Re: Need volunteer to draft definition ofreliablemessaging,wasRE:reliable messaging - hop by hop

David,  This has been my point all along. Although I disagree as to  the only .  Fairlure to receive a DFN is a fairly good assurance, not proof but possibly good enough for some. Receipt of the expected response message is another form of proof. If the response is typically generated with little or no delay, and the business request has a finite ttl, then maybe that's good enough.  Also, even if the message is delivered with BestEffort, a receipt (NRR) may be required to  prove  that the message was received.  I have not been arguing against a delivery receipt, just its role (or in my case its non-role) in RM protocol.  Cheers,  Chris David Fischer wrote: >
>
End-to-end acks don't ensure success, they can only report it. >
> The lack of an end-to-end ack does not necessarily mean failure but the arrival
> of an end-to-end ack is an assurance/report that there was success -- it is the
> only assurance. >
> David Fischer
> Drummond Group. >  >