schema files, the
XML schemas in the appendices, any pseudo-schema in the descriptive
text, and the normative descriptive text, the normative descriptive text
will take precedence over the separate XML Schema files. The separate
XML Schema files take precedence over any pseudo-schema and over any XML
schema included in the appendices.

XML Schemas only enforce a subset of constraints described in the
normative descriptive text. Hence, a WS-BPEL artifact, such as a process
definition, can be valid according to the XML Schemas only but not valid
according to the normative descriptive text.
---------------------

Regards,
Alex Yiu


Alex Yiu wrote:

  

Hi all,

Here is the draft text to describe the relationship XML Schema and Spec
(so far ...)
---------------------
XML Schemas are normative artifacts derived from the normative
descriptive text in the specification. XML Schemas only enforce a
subset of constraints described in the normative descriptive text.
Hence, a WS-BPEL artifact, such as a process definition, can be valid
according to the XML Schemas but not valid by the normative
specification text. When there is a conflict between XML schemas and
the normative descriptive text, the normative descriptive text will
take precedence over XML Schemas.
---------------------


Regards,
Alex Yiu