Re: Comments on primer please

Section 3 should be the example, and intro into "interaction-oriented" 
distributed system design and the components involved (e.g. participants, 
roles, etc).

Section 4 can then start to present more technical detail about the type 
definition constructs, followed by section 5 on interactions, etc (so the 
sections are pushed back by one).


----- Original Message ----- 
From: "Gary Brown" <gary@pi4tech.com>
To: "'WS-Choreography List'" <public-ws-chor@w3.org>
Sent: Tuesday, July 05, 2005 3:01 PM
Subject: Re: Comments on primer please


>
> Comments:
>
> I think the same document should be suitable for business analysts 
> (probably jusst the first few sections), choreography designers with the 
> various advanced situations that should be covered, but should also be 
> suitable for CDL implementors - i.e. providing clarifications on how 
> advanced situations should be handled, including how to potentially bind 
> with other standards (e.g. security, reliable messaging, coordination, 
> etc.).
>
> Section 3: Although I think the initial focus on interactions is a good 
> idea, as essentially CDL is promoting an 'interaction oriented' approach 
> to designing distributed systems, I am not sure that the interaction 
> syntax details (and code fragments) are appropriate here. I think section 
> 3 should be focusing on describing the example, highlighting the 
> interaction focus, but then establish the static type information required 
> to support the interaction-oriented description - including participants 
> which is currently missing. This then sets the scene for the flow 
> description of the activities.
>
> Possibly section 4 can be describing the interaction (with the details 
> about exchanges and record details).
>
> Then maybe section 5 on structural activities.
>
> Section 6 on more advanced choreo patterns
> - modularization (performing sub-choreos)
> - including async concurrent performs
> - dependent workunits
>
> Section 7 could then focus on Faults, Exception Handling and Finalization 
> (talk about finalization as opposed to compensation, as compensation is 
> only one possible use of the finalizer mechanism). Also could deal with 
> timeouts?
>
> Section 8 on "gotchas" - such as race conditions and the consequences of 
> global conditions that do not have the appropriate state management to 
> ensure that each participant have the appropriate state to support the 
> conditions. Also possibly discuss interference scenarios in terms of 
> linearity?
>
> Section 9 - WS-CDL with other standards - show examples of how CDL can be 
> bound to security, reliable message and coordination standards
>
> Regards
> Gary
>
>
>
>
>
>
> ----- Original Message ----- 
> From: "Steve Ross-Talbot" <steve@pi4tech.com>
> To: "'WS-Choreography List'" <public-ws-chor@w3.org>
> Sent: Monday, July 04, 2005 11:52 AM
> Subject: Comments on primer please
>
>
>>
>> All:
>>
>> I'd like to get some direction comments and even contributions to the 
>> primer.
>>
>> Can you take a look between now and tomorrow's meeting so that we can be 
>> focussed.
>>
>> One thing to think about is who are the readers (the roles they play) and 
>> what we want them to get out of the primer.
>>
>> Cheers
>>
>> Steve T
>>
>>
>>
>>
>>
>
>
>
>
>
>

Received on Tuesday, 5 July 2005 19:53:21 UTC