Re: Refinements of comments 7 and 9 [Re: [wsrm] Proposed clarificaitonresolutions to Ferris comments 1,2,24,5,7,9,16,19]

[Tom Rutt <tom@coastin.com>]

Doug:

I basically agree with the direction of your change,

however in some cases, the api may have the consumer invoke the deliver 
operation, in a polling
kind of api.

Thus, support is ambiguous enough to allow such an implementation.

Do you think the RMP is always the invoker of deliver?

If so , we should clarify that way,but does this cause any 
implementation problems?

I am not sure, that is why I am asking\

Tom Rutt

Doug Bunting wrote:

> Tom,
>
> For both of these comments, your suggestions need a bit more to solve 
> the issues Chris was kind enough to identify.
>
> On 14-Jul-04 13:58, Tom Rutt wrote:
>
> ...
>
>> Comment 7: Message Delivery is not the end of Receiving RMP processing
>>
>> Proposed clarification:
>>
>> Lines 204-206: change
>>
>> “
>>
>> Message Delivery:
>>
>> The action of invoking the “deliver” operation for a Reliable 
>> Message. This action marks the end of the RMP processing for this 
>> message.
>>
>> “
>>
>> to
>>
>> “
>>
>> Message Delivery:
>>
>> The action of invoking the “deliver” operation for a Reliable Message.
>>
>> “
>
>
> Actually, this does not cover the problems Chris identified with 
> failed Deliver operations nor the explicit connection we have made 
> between message delivery and sending an acknowledgement (when 
> required).  I suggest:
>
> "Message Delivery:
>
> Completion of the Deliver operation for a Reliable Message."
>
>> Comment 9: Use of the word implement with Abstract operation
>>
>> Proposed clarification:
>>
>> Lines 256-257: change
>>
>> “
>>
>> An RMP acting in the role of a Sending RMP MUST implement Submit, and 
>> notification of failure (Notify). An RMP acting in the role of a 
>> Receiving RMP MUST implement Deliver.
>>
>> “
>>
>> with
>>
>> “
>>
>> An RMP acting in the role of a Sending RMP MUST support use of Submit 
>> and Notify. An RMP acting in the role of a Receiving RMP MUST support 
>> use of Deliver.
>>
>> “
>
>
> An important part of the issue here was implications around who 
> implements what.  The phrase "support use" is not dissimilar enough to 
> avoid these implications.  I am also unsure why we are bothering with 
> the "acting in the role of" wordiness.  Finally, I see nothing in the 
> specification requiring a sending RMP to use AckRequested, meaning 
> that Notify should also be optional (for a more extreme reason, 
> admittedly, than Respond is optional).  I suggest:
>
> "A Sending RMP MUST support the Submit operation.  A Receiving RMP 
> MUST invoke the Deliver operation for every valid, in-order and 
> non-expired message it receives."
>
> ...
>
> thanx,
>     doug
>

-- 
----------------------------------------------------
Tom Rutt	email: tom@coastin.com; trutt@us.fujitsu.com
Tel: +1 732 801 5744          Fax: +1 732 774 5133