W3C home > Mailing lists > Public > www-voice@w3.org > October to December 2008

Re: Notification of implicit bridge teardowns - ISSUE-525

From: RJ Auburn <rj@voxeo.com>
Date: Mon, 24 Nov 2008 22:47:57 -0500
Cc: www-voice@w3.org, W3C Voice Browser Working Group <w3c-voice-wg@w3.org>
Message-Id: <5D15220F-0DCB-4D35-A3D2-B8CEB551DFF2@voxeo.com>
To: "Sanders, Derek (Derek)" <dsanders@avaya.com>


The working group reviewed your question and has the following response.

> CCXML platforms do not generate ‘conference.unjoined’ events as a  
> result of implicit teardowns when applications perform <join>/ 
> <unjoin> or equivalent actions (such as <createcall> with a joined),  
> regardless of whether these teardowns are partial or complete.  The  
> reason for this is that such teardowns are a direct consequence of  
> actions taken by the application, for which outcome events already  
> exist (‘conference.joined’/’conference.unjoined’ against the media  
> endpoints directly affected).  State variables for both directly and  
> implicitly affected media endpoints are updated when this primary  
> events fires; failing to do so would result in inconsistent session  
> state between the two events when bridges appeared to exist that in  
> actuality do not.  By contrast, the ‘conference.unjoined’ events  
> specified to 10.6.14 exist to ensure that media bridges are  
> determined entirely using ‘connection.joined’ and  
> ‘connection.unjoined’ events, rather than being derived from call  
> control events such as ‘connection.disconnected’.

Hopefully this clarification addresses your concern. If you have any  
follow up questions please let us know.

Best regards,


RJ Auburn
CTO, Voxeo Corporation
Chair, Editor and Chair, CCXML, VBWG, W3C

On Jul 17, 2008, at 10:33 AM, RJ Auburn wrote:

> Derek:
> This is being tracked as ISSUE-525. Thanks for the feedback and we  
> will have an answer for you shortly.
> Best regards,
> 	RJ
> On May 30, 2008, at 1:23 PM, Sanders, Derek (Derek) wrote:
>> The January 19th, 2007 CCXML Working Draft is not very clear on how  
>> implicit bridge teardowns resulting from a <join> should be  
>> handled.  Section 10.4.1 shows all of the possible outcomes of a  
>> <join> tag.  Some of these examples require a full or partial  
>> teardown of an existing bridge.  The spec does not state if a  
>> ‘conference.unjoined’ event should be generated when this occurs.   
>> It does state in section 10.6.14 that if a connection is dropped  
>> (as in a merge, disconnect, etc.), then the appropriate  
>> ‘conference.unjoined’ event(s) should be sent.  It may be an easy  
>> assumption that ANY implicit bridge teardowns should result in a  
>> ‘conference.unjoined’ event, but what about partial teardowns?  It  
>> starts to get a little more complicated there.  Is it enough to  
>> just update the connection state variables when bridges change as a  
>> result of a <join>?
>> Thanks,
>> -Derek Sanders
Received on Tuesday, 25 November 2008 03:48:18 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:07:40 UTC