W3C home > Mailing lists > Public > public-ws-chor@w3.org > June 2006

Re: Participants without roles, roles without behaviour

From: Gary Brown <gary@pi4tech.com>
Date: Mon, 12 Jun 2006 11:14:43 +0100
Message-ID: <448D3E93.8090605@pi4tech.com>
To: matthew@stickledown.com
CC: public-ws-chor@w3.org

Hi Matthew

 From my perspective, the list of participants you have provided sound 
more like roles than participants.

In WS-CDL a participant is really a logical grouping of roles, where 
each of those roles can then share some state information associated 
with that participant - i.e. so if a variable 'foo' is defined at two or 
more roles associated with a particular participant, then if that 
variable is set at one of the roles, its value will be available to all 
of the other roles within that participant.

All of the activities in WS-CDL are associated with roles - i.e. an 
interaction is performed between two roles, an assignment is performed 
at a role, etc. There are no direct associations between any activity 
and a participant, this is only indirectly inferred through the role.

One way to model what you require would be to create a different 
participant for each role (i.e. have a one to one relationship between 
them) - unless you require state information to be shared. So, 
participants are defined in terms of their roles, as to have a 
participant with no associated roles would be serve no purpose in WS-CDL.

Could you provide additional clarification regarding your last question.

Regards
Gary


Matthew Rawlings wrote:
>
> I have a few more questions around my implementations of WSCDL within 
> the financial services industry. Please help me with my understanding.
>
>  
>
> In WSCDL Participants are required to have at least one role, and 
> roles are required to have at least one behaviour. However, I wanted 
> to identify a Participant without any Roles, but find I cannot do this 
> in WSCDL.
>
>  
>
> My example comes from OTC derivatives trading. FpML 
> (http://www.fpml.org <http://www.fpml.org/>), defines 9 standard 
> participants in a OTC contract: orderer, introducer, executor, 
> confirmer, creditor, calculater, settler, beneficiary, accountant. I 
> wish to explicitly state that the beneficiary is a Participant that 
> has no Roles in the trade execution choreography. In WSCDL it seems 
> that to define a participant you are also required to define at least 
> one role, which is impossible if it has no roles.
>
>  
>
> Have I understood this correctly? How can I achieve my goal with WS-CDL?
>
>  
>
> Is a Participant in CDL defined solely in terms of its roles, or can 
> it be defined independently of its roles? I.e. if the roles change for 
> a participant is it a new participant?
>
>  
>
> Does WS-CDL have a closed or open world assumption about the base types?
>
>  
>
> *Matthew Rawlings*
>
> +44 791 539 7824
>
>  
>
Received on Monday, 12 June 2006 10:15:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 18 December 2010 01:01:45 GMT