RE: [ws-rx-implement] Phase1 Interop Endpoint for SAP (updated)

[Doug Davis <dug@us.ibm.com>]

AcksTo in general can be anon but there
may be certain cases where it doesn't make sense - like in 3.1 for the
Offered sequence -  where we know the Acks will never flow over an
http response flow because the RMD is non-addressable.  Another example
would be cases where the transport doesn't have a back-channel at all.
 The spec gives a warning that people need to be careful - it seems
good enough.  I doubt we could mention all possible cases.
thanks,
-Doug

"Videlov, Vladimir" <vladimir.videlov@sap.com>

10/25/2006 04:51 AM

To	Doug Davis/Raleigh/IBM@IBMUS
cc	<ws-rx-implement@lists.oasis-open.org>
Subject	RE: [ws-rx-implement] Phase1 Interop Endpoint for SAP (updated)

Hi Doug,
 
You are right and I have fixed
this. Do you think that its worth to be mentioned in the spec that in this
case AcksTo URI can't be anon. Its mentioned that it can't be none, but
in the current scenario (3.1) it may lead to misunderstandings? At least
my understanding is that flagging the AcksTo to anon may be considered
as sending the Acks within the HTTP requests (the one and only available
channel). What do you think ?
 
Vladimir

---

From: Doug Davis [mailto:dug@us.ibm.com]
Sent: 25 ïËÔÏÍ×ÒÉ 2006 Ç. 00:52
To: Videlov, Vladimir
Subject: RE: [ws-rx-implement] Phase1 Interop Endpoint for SAP (updated)


ok - your AcksTo on the CreateSequenceResponse can not be anonymous - it
has to be a real URI since your messages are flowing over HTTP response
flows. 
-Doug 


 <vv:snip />