Review of WS-Chor

I had an AI to review the WS-Choreography Description Language Version
1.0. 
                     (http://www.w3.org/TR/2004/WD-ws-cdl-10-20041217/)
I have grouped my comments under 3 categories:
A. Comments from the point of view of impact of WS-CDL on the
WS-Addressing specification
B. Comments on the WS-CDL specification generally, i.e. independent of
WS-Addressing considerations
C. Comments of editorial nature

For the purposes of this WG's requirement to review WS-CDL, I am listing
only category A comments for discussion by this WG.  (I will send
categories B & C comments directly to the WS-CDL Working Group unless
anyone from this WS-Addressing WG wants to see them.)

----------------------------

A. Comments in relation to WS-Addressing specification

1. The WS-CDL model consists of various entities, two of which, roughly
speaking, correspond to WS-Addressing's information models for Endpoint
References and for Message Addressing Properties.
   The two are Channels and Interactions:
   A Channel "realizes a point of collaboration between parties by
specifying where and how information is exchanged."
   An Interaction "results in an exchange of information between parties
and possible synchronization of their observable information changes and
the actual values of the exchanged information."

However, the "Choreography Description Language is not an "executable
business process description language" or an implementation language."
It is at an abstract level, providing "a contract containing a "global"
definition of the common ordering conditions and constraints under which
messages are exchanged."  It is at a different level from WS-Addressing
which is closer to the implementation level.  Hence there is no direct
impact on the WS-Addressing specification or vice-versa.


2. WS-CDL intentionally uses abstractions to avoid being coupled tightly
to specifics.  For example, it uses Information Types to "describe the
type of information used within a Choreography. By introducing this
abstraction, a Choreography definition avoids referencing directly the
data types, as defined within a WSDL document or an XML Schema
document."

Similarly, it uses a Token to "reference a document fragment within a
Choreography definition", and Token Locator to "provide a query
mechanism to select them. By introducing these abstractions, a
Choreography definition avoids depending on specific message types, as
described by WSDL, or a specific query string, as specified by XPATH.
Instead the document part and the query string can change without
affecting the Choreography definition."


3. Having said the above, there remains the issue of mapping (or
binding) from the higher abtract level of WS-CDL to the lower
WS-Addressing "implementation"-level.  The questions are:
a) Should mapping/binding of WS-CDL concepts to WS-Addressing concepts
be standardized?
b) If so, where should this mapping work be done?  WS-CDL or
WS-Addressing or a third Working Group?



4. In "Section 7 Relationship with the Addressing framework" of WS-CDL,
an old version of WS-Addressing specification Abstract is quoted.  It
should be updated to the most current.

The old version is:
The WS-Addressing specification [WSAD] provides transport-neutral
mechanisms to address Web services and messages, specifically, XML
[XML], [XMLNS] elements to identify Web service endpoints and to secure
end-to-end endpoint identification in messages. WS-Addressing enables
messaging systems to support message transmission through networks that
include processing nodes such as endpoint managers, firewalls, and
gateways in a transport-neutral manner.

The new version is:
Web Services Addressing provides transport-neutral mechanisms to address
Web services and messages. Web Services Addressing 1.0 - Core defines a
set of abstract properties and an XML Infoset [XML Information Set]
representation thereof to identify Web service endpoints and to
facilitate end-to-end identification of endpoints in messages. The
specification enables messaging systems to support message transmission
through networks that include processing nodes such as endpoint
managers, firewalls, and gateways in a transport-neutral manner.


5. "Section 7 Relationship with the Addressing framework" of WS-CDL
says:
"WS-Addressing can be used to convey the reference and correlation
information for normalizing expanded Channel Variable information into
an uniform format that can be processed independently of transport or
application." 

This is consistent with the conclusion stated in point 1.


6. "Section 7 Relationship with the Addressing framework" of WS-CDL
further says:
"The WS-Addressing specification is in progress and the WS-Choreography
Working Group will review and comment on developments in this effort on
an ongoing basis."

Big Brother is watching.

-----------------------

Please comment on comments above.

Yin Leng

Received on Friday, 14 January 2005 14:02:21 UTC