Bruce_Kahn@notesdev.ibm.com wrote:
In any case CAP 1.0 09-submitted is unclear as to what it means when no latency-param is specified and that needs to be rectified before it can pass WG last call.
I'll add: If LATENCY is not specified, then the initiator is indicating that any timeouts are up to the responder.
> It
> does not stop the responder from responding earlier
> and it does not stop the responder from sending some kind
> of error before the timeout, both of which terminate
> the command.
This case is not described in 09-submitted and I think it needs to be before this confusion can be considered resolved. Section 10.1.1 Bounded Latency covers many permuations but there is no talk of returning partial results. This needs to be covered in text in 10.1.1 where latency is covered.
Sure it is: If a CS can both start sending the reply to a command and guarantee that all of the results can be sent from a command (short of something like network or power falure), prior to the "LATENCY" timeout value, then the "LETENCY" time has not expired. (the above text being re-writted see separate e-mail on this list) Simply send a reply. -- Doug Royer | http://INET-Consulting.com -------------------------------|----------------------------- Doug@Royer.com | Office: (208)612-INET http://Royer.com/People/Doug | Fax: (866)594-8574 | Cell: (208)520-4044 We Do Standards - You Need Standards
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature