RE: Issue 662: specify at least two levels or types of Choreography description

Mmmmm ... interesting email ...

Monica you said ...

>>>David, you continue to discuss abstract choreography definition, 
references to business analysts, and an approach that infers your 
audience are business people. I would purport that WS-CDL is focused on 
the developer, pi-calculus and WSDL, and placement as a mechanism to 
help manage BPEL processes.<<<

You're right, I do see WS-CDL as being usable, probably via some graphical tool, by the business analyst who then hands the design over to developers to implement. I think there is a real need for this type of formal language, and if it's not WS-CDL, then there is a big gap in the WS-xx space, IMHO that I don't think is filled elsewhere - is this correct?

I also agree that we need to resolve this issue, as we are obviously approaching this problem from fundamentally different perspectives which would lead to different outcomes.

David

-----Original Message-----
From: Monica J. Martin [mailto:Monica.Martin@Sun.COM]
Sent: Sunday, May 02, 2004 7:39 AM
To: Burdett, David
Cc: UCorda@SeeBeyond.com; tony.fletcher@choreology.com;
public-ws-chor@w3.org
Subject: Re: Issue 662: specify at least two levels or types of
Choreography description


david.burdett@commerceone.com wrote:

> Ugo
>  
> Another way of thinking about this is from the perspecitive of where 
> you start when developing a choreography.
>  
> The first alternative is "top down" where you define the high level 
> flow of messages first and then work out what the web services you 
> need since they don't exist yet and finally create the WSDL 
> definitions for those services so that you can start building them.
>  
> The second alternative is "bottom up" where you have some Web services 
> that already exist and you need to define the flow of messages so that 
> you can "connect" them.
>  
> I think that both approaches are valid, however following the "top 
> down" approach is hard to do if you can't define the high level flow 
> until you've defined the web services to use as WSDL definitions.
>  
> Bottom line, if you have to have the WSDL definitions before you can 
> develop a choreography, a.k.a. message flow, you are prohibiting the 
> "top down" method of development that I think many, but not 
> all, business analysts would often want to follow.

mm1: David, you continue to discuss abstract choreography definition, 
references to business analysts, and an approach that infers your 
audience are business people. I would purport that WS-CDL is focused on 
the developer, pi-calculus and WSDL, and placement as a mechanism to 
help manage BPEL processes.  Regardless of tooling, I do not see the 
business analysts having any notion about what this language means. I 
will be providing a response to your comments to my concerns about the 
abstract definition. My concern is primarily based on two important 
premises:

    * The fully abstract definition is not well-suited for a WSDL-based
      paradigm.
    * Trying to further abstract WS-CDL may dilute the language [1].

Martin and Steve, I would suggest we resolve this issue quickly. I had 
thought we reached agreement and clarity in Cannes. We are limited in 
moving forward unless we can bound this issue and the continuous swirl 
could impact our capability to deliver. Thanks.

[1] I have seen several specifications that try to do too many things, 
rendering it either unusable or diluted.

>  
> Hope this helps.
>  
> David
>
>     -----Original Message-----
>     *From:* public-ws-chor-request@w3.org
>     [mailto:public-ws-chor-request@w3.org]*On Behalf Of *Ugo Corda
>     *Sent:* Tuesday, April 27, 2004 8:56 AM
>     *To:* Tony Fletcher; public-ws-chor@w3.org
>     *Subject:* RE: Issue 662: specify at least two levels or types of
>     Choreography description
>
>     Tony,
>      
>     The answer to your Why question in the case of BPEL might be as
>     simple as the one you find in "Goals of the BPEL4WS Specification"
>     (see [1]), where it says:
>      
>     "BPEL4WS is firmly set in the Web services world as the name
>     implies.  In particular, all external interactions occur through
>     Web service interfaces described using WSDL.  This has two
>     aspects: (1) the process interacts with Web services through
>     interfaces described using WSDL and (2) the process manifests
>     itself as Web services described using WSDL".
>     In any case, I am curious to see what types of answers your
>     question on whether BPEL abstract processes need WSDL is going to
>     get on the BPEL list.
>      
>     Regards,
>     Ugo
>      
>      
>     [1] 
>     http://www.oasis-open.org/committees/download.php/3249/Original%20Design%20Goals%20for%20the%20BPEL4WS%20Specification.doc
>
>      
>

Received on Monday, 3 May 2004 00:07:16 UTC