W3C home > Mailing lists > Public > www-ws-desc@w3.org > May 2002

Re: proposed issue: Support for non-named-operation oriented portTypes

From: Scott Hinkelman <srh@us.ibm.com>
Date: Thu, 23 May 2002 09:08:39 -0500
To: www-ws-desc@w3.org
Message-ID: <OF1F5BA92C.E6390976-ON86256BC2.004C01BE@raleigh.ibm.com>
I have come to a mind place that I believe that the RPC model and Document
model are so fundamentally different that the abstraction description level
of services should provide for clear distinction of such.

At the abstraction-level, we need to move toward having "groups" of the
current abstract messages as the common denominator across both models, and
utilize <operation name=> as a clear distinguishing element for an RPC
using what the masses tend to view as it's semantic -a "method". Document
interactions have no intentionality on the service interface and thus no
concern over intention overloading ("method overloading" as most see it) on
the service interface. Referencing the message groups from bindings can be
enabled via standard XML document mechanisms.

Comments?

Scott
______________
Scott Hinkelman, Senior Software Engineer
XML Industry Enablement
IBM e-business Standards Strategy
512-823-8097 (TL 793-8097) (Cell: 512-415-8490)
srh@us.ibm.com, Fax: 512-838-1074




                                                                                                                                                     
                      Scott                                                                                                                          
                      Hinkelman/Austin/        To:       "Sanjiva Weerawarana" <sanjiva@watson.ibm.com>                                              
                      IBM@IBMUS                cc:       www-ws-desc@w3.org                                                                          
                      Sent by:                 Subject:  Re: proposed issue: Support for non-named-operation oriented portTypes                      
                      www-ws-desc-reque                                                                                                              
                      st@w3.org                                                                                                                      
                                                                                                                                                     
                                                                                                                                                     
                      05/22/2002 10:20                                                                                                               
                      AM                                                                                                                             
                                                                                                                                                     
                                                                                                                                                     



Hi Sanjiva,
I wish I felt I was reading too much into the meaning of <operation>, but
do not. Clearly the spec indicates "Operation? an abstract description of
an action supported by the service.", and there are many wide spread
perceptions that it is intended to represent intentionality. Take the
thread/issue in this forum " Issue: should WSDL allow overloaded methods?"
-I don't even know what a "method" actually is in WSDL, but the discussion
refers to WSDL Operations. Beyond this, the spec does also refers to the 4
primitives as "operations" (which I would suggest we call 'interactions').

That said, make no mistake that I agree with you on the primary purpose of
<operation> at the abstraction level to allow reference from the
realization (concrete binding) layer to the associated <message>s.

However, I hold there is significant misconception and needed clarity on
this at a minimum, but more is needed. Consider a vertical consortium which
has no intention to formally standardize the realization layer and are
faced with defining <operation name=>s while reading WSDL 1.1. I assert
that it is far from clear that the primary purpose is to provide reference
targets that could be used in bindings.

I think we need an issue on this, if at this point it is clarification
only, but ultimately we need to support portTypes that clearly have no
action semantics in the abstract layer (not the case yet in my mind).

thanks,

Scott
______________
Scott Hinkelman, Senior Software Engineer
XML Industry Enablement
IBM e-business Standards Strategy
512-823-8097 (TL 793-8097) (Cell: 512-415-8490)
srh@us.ibm.com, Fax: 512-838-1074





                      "Sanjiva

                      Weerawarana"             To:       Scott
Hinkelman/Austin/IBM@IBMUS

                      <sanjiva@watson.i        cc:
<www-ws-desc@w3.org>

                      bm.com>                  Subject:  Re: proposed
issue: Support for non-named-operation oriented portTypes


                      05/21/2002 03:52

                      PM






Hi Scott,

I think you're reading too much into the operation "name." In
WSDL 1.1, while you're required to provide a name at the time
of defining the portType, there is NO requirement that that
translate into something on the wire. In particular, I believe
for doc/lit style bindings the operation name simply does not
apply.

The name exists primarily to allow bindings to refer to specific
operations so that binding information may be given. Without a
name, there's no way to refer to a particular operation.

Bye,

Sanjiva.

----- Original Message -----
From: "Scott Hinkelman" <srh@us.ibm.com>
To: "Sanjiva Weerawarana" <sanjiva@watson.ibm.com>
Cc: <www-ws-desc@w3.org>; <www-ws-desc-request@w3.org>
Sent: Friday, May 17, 2002 8:11 PM
Subject: proposed issue: Support for non-named-operation oriented portTypes


> Hello,
> I would like to propose an issue on the requirement for all portTypes to
be
> named-operation oriented.
>
> There are significant efforts in several industry/semi-vertical standards
> bodies that define business payload content through structures where all
> message intentionality is defined directly as part of the payload. Taking
> the view that the WSDL name attribute on operation is intended to reflect
> the intentionality of the message (which I believe is the popular view)
> fundamentally presents core model difference and appears can only result
in
> unnatural representations at best using WSDL. In my mind, named
operations
> on a Service is a key aspect to Service Orientation, but there is
> significant momentum in industry groups operating outside this
orientation
> aspect.
>
> I believe this to be a significant hurdle of WSDL adoption for groups
> operating under this type of design, where all message intentionality is
an
> intricate part of the payload structure.
>
> I'm not sure of the details for issue submission and thank you for this
> consideration.
>
> Scott
> ______________
> Scott Hinkelman, Senior Software Engineer
> XML Industry Enablement
> IBM e-business Standards Strategy
> 512-823-8097 (TL 793-8097) (Cell: 512-415-8490)
> srh@us.ibm.com, Fax: 512-838-1074
Received on Thursday, 23 May 2002 10:10:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:20 GMT