W3C home > Mailing lists > Public > www-voice@w3.org > January to March 2008

Re: CCXML: Dialog and Connection objects ISSUE-236 / ISSUE-237 [cc]

From: RJ Auburn <rj@voxeo.com>
Date: Thu, 27 Mar 2008 09:37:21 -0400
To: Petr Kuba <kuba@optimsys.cz>, Hrvoje Nezic <hrvoje.nezic@envox-lab.hr>
Message-Id: <3E7D8841-414B-4F05-A058-A7E9AD6560DD@voxeo.com>
Cc: www-voice@w3.org, W3C Voice Browser Working Group <w3c-voice-wg@w3.org>

Petr/Hrvoje,

We actually did a pretty major cleanup based on this thread and we  
feel that the current spec addresses a lot of the little  
inconsistencies that were in the spec prior to this.

A quick summary of the current state of affairs is below:

------------- START --------------
Dialog Object Properties:
	dialogid	This property is the ECMAScript string value of the
			Dialog Identifier, which uniquely identifies each
			instance of the Dialog class.
	input		As defined by the Media Endpoint Object in 10.4.3.
	outputs		As defined by the Media Endpoint Object in 10.4.3.
	objecttype	This property states the type of this object which must be  
'dialog'

Connection Object Properties (only properties of interest, not a  
complete list):
	connectionid	This property is the ECMAScript string value of the
			Connection Identifier, which uniquely identifies each
			instance of the Connection object
	input		As defined by the Media Endpoint Object in 10.4.3.
	outputs		As defined by the Media Endpoint Object in 10.4.3.
	objecttype	This property states the type of this object which must be  
'connection'

Media Endpoint Object (10.4.3)

	input		The identifier (connectiondid/conferenceid/dialogid) of the
			single media endpoint (connection, dialog or conference) providing
			the input stream to this media endpoint or undefined if
			there is no input stream. This property MUST be updated each
			time a <join>/<unjoin> or any other media operation changes
			the media source to this media endpoint. If the media endpoint
			does not have a media source the value MUST be ECMAScript undefined.
			For example the creation of a half duplex bridge
				<join id1="con1" id2="con2" duplex="half" />
			Result
				session.connections[con1].input = "con2"
				session.connections[con2].input = undefined

	outputs		An array containing the identifiers of all media endpoints  
to which
			the output stream of this media endpoint is sent. If the media  
endpoint
			does not have any media destinations the length of the array MUST  
be 0.
			For example the creation of a half duplex bridge
				<join id1="con1" id2="con2" duplex="half" />
			Result
				session.connections[con1].outputs.length = 0
				session.connections[con2].outputs.length = 1

------------- END --------------

Additionally we added references to the actual dialog/conference/ 
connection objects on the conference.join/unjoined events.

The end result is we feel that the current spec now presents a  
consistent model for inputs and outputs across all the different media  
objects.

Please let us know if these changes address your concerns. If there  
are still outstanding issues we would be happy to work with you future  
to try and address what needed to be done. If we do not hear back from  
you by April 17th 2008 we will consider this issue closed in our issue  
tracker system.

Thank you very much for your valuable comments on the CCXML  
specification.

Best regards,

	RJ


---
RJ Auburn
CTO, Voxeo Corporation
tel:+1-407-418-1800



On Mar 15, 2007, at 09:30:05, Hrvoje Nezic wrote:
>
> ----- Original Message ----- From: "Petr Kuba" <kuba@optimsys.cz>
> To: <www-voice@w3.org>
> Cc: "Hrvoje Nezic" <hrvoje.nezic@envox-lab.hr>
> Sent: Thursday, March 15, 2007 1:04 PM
> Subject: Re: CCXML: Dialog and Connection objects
>
>
>> Dear Hrvoje,
>> thanks for your comments and let me add some notes in-line.
>>
>> However I would still appreciate further clarification on my  
>> original comments.
>>
>> Hrvoje Nezic wrote:
>>>> Dialog object contains connectionid and/or conferenceid  
>>>> properties which
>>>> identify the connection/conference that is driving a media stream  
>>>> to the
>>>> dialog.
>>>
>>> The specification doesn't explicitly state that the connection/ 
>>> conference
>>> is driving a media stream to the dialog, but this is a reasonable  
>>> interpretation.
>>
>> See CCXML specification (version 20070119), section 7.4 Dialog  
>> object properties, description of the connectionid property which  
>> explicitly states:
>> "Identifies the connection that is driving a media stream to the  
>> dialog. ..."
>> The same holds for the conferenceid property.
>>
>>> However, other interpretations are also possible, because the  
>>> specification
>>> states that the connection/conference is "associated" with the  
>>> dialog.
>>
>> Statement like this is in the Connection object description, not  
>> Dialog object. The specification is very clear for the Dialog object.
>>
>>> For example, connection/conference could be the one specified in  
>>> <dialogprepare>
>>> or <dialogstart> although it doesn't drive a media stream to the  
>>> dialog.
>>
>> This is wrong.
>>
>>>> As we understand, after introducing input and outputs properties  
>>>> in the
>>>> last draft connectionid/conferenceid will always have the same  
>>>> value as
>>>> the input property.
>>>
>>> There are still no input and output properties in <dialogprepare>  
>>> and <dialogstart>.
>>> I think that the specification should introduce these properties  
>>> instead of
>>> connectionid / conferenceid.
>>
>> Input and outputs properties in the Connection and Dialog objects  
>> serve for storing information about all the endpoins bridged to the  
>> particular connection / dialog. Connectionid / conferenceid in  
>> <dialogprepare> /
>> <dialogstart> and joinid in <createcall> serve for establishing one  
>> new bridge. You do not need input / outputs properties in  
>> <dialogprepare> /
>> <dialogstart> / <createcall> tags.
>
> Petr,
>
> Of course, you are right. Please read my previous message: I was
> "too quick on trigger" and because these days I am not occupied with  
> CCXML
> at all, I confused some things. I apologize again.
>
> Hrvoje
>
>
Received on Thursday, 27 March 2008 13:39:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 27 March 2008 13:39:05 GMT