OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

ws-rx message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: RE: [ws-rx] State tables rev of July 25



From: Marc Goodner [mailto:mgoodner@microsoft.com]
Sent: Thursday, July 27, 2006 2:55 PM
To: Bob Freund-Hitachi; [WS-RX]
Subject: RE: [ws-rx] State tables rev of July 25


Ah, OK solves all on the table titles. I find that use case weird and only think of the client initiating this. You are right that it is either. I agree with your proposed change there.


Regarding SequenceTerminated, I still think the current table and explanatory text is sufficient. Would you be satisfied if the text that explains the blank cells included a note that a SequenceTerminated fault may result from entering into one of those states?


From: Bob Freund-Hitachi [mailto:bob.freund@hitachisoftware.com]
Sent: Thursday, July 27, 2006 11:49 AM
To: Marc Goodner; Bob Freund-Hitachi; [WS-RX]
Subject: RE: [ws-rx] State tables rev of July 25



Referencing section 3.7 therein it is stated:

A common usage will be a client RM Destination sending MakeConnection to a server RM Source for the purpose of receiving asynchronous response messages”

That seems more illustrative than prescriptive.

Looking at it without assumptions, it now looks to me that it is not RMD and RMS, but rather Anon endpoint and non-Anon endpoint which seems could be either rms or rmd roles.

Yes, the titles of the tables are wrong.

The first new table should be table 3 anon endpoint
The second new table should be table 4 non-anon endpoint.

As for Sequence Terminated:

This fault seems to be the mechanism to declare the occurrence of a protocol violation.

Are protocol violations just informative without consequence?

What is a protocol violation?





From: Marc Goodner [mailto:mgoodner@microsoft.com]
Sent: Thursday, July 27, 2006 1:44 PM
To: Bob Freund-Hitachi; [WS-RX]
Subject: RE: [ws-rx] State tables rev of July 25




On the MakeConnection tables I think the table labeled RMS is really RMD, and likewise the RMS table is the RMD. I don’t think there is anything to say about the non-response situation here, maybe [none].


The other changes look good to me.


Regarding your new question, no. I think there are a number of blanks where that would not make sense. I think leaving it as is with the explanation that it is not normal protocol behavior is sufficient.



From: Bob Freund-Hitachi [mailto:bob.freund@hitachisoftware.com]
Sent: Wednesday, July 26, 2006 12:54 PM
To: [WS-RX]
Subject: [ws-rx] State tables rev of July 25


I have incorporated the results of the resolution last meeting to the message number rollover discussion.

We will need to update these tables when there is a resolution to the expiry issue which hopefully will occur this week.



There has also been an attempt made to incorporate the results of make connection to fulfill my action item on that subject.

Since Make Connection is not related solely to a specific sequence, it did not fit into the fabric of the other state tables previously submitted.

I chose to add a couple more tables to demonstrate a simplistic underlying message transfer engine.  Section 3.7 is not clear (to me) as to what conditions cause the initiation of polling, what might trigger a specific poll (interval or event), or how long such polling might continue.

I also do not see what the negative response to a poll might be and what the RMD might do upon receipt of the negative (or non) response.

So please take it for now for discussion and hopefully out of the fire there might be found an answer or two.


Other changes:

There has been a new event added to the RMD table indicating the receipt of an invalid acknowledgement fault.  In Section 4 the fault is described but I cannot locate what its implications might be.


New Question.

All of the blank cells in the table are unspecified behavior in the specification.  Should they be filled with “generate Sequence Terminated Fault”?




[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]