Voice Browser Call Control: CCXML Version 1.0

Abstract

This document describes CCXML, or the Call Control eXtensible Markup Language. CCXML is designed to provide telephony call control support for dialog systems, such as VoiceXML [ VOICEXML ]. While CCXML can be used with any dialog systems capable of handling media, CCXML has been designed to complement and integrate with a VoiceXML interpreter. Because of this there are many references to VoiceXML's capabilities and limitations. There are also details on how VoiceXML and CCXML can be integrated. However, it should be noted that the two languages are separate and are not required in an implementation of either language. For example, CCXML could be integrated with a more traditional Interactive Voice Response (IVR) system or a 3GPP Media Resource Function (MRF), and VoiceXML or other dialog systems could be integrated with other call control systems.

Status of this Document

This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.

This is the 1 April 2010 Candidate Recommendation of "Call Control eXtensible Markup Language (CCXML) Version 1.0". Changes from the Last Call Working Draft can be found in Appendix F - Changes . A diff-marked version of this document is also available for comparison purposes.

This specification describes the Call Control XML (CCXML) markup language that is designed to provide telephony call control support for VoiceXML or other dialog systems.

This document has been produced as part of the W3C Voice Browser Activity , following the procedures set out for the W3C Process . The authors of this document are members of participants in the Voice Browser Working Group ( W3C Members only ). This is . For more information see the Last Call Working Draft of CCXML 1.0. Voice Browser FAQ . The Last Call period is set to last 3 weeks. Therefore comments on the documents are welcome until 7 February 2007. There are no changes since the previous Working Draft , except that Group expects to advance this is a Last Call WD. document to Recommendation status.

This document is also available a W3C Candidate Recommendation for review by W3C Members and other interested parties. W3C publishes a technical report as a non-normative HTML page showing Candidate Recommendation to indicate that the changes since previous revisions . For a detailed list please see Appendix F - Changes . document is believed to be stable, and to encourage implementation by the developer community.

An Implementation Report Plan is currently being developed for this specification. The Working Group currently expects entrance criteria to the Proposed Recommendation phase require at least two independently developed interoperable implementations of each required feature, and at least one implementation or two implementations of each feature, optional feature depending on whether the feature's conformance requirements have an impact on interoperability. Detailed implementation requirements and the invitation for participation in order to exit the next phase Implementation Report are provided in the Implementation Report Plan . We expect to meet all requirements of this document, that report within the Candidate Recommendation phase. To help the period closing 28 May 2010 . The Voice Browser Working Group build such a report, reviewers are encouraged will advance CCXML 1.0 to implement this Proposed Recommendation no sooner than 28 May 2010 .

Several of the features in the current draft specification and are considered to indicate be at risk of removal due to W3C which features have been implemented, and any problems that arose. potential lack of implementations.

• <move> - Section 9.2.4
• Advanced <join> use cases (dialog to dialog, conf to conf) - Section 7 , combined with Section 10.4
• Advanced entertone / exittone (URI) - Section 10.5.7
• dialog.transfer eventing and use case - Section 7.3.5 and Appendix D.9

This document is for public review. Comments and discussion are welcomed welcome on the public mailing list < www-voice@w3.org >. To subscribe, send an email to < www-voice-request@w3. org ( archive > with the word subscribe in the subject line (include the word unsubscribe if you want to unsubscribe). The ). See W3C mailing list and archive usage guidelines . Please check the disposition of comments for received during the list is accessible on-line. Last Call period.

Publication as a Working Draft Candidate Recommendation does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress.

This document was produced by a group operating under the 5 February 2004 W3C Patent Policy . W3C maintains a public list of any patent disclosures made in connection with the deliverables of the group; that page also includes instructions for disclosing a patent. An individual who has actual knowledge of a patent which the individual believes contains Essential Claim(s) must disclose the information in accordance with section 6 of the W3C Patent Policy .

Conventions of this Document

In this document, the key words "must", "must not", "required", "shall", "shall not", "should", "should not", "recommended", "may", and "optional" are to be interpreted as described in [RFC2119] and indicate requirement levels for compliant CCXML implementations.

Table of Contents

• 1: Introduction
• 2: Motivation (Informative)
• 3: Concepts and Architecture
• • 3.1: Event Processing
  • 3.2: Conferencing
  • 3.3: Scripting
  • 3.4: Definitions
  • 3.5: Session Life-Cycle
  • • 3.5.1: Startup
    • 3.5.2: Shutdown
    • 3.5.3: Session Life-Cycle Diagrams
• 4: Simple Examples
  • 4.1: Hello World
  • 4.2: Accept or Reject a Call
  • 4.3: Simple Dialog
• 5: CCXML Elements Listing
• 6: Document Control Flow and Execution
• • 6.1: Overview
  • 6.2: Elements
  • • 6.2.1: <ccxml>
    • 6.2.2: <meta>
    • 6.2.3: <metadata>
    • 6.2.4: <if>
    • 6.2.5: <elseif>
    • 6.2.6: <else>
    • 6.2.7: <fetch>
    • 6.2.8: <goto>
    • 6.2.9: <createccxml>
    • 6.2.10: <exit>
    • 6.2.11: <log>
  • 6.3: Events
  • • 6.3.1: Overview
    • 6.3.2: fetch.done
    • 6.3.3: error.fetch
    • 6.3.4: ccxml.exit
    • 6.3.5: ccxml.loaded
    • 6.3.6: ccxml.kill
    • 6.3.7: ccxml.created
    • 6.3.8: error.createccxml
    • 6.3.9: error.unsupported
• 7: Dialogs
• • 7.1: Overview
  • 7.2: Elements
  • • 7.2.1: <dialogprepare>
    • 7.2.2: <dialogstart>
    • 7.2.3: <dialogterminate>
  • 7.3: Events
  • • 7.3.1: Overview
    • 7.3.2: dialog.started
    • 7.3.3: dialog.exit
    • 7.3.4: dialog.disconnect
    • 7.3.5: dialog.transfer
    • 7.3.6: dialog.terminatetransfer
    • 7.3.7: error.dialog
    • 7.3.8: error.dialog.notstarted
    • 7.3.9: blank
    • 7.3.10: dialog.user.*
    • 7.3.11: dialog.prepared
    • 7.3.12: error.dialog.notprepared
  • 7.4: Dialog Object Properties
  • 7.5: Dialog Class
• 8: Variables and Expressions
• • 8.1: Overview
  • 8.2: Elements
  • • 8.2.1: <assign> and <var>
    • 8.2.2: <script>
  • 8.3: Session Variables
  • 8.4: Application Variables
• 9: Event Handling
• • 9.1: Overview
  • 9.2: Elements
  • • 9.2.1: <eventprocessor>
    • 9.2.2: <transition>
    • 9.2.3: <send>
    • 9.2.4: <move>
    • 9.2.5: <cancel>
  • 9.3: Events
  • • 9.3.1: error.notallowed
    • 9.3.2: error.semantic
    • 9.3.3: error.send.targetunavailable
    • 9.3.4: error.send.targettypeinvalid
    • 9.3.5: error.send.failed
    • 9.3.6: send.successful
    • 9.3.7: move.successful
    • 9.3.8: cancel.successful
    • 9.3.9: error.move
  • 9.4: Standard Events
  • • 9.4.1: Overview
    • 9.4.2: Standard Event Attributes
    • • 9.4.2.1: Overview
      • 9.4.2.2: Standard Event Attribute Table
    • 9.4.3: Connection Events
    • 9.4.4: Language Events
    • 9.4.5: Error Events
  • 9.5: Error Handling
  • • 9.5.1: Fetching & compilation errors
    • 9.5.2: Document Initialization Errors
    • 9.5.3: Error In <eventprocessor> attributes
    • 9.5.4: Error in <transition> attributes
    • 9.5.5: Errors While Handling Error Events
    • 9.5.6: Errors in hints attributes
• 10: Telephony Operations and Resources
• • 10.1: Overview
  • • 10.1.1: Concepts Background (INFORMATIVE)
  • 10.2: Connections
  • • 10.2.1: Connection State
    • 10.2.2: Connection Object
    • 10.2.3: Connection Events
    • 10.2.4: Connection Operations
    • 10.2.5: Connection Class
  • 10.3: Conferences
  • • 10.3.1: Conference Object Properties
    • 10.3.2: Conference Class
  • 10.4: Bridges
  • • 10.4.1: <join> Outcomes
    • 10.4.2: Invariant Examples
    • 10.4.3: Media Endpoint Properties
  • 10.5: Elements
  • • 10.5.1: <accept>
    • 10.5.2: <redirect>
    • 10.5.3: <reject>
    • 10.5.4: <createcall>
    • 10.5.5: <createconference>
    • 10.5.6: <destroyconference>
    • 10.5.7: <join>
    • 10.5.8: <unjoin>
    • 10.5.9: <disconnect>
    • 10.5.10: <merge>
  • 10.6: Events
  • • 10.6.1: Overview
    • 10.6.2: connection.alerting
    • 10.6.3: connection.progressing
    • 10.6.4: connection.connected
    • 10.6.5: connection.disconnected
    • 10.6.6: connection.redirected
    • 10.6.7: connection.merged
    • 10.6.8: connection.failed
    • 10.6.9: error.connection
    • 10.6.10: connection.signal
    • 10.6.11: conference.created
    • 10.6.12: conference.destroyed
    • 10.6.13: conference.joined
    • 10.6.14: conference.unjoined
    • 10.6.15: error.conference.create
    • 10.6.16: error.conference.destroy
    • 10.6.17: error.conference.join
    • 10.6.18: error.conference.unjoin
    • 10.6.19: connection.merge.failed
    • 10.6.20: error.connection.wrongstate
    • 10.6.21: error.conference
    • 10.6.22: connection.redirect.failed
    • 10.6.23: connection.accept.failed
    • 10.6.24: connection.reject.failed
• 11: Complex Examples
• • 11.1: Calling Card Application
  • 11.2: Conferencing application
  • 11.3: Personal Assistant
• Appendix A - Related Work
• Appendix B - CCXML DTD
• Appendix C - CCXML Schema
• Appendix D - VoiceXML 2.0 Integration Details
• Appendix E - Recommended Telephony Protocol Names
• Appendix F - Changes
• Appendix G - Acknowledgments
• Appendix H - References
• Appendix J - Conformance
• Appendix K - Basic HTTP Event I/O Processor
• Appendix L - Session Creation Event I/O Processor
• Appendix M - MIME Type and File Suffix

1: Introduction

This document describes CCXML, the Call Control eXtensible Markup Language. CCXML provides declarative markup to describe telephony call control. CCXML is a language that can be used with a dialog system such as VoiceXML [ VOICEXML ].

CCXML can provide a complete telephony service application, comprised of Web server CGI compliant application logic, one or more CCXML documents to declare and perform call control actions, and to control one or more dialog applications that perform user media interactions

Since platforms implementing CCXML may choose to use one of many telephony call control definitions (JAIN Call Control [ JSR021 ], ECMA CSTA [ CSTA ], S.100 [ S.100 ], etc.), the call control model in CCXML has been designed to be sufficiently abstract so that it can accommodate all major definitions. For relatively simple types of call control, this abstraction is straightforward. The philosophy in this regard has been to "make simple things simple to do." Outdial, transfer (redirect), two-party bridging, and many forms of multi-party conferences fall within this classification.

Figure 1 shows the architecture of a telephony implementation consisting of four primary components:

• a caller (along with the telephone network),
• a dialog system (e.g. a VoiceXML implementation),
• a conference server used to mix media streams,
• and the CCXML implementation which manages the Connections between the first two components.

The Telephony Web Application may or may not be integrated with the Voice Web Application.

The Telephony Control and Dialog Control Interfaces may be implemented as an API or protocol.

The components as shown in the figure below represent logical functions, and are not meant to imply any particular architecture.

Figure 1

2: Motivation (Informative)

CCXML is designed to complement dialog systems such as VoiceXML by providing advanced telephony functions. It also can be used as a third-party call control manager in any telephony system. This document contains references to VoiceXML's capabilities and limitations, as well as details on how VoiceXML and CCXML can be integrated.

The CCXML specification originated from the desire to handle call control requirements that were beyond the scope of the VoiceXML specification. The following requirements are addressed by this specification:

• Support for multi-party conferencing, with advanced conference and audio control. A conferencing application involves multiple participants, and is dependent upon call control to establish relationships between those participants.
• The ability to give each active call leg its own dedicated VoiceXML interpreter. For example, in VoiceXML, the second leg of a transferred call lacks a VoiceXML interpreter of its own, limiting the scope of possible applications.
• Sophisticated multiple-call handling and control, including the ability to place outgoing calls.
• Handling for a richer class of asynchronous events. Advanced telephony operations involve substantial amounts of signals, status events, and message-passing. VoiceXML 2.0 does not integrate asynchronous "external" events into its event-processing model.
• VoiceXML lacks the external interfaces required to interact with an outside call queue, or place calls on behalf of an external document server

CCXML and VoiceXML implementations are not mutually dependent. A CCXML implementation may or may not support voice dialogs, or may support dialog languages other than VoiceXML.

3: Concepts and Architecture

A CCXML application consists of a collection of CCXML documents that control and manage the objects listed below:

• CCXML Session: A CCXML session is comprised of an executing CCXML document, or sequence of CCXML documents; each concurrently executing CCXML document is a separate session, and can be uniquely identified and referenced.
• Connection: Connections can be "call legs" (real-world phone connections) or system resources to facilitate interaction with a voice dialog . Media streams between Connections , or between Connections and Conference objects , need to be tracked by the CCXML interpreter and will take real system resources, but do not need a dedicated identifier because they are identified by their endpoints. Ownership of Connections can be moved from one session to another using <move> .
• Conference object: A Conference Object models a resource for mixing media streams. In order to accommodate the widest range of underlying telephony call control definitions, CCXML assumes a separate Conference Object , realizing that this abstraction may not have a direct counterpart in all telephony platforms. See <createconference> and <destroyconference> for further information.
• Dialog: When active, a Dialog may interact with other Connections or Conferences using one-way or two-way media streams, often under the control of dialog environment such as VoiceXML.

CCXML programs manipulate these entities through elements defined in the CCXML language. They can also send and/or receive asynchronous events associated with these entities.

CCXML programs directly manipulate Connection Objects and Conference Objects with various elements in the language, such as <accept> , <createconference> , and <join> . CCXML may also receive events from Connection and Conference Objects , in the case of line signaling, line-status informational messages, or error and failure scenarios.

CCXML programs can start and kill Voice Dialogs using language elements. It can receive events from Voice Dialogs , which may be standardized events such as dialog.exit , or application-specific ones. CCXML can support sending of an event to a Voice Dialog .

CCXML programs can create other CCXML sessions using <createccxml> . This is the only guaranteed control mechanism a CCXML Session ever wields over another. Any other interaction takes place through the event mechanism. CCXML Sessions can both send and receive events between one another.

3.1: Event Processing

Telephone applications need to receive and process large numbers of events in real-time. These events arrive from outside the program itself - either the underlying telephony platform, or from other sources of events.

A CCXML program includes event handlers which are executed when certain events arrive. There are mechanisms for passing information back and forth between Voice Dialogs (such as VoiceXML) and CCXML, but the important points are that CCXML:

• lives on its own thread, and
• carries the burden of rapid asynchronous event handling

Note: References to threads are meant as logical threads and do not imply any specific platform implementation.

3.2: Conferencing

CCXML provides a powerful and flexible method of creating multi-party calls based on the following concepts:

• Call legs are audio sinks and sources which can be combined to form arbitrary networks.
• A conference is an audio stream that mixes together all the speakers' audio outputs. CCXML conference networks allow for conference objects , which mix inputs into a single output channel.
• A conference's architecture may be modified dynamically, thus allowing for moderation, floor control, mute, etc.
• Media splitters, which take a single input channel and emit it over several outputs, may be used by an implementation to realize the CCXML model in a given network environment.
• Conferencing technology is implementation-dependent. CCXML implementations may not support all conferencing features mentioned above.

3.3: Scripting

The computational semantics of CCXML language is based on the ECMAScript Compact Profile (ES-CP, also known as ECMA-327) [ ECMA327 ]. ES-CP is a strict subset of the third edition of ECMA-262 [ ECMASCRIPT ]. Execution efficiency is a primary goal of CCXML implementations, and ES-CP was chosen to ensure that CCXML implementations can operate in a variety of execution environments and without excessive execution overhead.

The ES-CP document specification states:

'ECMAScript Compact Profile is a subset of ECMAScript 3rd Edition tailored to resource-constrained devices such as battery powered embedded devices. Therefore, special attention is paid to constraining ECMAScript features that require proportionately large amounts of system memory (both for storing and executing the ECMAScript language features) and continuous or proportionately large amounts of processing power.'

While CCXML implementations are not necessarily intended for battery powered embedded devices, it is intended to be used in large, real-time telephony platforms managing thousands of lines. The constraints of ES-CP emphasize CCXML's ongoing concern for execution efficiency.

Even though ES-CP tends to be implemented using interpreters, CCXML does not require an interpretive implementation. ES-CP can be compiled to a target language such as C, and thus in turn to machine code, so that CCXML documents which are static can be rendered once in machine code. For example, a CCXML implementation, for optimization purposes, could translate and compile frequently used CCXML documents on their way from the document server to the CCXML execution environment in order to avoid multiplying interpretive overhead by the number of lines that execute the same document.

The emphasis on efficiency in CCXML language is also shown by the avoidance of requirements which can only be implemented either by interpretation or by run-time evaluation.

The choice of an implementation strategy is up to the CCXML implementer and CCXML language is aimed to allow a range of design choices in order to accommodate implementations on a wide variety of platforms.

A CCXML implementation MUST support the ECMAScript Compact Profile.

3.4: Definitions

The following terms, which are used throughout this specification, are defined as:

• ECMAScript left-hand-side expression - defined in ECMA-262 [ ECMASCRIPT ] 11.2; this is an expression which produces a result to which a value can be assigned; an expression which is valid as the left hand operand of an assignment (=) operator;

  Several examples of left-hand-side expressions are as follows (left-hand-side expression in red):

  <?xml version="1.0" encoding="UTF-8"?>
  <ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <script>
  var simpleVar;
  var aaVar = Object();
  simpleVar            = 'Simple Expr';
  aaVar[0]          = 'Simple Expr';
  aaVar['arrayKey'] = 'Simple Expr';
  aaVar             = {callingDevice: 'notSpecified', callCharacteristics: 'voiceUnitCall'};
  </script>
  </ccxml>
  

• ECMAScript expression - defined in ECMA-262 [ ECMASCRIPT ] 11.1; this is an expression which produces a value; an expression which is valid on the right hand side of an assignment operator;

  Several examples of ECMAScript expressions are as follows (ECMAScript expression in red):

  <?xml version="1.0" encoding="UTF-8"?>
  <ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <script>
  var simpleVar;
  var aaVar = Object();
  simpleVar = 'hello world'; // simple string expression
  simpleVar = 'hello world'.length; // Calling a method that returns a 
                                    // number on the simple string object
  simpleVar = 5; // Simple number expression
  simpleVar = aaVar[0]; // Associative Array position expression
  simpleVar = aaVar['key']; // Associative Array named value expression
  simpleVar = myCoolFunction(); // Function return expression
  </script>
  </ccxml>
  

• ECMAScript variable name - defined in ECMA-262 [ ECMASCRIPT ] 7.6; this is any valid sequence of characters, known as an identifier, which can be used as a variable name, a property name, or a function name; this does not include any qualifiers, such as array or property accessors;

  Several examples of ECMAScript variable names are as follows (ECMAScript variable name in red):

  <?xml version="1.0" encoding="UTF-8"?>
  <ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <script>
  var simpleVar;
  var arrayVar = Object();
  simpleVar            = 'Simple Expr';
  arrayVar[0]          = 'Simple Expr';
  arrayVar['arrayKey'] = 'Simple Expr';
  arrayVar             = {callingDevice: 'notSpecified', callCharacteristics: 'voiceUnitCall'};
  </script>
  </ccxml>
  

• Empty ECMAScript object - an object returned by the new Object() ECMAScript expression; an object with no properties.
• CCXML variable name - a variable name, with optional dot separated qualification, declared in a CCXML <var> element or defined within ECMAScript code using the var keyword.
• Scope element - a CCXML element which defines a variable scope; <ccxml> and <transition> are CCXML scope elements. Variables which are defined within a scope element are not visible to variables defined in other scope elements.
• Time interval - CCXML uses the Cascading Style Sheets, Level 2 [ CSS2 ] time format. Time designations consist of a non-negative real number followed by a time unit identifier. The time unit identifiers are:
  • ms : milliseconds
  • s : seconds
  Examples include: "3s", "850ms", "0.7s", ".5s" and "+1.5s".
• CCXML Application/Program - A collection of CCXML documents that together create a complete application/program.
• CCXML Interpreter - The software that processes CCXML documents, executes commands and dispatches events. Also referred to as the "CCXML Platform".
• CCXML Platform - An implementation of CCXML for a specific network and environment, generally including a CCXML Interpreter as a component.
• CCXML Session - A single instance of a CCXML application. Can It can span multiple documents and phone calls. See Section 3.5 Session Life-Cycle for details.
• CCXML Parent Session - The CCXML session that created the currently running session using the <createccxml> element.
• Event - An action or occurrence to which an application can respond. Examples of events are incoming phone calls, dialog actions or user defined events. Events in CCXML are modeled as ECMAScript objects and can contain complex values. All events in CCXML have a standard set of properties as defined in section 9.4.2: Standard Event Attributes .
• Associative array - An associative array (also known as a map, lookup table, or dictionary) is an abstract data type composed of a collection of keys and a collection of values, where each key is associated with one value. Associative arrays in ECMAScript are implemented using an Object with sub properties.
• Class - CCXML Classes are ECMAScript Constructor objects (see section 4.2.1 of the ECMAScript specification for more information) to help in creating the appropriate type of object instances (such as connection, conference and dialog). The Constructor object may contain properties that are not inherited by children objects (such as Connection.states). The constructor object MAY also contain a Prototype property to allow properties that are inherited. Platforms MAY choose to add properties to classes. By convention, the properties MUST begin with an underscore, "_", to identify them as platform-dependent. All defined objects in this specification must be initiated via one of these classes. An example would be 'MyConnection = new Connection()'. Reserved ECMAscript ECMAScript property ' prototype.constructor ' MUST reference the class constructor object, so in the example above, 'MyConnection.prototype.constructor == Connection'. The Connection class MUST initiate connection objects, the Dialog class MUST initiate dialog objects and the Conference class MUST initiate conference objects.

3.5: Session Life-Cycle

3.5.1: Startup

A CCXML session can be started for the following reasons:

• A new incoming phone call coming into the platform.
• A CCXML application executing a <createccxml> .
• An external session launch request coming into the platform.

To create a CCXML session, the URI for the initial CCXML document must be known, along with any fetching parameters affecting how that CCXML document is retrieved. For incoming calls, the selection of the initial URI and fetching parameters is platform-dependent, and MAY be based on information from the incoming call. Sessions created via <createccxml> and the session creation event I/O processor determine the initial URI and fetching parameters as stated in this specification.

When a session is started due to an incoming call it has ownership of the new Connection that caused it to be created . created. The new CCXML session will be responsible for processing the Connection state events and performing the Connection actions. If the session was started because of a <createccxml> , it will start without ownership of any event endpoints. In the case of an external session launch the session will not own any event endpoints.

A CCXML application can determine the reason its session was started by evaluating the contents of the session.startupmode session variable that is defined in the Session Variables section.

3.5.2: Shutdown

A CCXML session can end in one of the following ways:

• The CCXML application executes an <exit> .
• An unhandled "error.*" event.
• An unhandled "ccxml.kill" event.
• A "ccxml.kill.unconditional" event.

When a CCXML session ends, all active connections, conferences and dialogs that are owned by that session are automatically terminated by the platform.

3.5.3: Session Life-Cycle Diagrams

The following diagrams illustrate the session life-cycle of several different scenarios. These diagrams do not show all possible scenarios but rather show some of the most common ones that CCXML applications may encounter.

3.5.3.1: session can live before and after active connections (or no connections at all)

A CCXML session does not necessarily need to have any connections associated with it. After starting, a session may acquire connections as a result of <createcall> or <move> requests.

3.5.3.2: connection life shorter than session

In this example, the session is started due to an incoming call. A connection is typically shorter than a session. A session does not end when a connection terminates.

3.5.3.3: session ends, kills all active connections

When a session ends, any resources, including connections owned by that session are terminated.

3.5.3.4: session can have multiple sequential connections

A session can have multiple sequential connections

3.5.3.5: session can have multiple sequential connections and multiple concurrent connections

In addition to having multiple sequential connections, a session can have multiple concurrent connections.

3.5.3.5: move a connection to a newly created session

A connection can be moved from one CCXML session to another session. In the figure below, CCXML session (1) creates a new CCXML session (2) via <createccxml> . Then, the connection is moved from the original CCXML session to the new session.

3.5.3.7: move a connection to a "master" session

A connection can be moved from one CCXML session to another session, such as a "master" session.

3.5.3.8: optional "master" session for inbound call handling

Implementations MAY , as a platform-specific optimization, choose to deliver more than one inbound call to a single "master" session. This can be viewed as equivalent to sessions handling incoming calls performing a <move> , as described in 3.5.3.7, of the new Connection (including the connection.alerting event) to the single "master" CCXML session.

The default inbound call handling behavior for CCXML implementations is to create a new CCXML session and deliver the connection.alerting event to it. If a platform supports delivery of multiple inbound calls to a single session, the way this is configured is implementation specific.

3.5.3.9: ccxml.kill.unconditional event raised

If at anytime a ccxml.kill.unconditional event is raised by the underlying implementation, the CCXML session is immediately terminated and all active connections, conferences and dialogs that are owned by that session are automatically terminated by the platform.

3.5.3.10: Normal session shutdown requested by the platform

If at anytime the platform wishes to terminate a CCXML session it MUST raise a ccxml.kill event to inform the CCXML application. The normal response to this event is for the CCXML application to perform any clean up and termination of current active connections, conferences or dialogs and then execute an <exit> element.

If the CCXML application does not respond to the ccml.kill event in a timely manner the platform MAY then raise a ccxml.kill.unconditional event to immediately terminate the CCXML session and all active connections, conferences, and dialogs that are owned by the session.

4: Simple Examples

4.1: Hello World

This simple CCXML document shows an example of a "hello world" application that is started due to an incoming call where the application simply assigns a value to a variable, prints a message to the platform log and exits:

<?xml version="1.0" encoding="UTF-8"?>

  <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <eventprocessor>  
    <transition event="connection.alerting">
      <var name="MyVariable" expr="'This is a CCXML Variable'"/>
      <log expr="'Hello World. I just made a variable: ' + MyVariable"/>
      <log expr="'Lets hang up on this incoming call.'"/>
      <exit/>
    </transition>    
  </eventprocessor>
</ccxml>

4.2: Accept or Reject a Call

This CCXML document shows an example of how to process a incoming call event and answer or reject the call based on the phone number of the calling party:

<?xml version="1.0" encoding="UTF-8"?>

      <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <eventprocessor>
  
    <transition event="connection.alerting">
      <log expr="'The number called is' + event$.connection.remote + '.'"/>
      <if cond="event$.connection.remote == 'tel:+18315551234'">
        <log expr="'Go away! we do not want to answer the phone.'"/>
        <reject/>
      <else/>
        <log expr="'We like you! We are going to answer the call.'"/>
        <accept/>
      </if>
    </transition>
    <transition event="connection.connected">
      <log expr="'Call was answered,Time to disconnect it.'"/>
      <disconnect/>
    </transition>
    <transition event="connection.disconnected">
      <log expr="'Call has been disconnected. Ending CCXML Session.'"/>
      <exit/>
    </transition>
    
  </eventprocessor>
</ccxml>

</ccxml>  

4.3: Simple Dialog

This is an example of running a simple VoiceXML dialog from CCXML. The application answers an incoming phone call and then connects it to a VoiceXML dialog that returns a value that is then logged to the platform:

dialog.ccxml:

<?xml version="1.0" encoding="UTF-8"?>

  <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <!-- Lets declare our state var -->
  <var name="state0" expr="'init'"/>
  
  <eventprocessor statevariable="state0">
    <!-- Process the incoming call -->  
    <transition state="init" event="connection.alerting">
      <accept/>      
    </transition>
    <!-- Call has been answered -->  
    <transition state="init" event="connection.connected">
      <log expr="'Houston, we have liftoff.'"/>
      <dialogstart src="'dialog.vxml'"/>
      <assign name="state0" expr="'dialogActive'" />
   
    </transition>
    <!-- Process the incoming call -->  
    <transition state="dialogActive" event="dialog.exit">
      <log expr="'Houston, the dialog returned [' + event$.values.input + ']'" />
      <exit /> 
    </transition>
    <!-- Caller hung up. Lets just go on and end the session -->
    <transition event="connection.disconnected">
      <exit/>
    </transition>
    <!-- Something went wrong. Lets go on and log some info and end the call -->
    <transition event="error.*" >
      <log expr="'Houston, we have a problem: (' + event$.reason + ')'"/>
      <exit/>
    </transition>
  </eventprocessor>    
</ccxml>
  

dialog.vxml:

<?xml version="1.0"?>

  <?xml version="1.0"?>

<vxml xmlns="http://www.w3.org/2001/vxml" version="2.0">
  <form id="Form">
    <field name="input" type="digits">
      <prompt>
        Please say some numbers ...
      </prompt>
      <filled>                            
        <exit namelist="input"/>
      </filled>
    </field>
  </form>
</vxml>

</vxml>  

5: CCXML Elements Listing

6: Document Control Flow and Execution

6.1: Overview

A CCXML session begins with the execution of a CCXML document. The flow of the execution can be changed with the help of <if> , <elseif> , <else> , <fetch> , and <goto> . Most of a CCXML session's execution will take place within an <eventprocessor> , which processes a stream of incoming events.

A CCXML session can consist of multiple CCXML documents, traversed by use of <goto> and <fetch> .

A new CCXML session has a new session object (session.*). ( session.* ). A CCXML session can contain multiple active connections.

A CCXML session may launch a new CCXML session using <createccxml>. <createccxml> . The new CCXML session executes in an independent context and variable space from the original CCXML session, completely independent of the lifetime of the original session. Sessions can communicate by sending messages via <send> .

This media type should be used for a XML document containing CCXML content. content, see Appendix M.

6.2: Elements

This section details the CCXML elements for control flow and execution.

6.2.1: <ccxml>

6.2.1.1: Overview

This is the parent element of a CCXML document and encloses the entire CCXML script in a document. When a <ccxml> is executed, its child elements are collected logically together at the beginning of the document and executed in document order before the target <eventprocessor> . This is called document initialization.

The <ccxml> can designate the CCXML namespace. This can be achieved by declaring an xmlns attribute or an attribute with an " xmlns " prefix. See [XMLNS] for details. Note that when the xmlns attribute is used alone, it sets the default namespace for the element on which it appears and for any child elements. The namespace URI for CCXML is "http://www.w3.org/2002/09/ccxml".

6.2.1.2: <ccxml> Attribute Details

6.2.2: <meta>

6.2.2.1: Overview

The <metadata> and <meta> are containers in which information about the document can be placed. The <metadata> provides more general and powerful treatment of metadata information than <meta> by using a metadata schema.

A <meta> declaration associates a string to a declared meta property or declares " http-equiv " content. Either a name or http-equiv attribute is REQUIRED . It is an error to provide both name and http-equiv attributes. A content attribute is REQUIRED . The http-equiv attribute has a special significance when documents are retrieved via HTTP . Although the preferred method of providing HTTP header information is by using HTTP header fields, the " http-equiv " content MAY be used in situations where the CCXML document author is unable to configure HTTP header fields associated with their document on the origin server, for example, cache control information. Note that, as with <meta> in HTML documents [HTML] , HTTP servers and caches are not REQUIRED to introspect the contents of <meta> in CCXML documents and thereby override the header values they would send otherwise.

Informative: This is an example of how <meta> can be included in a CCXML document to specify a resource that provides additional metadata information and also indicate that the document MUST NOT be cached.

<?xml version="1.0" encoding="UTF-8"?>

  <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0"
       xmlns="http://www.w3.org/2002/09/ccxml">
       <meta name="seeAlso"
content="http://example.com/my-ccxml-metadata.xml"/>

       <meta http-equiv="Cache-Control" content="no-cache"/>
</ccxml>
  

<meta> is an empty element.

6.2.2.2: <meta> Attribute Details

6.2.3: <metadata>

6.2.3.1: Overview

<metadata> is a container in which information about the document can be placed using a metadata language. Although any metadata language can be used within <metadata> , it is recommended that the Resource Description Format [RDF] be used in conjunction with the general metadata properties defined by the Dublin Core Metadata Initiative [DC] .

RDF [RDF-SYNTAX] is a declarative language and provides a standard way for using XML to represent metadata in the form of statements about properties and relationships of items on the Web. A recommended set of generally applicable metadata properties (e.g., " title ", " creator ", " subject ", " description ", " copyrights ", etc.) is the Dublin Core Metadata Element Set [DC] , used in the example below.

Document properties declared with <metadata> can use any metadata schema.

Informative: This is an example of how <metadata> can be included in a CCXML document using the Dublin Core version 1.0 RDF schema [DC] describing general document information such as title, description, date, and so on:

<?xml version="1.0" encoding="UTF-8"?>
<ccxml version="1.0"
       xmlns="http://www.w3.org/2002/09/ccxml">
  <metadata>
   <rdf:RDF
       xmlns:rdf = "http://www.w3.org/1999/02/22-rdf-syntax-ns#"
       xmlns:dc = "http://purl.org/dc/elements/1.1/">
   <!-- Metadata about CCXML document -->
   <rdf:Description rdf:about="http://www.example.com/meta.ccxml"
       dc:title="Hamlet-like Soliloquy"
       dc:description="Aldine's Soliloquy in the style of Hamlet"
       dc:publisher="W3C"
       dc:language="en"
       dc:date="2002-11-29"
       dc:rights="Copyright 2002 Aldine Turnbet"
       dc:format="application/ccxml+xml" >
       <dc:creator>William Shakespeare</dc:creator>
       <dc:creator>Aldine Turnbet</dc:creator>
   </rdf:Description>
  </rdf:RDF>
 </metadata>
</ccxml>

The following CCXML elements can occur within the content of <metadata> : none .

6.2.3.2: <metadata> Attribute Details

6.2.4: <if>

6.2.4.1: Overview

<if> is a container for conditionally executed elements. <else> and <elseif> can optionally appear within an <if> as immediate children, and serve to partition the elements within an <if> . <else> and <elseif> have no content. <else/> is a synonym for <elseif cond="true"/> .

Each partition within an <if> is preceded by an element having a cond attribute. The initial partition is preceded by the <if> and subsequent partitions by <elseif> s (or <else> s). The first partition in document order with a cond that evaluates to true is selected. <else> always evaluate to true . A partition MAY be empty.

If an <if> has no immediate <elseif> or <else> children, the full contents of the <if> will be selected when the cond attribute is true .

<else> was chosen to match similar concepts in other languages, and supports examples such as

<if cond="...">
  <!-- selected when <if cond> is true -->
  <else/>

<else/>

    <!-- selected when <if cond> is false -->
  </if>.

</if>

However, <else> is a synonym for <elseif cond="true"/> , so an example such as

<if cond="...">
  <!-- selected when <if cond> is true -->
<else/>
  <!-- selected when <if cond> is false -->
<else/>
  <!-- never selected -->
</if>

<if cond="...">
  <!-- selected when <if cond> is true -->
<elseif cond="true"/>
  <!-- selected when <if cond> is false -->
<elseif cond="true"/>
  <!-- never selected -->
</if>.

</if>

With this definition for <else> , CCXML provides familiar if/elseif/else semantics, but conforms to the rules of valid XML [XML] documents.

6.2.4.2: <if> Attribute Details

6.2.5: <elseif>

6.2.5.1: Overview

An <elseif> partitions the content of an <if> , and provides a condition that determines the selection of the partition it begins. <elseif> can appear optionally as an immediate child of an <if> .

6.2.5.2: <elseif> Attribute Details

6.2.6: <else>

6.2.6.1: Overview

<else> is a synonym for <elseif cond="true"/> .

6.2.6.2: <else> Attribute Details

6.2.7: <fetch>

6.2.7.1: Overview

<fetch> is used to asynchronously fetch content identified by the attributes of the <fetch> . The fetched content may either be a CCXML document, or script content. content or other document types supported by the CCXML platform. Content that has been acquired using <fetch> is accessible through other elements defined by CCXML. CCXML and via the content attribute of the fetch.done event. Execution returns from the element immediately, and the CCXML application can continue on while the platform works to fetch the identified resource. When the fetch request has been completed, an event is generated against the session that initiated the fetch. The event is one of fetch.done , which indicates that If the identified requested content was fetched successfully, or error.fetch , indicative and in the case of CCXML content, was successfully parsed and validated, a failure to fetch fetch.done MUST be generated. If the requested content. content failed to be fetched, an error.fetch MUST be generated. Note that even if content is successfully fetched, errors in processing fetched content (for instance, a CCXML document with a syntax error) may result in an error.fetch being thrown.

The fetch request is local to the session that initiated the <fetch> , and is referenced through a unique identifier generated by the CCXML platform. The application may obtain the unique identifier for a fetch request by providing an ECMAScript left-hand-side expression in the fetchid attribute when the fetch is performed. The fetch identifier can also be obtained as a property of the fetch.done event. The application uses the fetch identifier in any CCXML elements that reference fetched content, currently <goto> and <script> .

Fetched content that could be referenced via a fetchid (eg, a CCXML or ECMAScript Document) that was fetched in processed mode has a lifetime that is limited to that of the document in which it is fetched. Therefore, following a transition to a new CCXML document using <goto> , content fetched in the scope of the current document is no longer accessible. Note that this should not be taken to preclude platform-level optimizations or caching of resources that are fetched multiple times.

All other content is only accessible via the content attribute of the fetch.done event. If the application needs access to the content outside of the fetch.done event transition it is the responsibility of the application to save a reference in an appropriately scoped variable.

Content referenced via fetchid (for example via <script> or <goto> ) MUST be fetched using processed mode.

The use of <fetch> to obtain content does not compel the application to make use of that content. However, it is wasteful of system resources to fetch resources that are not used. Platforms are responsible for clearing out unused fetch resources, and may impose limits on the resources that can be fetched by a single session.

The "http" URI scheme MUST be supported by CCXML platforms, the "https" protocol should be supported and other URI protocols may be supported.

If the platform does not support the content type returned from a <fetch> request but the fetch does successfully complete (for example HTTP 2xx response code) the platform MUST still throw a fetch.done event for the fetchid .

If the platform implements a security model (such as Access Control for Cross-site Requests [ACCESS-CONTROL] ) and the request is denied due to the security model an error.notallowed event MUST be thrown.

INFORMATIVE NOTE: It is expected in future versions of CCXML additional forms of processed fetch modes will be added. Examples of these could be dom or e4x mappings of raw XML content.

6.2.7.2: <fetch> Attribute Details

6.2.8: <goto>

6.2.8.1: Overview

<fetch> , in conjunction with <goto> , is used to transfer execution to a different CCXML document in a multi-document CCXML application. The <fetch> tells the platform to find, load, and parse a given CCXML document. After the fetch completes, the CCXML application can then issue a <goto> to execute the now-fetched document.

Below is a small snippet of code from the CCXML application's event handler. We execute a <fetch> operation, and continue on to assign to a state variable, and maybe handle more events. Eventually, the fetch completes, the CCXML platform services the event, and the application performs the <goto> .

<fetch next="'http://www.example.com/control.ccxml'"/>
<--control continues here->
<assign name="state_var" expr="'fetch_wait'"/>
</transition>
<!-- ……… -->

<!-- ……… -->

<transition state="fetch_wait" event="fetch.done"/>
<goto fetchid="event$.fetchid"/>
</transition>

A <goto> transfers control to the document obtained through a fetch request, using the platform-generated unique identifier associated with that fetch request. The fetch completion event MUST have arrived before the <goto> is executed, otherwise, an error.semantic event is generated. If the fetched content referenced by the fetch identifier is not a CCXML document, or the fetch identifier is invalid and does not correspond to any fetch request, this also results in an error.semantic event.

When a <goto> is executed, the target document replaces the current document in its session. Event sources associated with this session are inherited by the target document. Execution of the current document terminates.

6.2.8.2: <goto> Attribute Details

6.2.8.3: <fetch> and <goto> Example

The following code shows the use of the <fetch> and <goto> elements along with the fetchid attribute to handle more complex fetching situations:

<?xml version="1.0" encoding="UTF-8"?>

  <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <!-- var to hold the value of the fetch 
          identifier that we care about -->
  <var name="myGoodFetchID"/>
  <eventprocessor>  
    <transition event="ccxml.loaded">
       <!-- stick the value of the fetch 
               identifier in the myGoodFetchID var -->
       <fetch fetchid="myGoodFetchID" 
                 next="'http://www.example.com/goodfetch.ccxml'"/>
       <!-- do not bother saving the fetch id's for these, 
               we would just ignore them anyway -->
       <fetch next="'http://www.example.com/fakefetch1.ccxml'"/>
       <fetch next="'http://www.example.com/fakefetch2.ccxml'"/>
    </transition>
    <transition event="fetch.done">
       <if cond="myGoodFetchID == event$.fetchid">
          <!-- only matched if we have fetched 
                  http://www.example.com/goodfetch.ccxml -->
          <goto fetchid="event$.fetchid"/>
       </if>
    </transition>
    <transition event="error.fetch">
          <!-- Ignore bad fetches in this example -->
    </transition>
  </eventprocessor>
</ccxml>
  

6.2.9: <createccxml>

6.2.9.1: Overview

<createccxml> is used to create another CCXML session, which begins execution with the document identified by this element. The new CCXML session has no relation to its creator once spawned, and has a wholly separate lifetime and address space.

Execution returns from the <createccxml> element immediately, and the CCXML interpreter can continue on while the new CCXML session is established and loads its initial document. If the new session is successfully established or a failure occurs an event is generated and is delivered to the session that executed the <createccxml> element.

6.2.9.2 <createccxml> Attribute Details

6.2.10: <exit>

6.2.10.1: Overview

<exit> ends execution of the CCXML session. All pending events are discarded, and there is no way to restart CCXML execution.

6.2.10.2 <exit> Attribute Details

A CCXML document executing the <exit> will generate a ccxml.exit event to the parent session. The exiting document will be identified on the exit event by its session ID.

6.2.11: <log>

6.2.11.1: Overview

<log> allows an application to generate a logging or debug message which a developer can use to help in application development or post-execution analysis of application performance. The manner in which the message is displayed or logged is platform-dependent. The usage of label is platform-dependent. The use of <log> SHOULD have no other side-effects on interpretation. <log> is an empty element.

6.2.11.2 <log> Attribute Details

6.3: Events

6.3.1: Overview

CCXML allows operations such as document fetching, startup and shutdown to execute independently. CCXML events that describe these operations are defined below:

6.3.2: fetch.done - Fetch Completion Event

This event is generated when a fetch request completes successfully . successfully. It is delivered to the document which issued the request.

The fields of this event are:

6.3.3: error.fetch - Fetch Error Event

This event is generated when a fetch request does not successfully complete. It is delivered to the document which issued the request.

The fields of this event are:

6.3.4: ccxml.exit - CCXML Document Exit Event

This event is generated when a CCXML document executes an <exit> , having an unhandled "error.*" or ccxml.kill event. This event session is only generated when terminated for any reason, if the terminated session executing the <exit/> has a parent session. This event is sent to the parent session and not the session executing the <exit/> . that was terminated.

The fields of this event are:

6.3.5: ccxml.loaded - CCXML Document Loaded Event

This event is thrown once the document is parsed and ready for execution (document initialization occurs between the fetched and loaded events). The CCXML platform MUST generate this event when the CCXML document is first loaded, both at session startup and after transferring control to a new document with the <goto> . This event would be processed after the platform had executed the document initialization including executing any elements under the <ccxml> and before events such as connection.alerting which may have triggered creation of the session . session.

The fields of this event are:

6.3.6: ccxml.kill - CCXML kill Event

The kill event can be used by the platform to terminate a session without an explicit <exit>. <exit> . There are two versions of this event: catchable, and non-catchable.

The ccxml.kill event can be caught, typically to perform a clean-up operation at the end of a session. If the event is caught the session will not be terminated unless the an <exit> element is processed. If the event is not caught the session will be terminated and all active connections, conferences and dialogs that are owned by that session will be automatically terminated by the platform.

Unlike other events, the ccxml.kill.unconditional event is the only event that cannot be caught by an application; it will unconditionally terminate the session and all active connections, conferences and dialogs that are owned by that session will be automatically terminated by the platform.

Note that while the normal cause of a ccxml.kill or ccxml.kill.unconditional event being queued to a session is that the platform wishes to terminate the session, it is legal for any event I/O processor to generate a ccxml.kill or ccxml.kill.unconditional event. For instance, it is legal for one CCXML session to unconditionally kill another session by sending a ccxml.kill.unconditional event using <send> . Note, however, that platforms may impose rules that prevent one session from arbitrarily killing another (to prevent malicious applications, for instance).

The fields of this event are:

6.3.7: ccxml.created - CCXML Session Create Completion Event

This event is generated when a <createccxml> request completes successfully. It is delivered to the document which issued the request and indicates that the new session has retrieved the specified initial CCXML document, parsed and has begun execution of it by sending the ccxml.loaded event to the new session.

The fields of this event are:

6.3.8: error.createccxml - CCXML Session Create Failed Event

This event is generated when a <createccxml> request fails to complete. It is delivered to the document which issued the request and indicates that the new session has not been created.

The fields of this event are:

6.3.9: error.unsupported - CCXML Unsupported Operation

This event is generated when an operation that is not supported by the platform is executed.

The fields of this event are:

7: Dialogs

7.1: Overview

CCXML does not provide any mechanism for interacting with callers but relies on separate dialog environments such as VoiceXML [ VOICEXML ]. Whenever interaction with a caller is required, a CCXML session can create a separate dialog to perform that interaction. After the dialog interaction is complete, an asynchronous event is sent to the CCXML session which can use any results returned by the dialog environment to decide what should happen next.

Dialogs initiated by CCXML sessions are not tied to any single dialog language or technology. Any dialog system which fulfils CCXML's requirements MAY be used for interaction with the caller. Examples of dialog systems include VoiceXML, SALT [ SALT ], SALT, traditional IVR , 3GPP MRF as well as simple media handling systems for fax, media playback and recording, DTMF detection, answer-machine detectors, etc. A CCXML platform MAY support interaction with several dialog systems with the selection of the particular technology being based on the MIME type specified when the dialog is initiated.

All CCXML elements that manipulate dialogs are asynchronous with control returning immediately to the CCXML session after the operation is initiated. The CCXML session is notified when the dialog operation successfully completes, or fails, by an asynchronous event. Execution of a <dialogprepare> or <dialogstart> without preparation creates a Dialog Object immediately, allowing the Dialog Object to be accessed following the <dialogprepare>/<dialogstart> , within the same transition. However, existence of the Dialog Object does not guarantee anything about the success of the <dialogprepare>/<dialogstart> request.

A CCXML program initiates a dialog using the <dialogstart> element. Execution of this element connects a dialog environment to a connection and instructs it to start interacting with the caller. For some dialog environments it may take some time to initialize the dialog environment and thus the use of the <dialogstart> element alone may cause the caller to hear silence, or "dead air". To avoid this situation CCXML provides an ability to ready a dialog environment prior to connecting and starting it, this is done using the <dialogprepare> element. Any dialog that has been either started with <dialogstart> , or prepared with <dialogprepare> can be terminated using the <dialogterminate> element. CCXML implementations MUST support the <dialogprepare> , <dialogstart> , and <dialogterminate> elements though the exact behaviour behavior may vary depending on the dialog environments supported.

The following examples illustrate the valid use patterns for these three elements. Firstly the normal case of preparing a dialog, starting it, then optionally terminating it before normal completion. This example illustrates the use of <dialogprepare> to ready a dialog while the call is left in alerting state. When the alerting notification arrives the script executes a <dialogprepare> to prepare a dialog and associate it with the connection. When the dialog is prepared the script executes an <accept> to connect the call and then when the connection transitions to connected state, a <dialogstart> element is used to execute the previously prepared dialog.

<transition event="connection.alerting">
    <dialogprepare src="..." connectionid="event$.connectionid"/>
</transition>
<transition event="dialog.prepared">
    <accept connectionid="session.dialogs[event$.dialogid].connectionid"/>
</transition>
    
<transition event="connection.connected" >
    <dialogstart prepareddialogid="event$.connection.dialogid" 
                    connectionid="event$.connectionid"/>
</transition>
    
(optionally)
<transition event="???">
    <dialogterminate dialogid="..." />
</transition>

The next example shows a single step dialog invocation without dialog preparation. In this case a connection in alerting state is accepted and, when the transition to connected state occurs, a <dialogstart> element is used to start the dialog.

<transition event="connection.alerting">
    <accept connectionid="event$.connectionid"/>
</transition>
    
<transition event="connection.connected">
    <dialogstart src="..." connectionid="connectionid"/>
</transition>
    
(optionally)
<transition event="???">
    <dialogterminate dialogid="..." />
</transition>

This next example shows the preparation of a dialog for an outbound call:

<transition event="...">
    <dialogprepare src="..."> <!-- no connectionid -->
</transition>
<transition event="dialog.prepared" >
    <createcall dest="..." joinid="event$.dialogid"> <!-- use joinid -->
</transition>
<transition event="connection.connected">
    <dialogstart prepareddialogid="event$.connection.dialogid"
                 connectionid="event$.connectionid"/>
</transition>

The final example shows the case where a dialog which has been previously prepared is cancelled before a <dialogstart> has been issued. A dialog may be terminated when it is in the prepared state or while it is being prepared such as might be the case if the caller hangs up at some arbitrary point. In this case the <dialogterminate> may be executed before or after the dialog.prepared event is processed.

<transition event="connection.connected">>
    <dialogprepare src="..." connectionid="event$.connectionid"/>
</transition>
<transition event="connection.disconnected" >
    <dialogterminate dialogid="event$.connection.dialogid" />
</transition>

It is possible for Dialogs to exist that are not joined to a Connection or a Conference. For example, this could be due to a Connection disconnecting, or due to the application performing an <unjoin/> operation.

7.2: Elements

7.2.1: <dialogprepare>

7.2.1.1: Overview

<dialogprepare> is used to get an appropriate dialog handler ready to process, it is used as the precursor to a <dialogstart> request. The element includes a URI reference to the initial document for the dialog. The new dialog is prepared on a separate logical execution thread (this MAY may be a thread, process, or system depending upon platform implementation) and does not block the processing of further events by the CCXML session. The use of the <dialogprepare> element is entirely optional, applications may choose to simply use <dialogstart> without prior preparation.

Optionally the new dialog may be associated with a connection by specifying the connectionid attribute, or with a conference by specifying the conferenceid attribute.

When preparation of the dialog completes successfully a dialog.prepared event MUST be posted to the event queue of the CCXML session. If however the dialog cannot be prepared for any reason, an error.dialog.notprepared event MUST be posted. However should the dialog be terminated using <dialogterminate> while it is being prepared the platform MUST only post a dialog.exit event.

CCXML implementations MUST support dialog preparation though the processing carried out as part of a <dialogprepare> request is dialog manager specific. In the case of a dialog manager that does not support preparation, the CCXML implementation MUST as a minimum, note the values provided in the <dialogprepare> attributes, create a Dialog object, and return a new unique value to the location defined by the dialogid attribute and throw a dialog.prepared event.

If the connectionid or conferenceid attributes are specified on <dialogprepare> the Dialog Object's connectionid/conferenceid properties MUST be set to the appropriate values. Additionally the dialogid attribute of the Connection Object MUST be set to the dialogid of the new Dialog.

The CCXML session selects what it believes to be the appropriate dialog manager based on the MIME type specified by the type attribute without retrieving the resource specified by the src URI. If, when the dialog manager retrieves the content, it finds the MIME type, as specified by the HTTP headers, differs from that specifed by the type attribute, it MUST raise an error.dialog.notprepared event with a reason indicating the type mismatch. The dialog manager MUST NOT ignore the type mismatch or render the resource as a different type based on the HTTP headers or on inspection of the document data. Refer to the W3C guidelines for client handling of MIME types [ MIME-TAG ] for further information.

7.2.1.2: <dialogprepare> Attribute Details

7.2.2: <dialogstart>

7.2.2.1: Overview

<dialogstart> is used to start a dialog and associate the dialog with a connection or conference. (See Section 10 for a discussion of connections and bridges). The element includes MUST include either a URI reference to the initial document for the dialog or the identity of a previously prepared dialog. The dialog executes on a separate logical execution thread (this MAY may be a thread, process, or system depending upon platform implementation) and does not block the processing of further events by the CCXML session.

If the dialog cannot be started for any reason, an error.dialog.notstarted event MUST be posted to the event queue of the CCXML session that processed the <dialogstart> request otherwise, request, otherwise a dialog.started event MUST be posted to indicate that the dialog has started successfully. When the dialog completes, a dialog.exit event MUST be posted.

If the connectionid attribute of <dialogstart> is specified, and if the dialog language allows access to telephony variables such as ANI, DNIS and UUI, values of these variables will MUST be propagated from the specified connection to the dialog application.

If the prepareddialogid attribute is specified and any attribute values conflict with the values specified in the <dialogprepare> element this MUST result in the throwing of an error.dialog.notstarted event.

Dialogs MAY only be started on connections while they are in the ALERTING or CONNECTED states. For connections in the ALERTING state media MAY not be sent or received by the connection making the bridge effectively full-duplex, half-duplex or silent depending on the underlying telephony environment and the restrictions imposed by it.

It is not possible to start a dialog that is not joined to a connection or a conference. For dialogs prepared using <dialogprepare> with neither connectionid nor conferenceid specified and dialogs being started directly without preparation, if neither connectionid nor conferenceid are specified then execution of the dialogstart MUST fail, with an error.dialog.notstarted event generated.

The CCXML session selects the appropriate dialog manager based on the MIME type specified by the type attribute without retrieving the resource specified by the src URI. If, when the dialog manager retrieves the content, it finds the MIME type, as specified by the HTTP headers, differs from that specifed by the type attribute, it MUST raise an error.dialog.notstarted event with a reason indicating the type mismatch. The dialog manager MUST NOT ignore the type mismatch or render the resource as a different type based on the HTTP headers or on inspection of the document data. Refer to the W3C guidelines for client handling of MIME types [ MIME-TAG ] for further information.

7.2.2.2: <dialogstart> Attribute Details

7.2.3: <dialogterminate>

7.2.3.1: Overview

A CCXML document may decide that it wants to terminate a currently executing dialog, to throw away a previously prepared dialog, or to terminate the preparation of a dialog. This is accomplished using the <dialogterminate> element. When the CCXML interpreter encounters a <dialogterminate> element, it MUST send a terminate request to the specified dialog.

A dialog terminated due to the processing of a <dialogterminate> element MAY still return data to the CCXML application using a dialog.exit event if the value of the immediate attribute is false or unspecified. The details of the data returned are dialog environment specific.

If the immediate attribute is set to true the dialog MUST NOT return data to the CCXML application and the CCXML interpreter MUST post a dialog.exit event immediately.

The platform MUST implicitly tear down any existing bridges to the dialog and send a conference.unjoined to the CCXML document once the media paths have been freed.

7.2.3.2: <dialogterminate> Attribute Details

7.3: Events

7.3.1: Overview

The majority of communication between CCXML interpreter sessions and dialogs is by way of events. Dialog environments post events to the CCXML interpreter event queue and a CCXML application MAY send an event to a dialog. How this is handled on the dialog side is dialog manager and CCXML interpreter dependent. On the CCXML side it is done by using <send> and passing in the dialogid that was received as a result of processing a <dialogstart> or a <dialogprepare> . .

The following are the CCXML events related to dialogs:

7.3.2: dialog.started

The dialog.started event MUST be thrown when a dialog is successfully started. The fields available in the event are:

7.3.3: dialog.exit

The dialog.exit event MUST be thrown when a dialog terminates either normally or following a <dialogterminate> request. Termination of a dialog always results in a single dialog.exit event; if normal dialog termination and a <dialogterminate> occur simultaneously, the dialog.exit event will reflect the condition that the platform processes first. For example, if a dialog.exit event is thrown based on normal termination, and the platform subsequently processes a <dialogterminate> request made by the application, it will discard the <dialogterminate> and will not generate a second dialog.exit event. The fields available in the event are:

7.3.4: dialog.disconnect

The dialog.disconnect event represents a request by a dialog to disconnect the Connection or Conference with which it is presently associated. The actual handling of this event is determined entirely by the running CCXML application, which may disconnect the associated Connection, detach the dialog from the Connection/Conference, or elect to ignore the dialog.disconnect event altogether. The fields available in the event are:

7.3.5: dialog.transfer

The dialog.transfer event represents a request by a dialog to transfer the Connection or Conference with which it is presently associated to a new destination. The actual handling of this event is determined entirely by the running CCXML application, which may perform a <redirect> on the associated Connection, use <createcall> and <join> to establish a bridged transfer, handle the transfer request in some other way, or elect to ignore the dialog.transfer event altogether. The properties of the dialog.transfer event reflect the information dialog environments are most likely to supply when requesting a transfer, but are up to the running CCXML application to interpret. The fields available in the event are:

7.3.6: dialog.terminatetransfer

The dialog.terminatetransfer event represents a request by a dialog to terminate an ongoing transfer for example due to a "hotword" recognition. The actual handling of this event is determined entirely by the running CCXML application, which may terminate the outgoing call leg and return the media stream of the original call to the dialog using the <join> element, handle the terminate transfer request in some other way, or elect to ignore the dialog.terminatetransfer event altogether. The properties of the dialog.terminatetransfer event reflect the information dialog environments are most likely to supply when terminating a transfer, but are up to the running CCXML application to interpret. The fields available in the event are:

7.3.7: error.dialog

The error.dialog event MUST be thrown when there was an unspecified error with the dialog (for example the dialog server crashed). The fields available in the event are:

The platform MUST implicitly tear down any existing bridges to the dialog and send a conference.unjoined to the CCXML document once the media paths have been freed. The fields available in the event are:

7.3.8: error.dialog.notstarted

The error.dialog.notstarted event MUST be thrown when the processing of a <dialogstart> element fails because the dialog cannot be started for some reason. The fields available in the event are:

7.3.9: blank

This section is intentionally left blank.

7.3.10: dialog.user.*

The dialog.user.* (where * is the name of the user defined event) MUST be thrown when a dialog sends a user/application-defined event. The fields available in the event are:

7.3.11: dialog.prepared

The dialog.prepared event MUST be thrown when a dialog has been successfully prepared following the execution of a <dialogprepare> element. The fields available in the event are:

7.3.12: error.dialog.notprepared

The error.dialog.notprepared event MUST be thrown when the processing of a <dialogprepare> element fails. The fields available in the event are:

7.4: Dialog Object Properties

An instance of the Dialog object is associated with each dialog created by <dialogstart> or <dialogprepare> and referenced in the session.dialogs associative array.

The Dialog Object is an extension of the Media Endpoint Object as defined in section 10.4.3.

7.5: Dialog class

All dialog objects MUST be initiated via the Dialog class. The Dialog class currently has no defined properties.

8: Variables and Expressions

8.1: Overview

CCXML expressions are valid ECMAScript [ ECMASCRIPT ] expressions, assignable to variables with valid ECMAScript names. For further details please see section 3.4 .

Many CCXML elements have multiple attributes that are ECMAScript expressions. The CCXML language does not guarantee a specific order of evaluation of these expressions. Also, some elements such as <transition> may or may not evaluate all attributes. Hence, attributes containing expressions with side-effects can lead to implementation specific behavior. It is RECOMMENDED that applications do not use ECMAScript expressions with side-effects in attributes.

If the implementation is unable to evaluate an ECMAScript expression it MUST throw an error.semantic event.

8.2: Elements

8.2.1: <assign> and <var>

8.2.1.1: Overview

Variables are declared using the <var> element and are initialized with the results of evaluating the OPTIONAL expr attribute as an ECMAScript expression. If the expr attribute is not present in the <var> declaration, the variable is initialized to ECMAScript undefined . The values of variables MAY be subsequently changed with <assign> .

Variables are declared explicitly by <var> :

<var name="sessionid" />
<var name="currentstate" expr="'initial'" />

Variables declared without an explicit initial value MUST be initialized to the ECMAScript value undefined by the implementation.

It is illegal to make an assignment to a variable that has not been explicitly declared using <var> or a var statement within a <script> . Attempting to assign to an undeclared variable causes an error.semantic event to be thrown. Please see Section 9.5 for a detailed description of error events.

Note that when an ECMAScript object, e.g. " obj ", has been properly initialized then its properties, for instance " obj.prop1 ", can be assigned without explicit declaration. An attempt to declare ECMAScript object properties such as " obj.prop1 " results in an error.semantic event being thrown.

CCXML uses an ECMAScript scope chain (please see section 3.4 ) to allow variables to be declared at different levels of hierarchy in an application. For instance, a variable declared at ccxml (document) scope can be referenced anywhere within that document, whereas a local variable declared in a <transition> is only available within that element.

The implementation MUST define four scopes - session , application , ccxml and transition . The relationship between these scopes is shown below.

Variable Scoping

A description of the scopes is provided in the table below.

The implementation MUST instantiate a variable within the scope of the closest containing scope element. The fully-qualified name of a variable is the name of the variable's scope object prepended with a dot to the name of the variable. The implementation MUST allow reference to variables by their fully qualified names. The implementation MUST allow reference to variables without requiring use of their fully qualified names. In the case of like-named variables declared in different scopes, the implementation MUST reference the variable in the closest containing scope, unless the fully-qualified variable name is used.

The implementation MUST resolve variables by searching the enclosing transition scope first (if applicable) followed by the ccxml scope, the application scope and then the session scope, unless the variable reference is qualified with a scope prefix.

If the variable includes a scope prefix, the implementation MUST resolve the variable by searching the named scope.

If a variable is declared more than once, the implementation MUST evaluate the expr attribute of each subsequent declaration, and assign the result to the variable declared by the first <var> .

Variables can be assigned new values using <assign> :

        <assign name="currentstate" expr="'cleanup'" />

The implementation MUST evaluate the ECMAScript expression contained in the expr attribute of <assign> , and assign the results to the variable referenced in the name attribute.

The variable naming convention is as in ECMAScript, but names beginning with the underscore character ("_") and names ending with a dollar sign ("$") are reserved for internal use. CCXML variables MUST NOT contain ECMAScript reserved words. They MUST also follow ECMAScript rules for referential correctness. For example, variable names must be unique and their declaration MUST NOT include a dot - "var x.y" " var x.y " is an illegal declaration in ECMAScript. Variable names which violate naming conventions or ECMAScript rules MUST cause an ' error.semantic ' event to be thrown.

8.2.1.2: <var> Attribute Details

8.2.1.3: <assign> Attribute Details

8.2.2: <script>

8.2.2.1: Overview

<script> encloses computations written in the ECMAScript Compact Profile [ ECMA327 ] scripting language. The ECMAScript Compact Profile is a strict subset of the third edition of ECMA-262 [ ECMASCRIPT ]. It has been designed to meet the needs of resource-constrained environments. Special attention has been paid to constraining ECMAScript features that require proportionately large amounts of system memory, and continuous or proportionately large amounts of processing power. In particular, it is designed to facilitate prior compilation for execution in a lightweight environment. For specific details on what ECMAScript functions are not supported please take a look at ECMA-327 specification . [ ECMA327 ].

An implementation MUST support the ECMAScript Compact Profile and MAY support the full ECMA-262 ECMAScript specification.

The example <script> below defines a function that computes the greatest common factor of two integers:

<script>
<![CDATA[
function gcd(a, b)
{
 var t;
 if (a < 1 || b < 1)
   return -1;
 do
  {
   t = a % b;
   a = b;
   b = t;
  }
 while (b > 1);
 return (b == 0) ? a : b;
}
]]>
</script>

An implementation MUST support <script> within the <ccxml> element and in executable content. <transition> and <if> contain executable content. The implementation MUST evaluate <script> in a <ccxml> immediately after the document is loaded, along with any <var> and <assign> elements, in document order. When used as a child of the <ccxml> element, <script> cannot be used to execute dynamically fetched content obtained using <fetch> . The implementation MUST evaluate <script> in executable content as it is processed.

The ECMAScript contained within the <script> can declare variables with the ECMAScript var statement. Variables declared in this manner are declared in the scope of the closest containing scope CCXML element. They are known from the point of declaration to the end of the containing scope. The implementation MUST allow reference to these variables from CCXML and from ECMAScript, using either the fully-qualified variable name, or the declared name.

If the implementation is unable to run the script referenced it MUST throw an error.semantic event.

INFORMATIVE NOTE: The <script> element's resource loading model is a bit different than the rest of CCXML for a number of reasons. Because CCXML and ECMAScript applications can be CPU intensive to compile we define <script> 's src attribute (defining the URI of the document to load) to be a static string instead of a dynamically valued ECMAScript result. This allows implementations to load ECMAScript content at CCXML document load time and perform compiling and/or caching of the resulting ECMAScript code. We do however recognize that there are cases where a CCXML application needs to load a dynamic ECMAScript resource, for this reason applications can use the the <fetch> element to asynchronously load a resource and then execute it by referencing it's fetchid in the the <script> element.

8.2.2.2: <script> Attribute Details

8.3: Session Variables

Every CCXML session has a set of standard ECMAScript variables that are available to the program during execution called session variables. The session variables are defined by the CCXML implementation when the CCXML session is created and are read-only to the running script and cannot be modified by the CCXML program. New session variables cannot be declared by CCXML programs.

Session variable values visible to a CCXML application reflect the state of the executing CCXML session, current as of the occurrence of the event being processed. For example, when an application processes a Connection event such as connection.alerting , the value of the state property of the corresponding Connection object will be updated by the CCXML implementation so that if the CCXML program's event handler evaluates the state variable, it will evaluate to ALERTING , since connection.alerting always results in a transition to the ALERTING state. This is true even if the underlying connection is actually DISCONNECTED , in which case a connection.disconnected event would be queued. It is the responsibility of the CCXML implementation to control and update the session changes as they occur in the CCXML session. It is assumed that session changes are visible to the CCXML program as they occur. However, it is permissible for a CCXML implementation to optimize session changes by "lazy-binding" values as they are accessed or evaluated by a CCXML program, so as to minimize processing time. For example, an implementation might only update the current Connection states when a CCXML program evaluates the variable during execution time versus continually updating the Connection states inside ECMAScript scope as state changes. Regardless of when session variables are updated to reflect changes, the CCXML implementation is REQUIRED to provide the correct values when accessed by a CCXML program.

Variables defined in the session scope are subject to the parent scope chain delegation model but do not have a parent scope defined.

The following are the list of standard session variables:

session.startupmode

String that indicates the startup mode that the script was started as:

Value	Details
newcall	Session was started due to a new incoming call.
external	Session was started due to a external session launch request.
createccxml	Session was started due to a <createccxml> request.

session.id

A globally unique string that indicates the session identifier of the executing CCXML session.

session.uri

URI that was used when creating the current CCXML session. If the interpreter fetched multiple documents as a result of one or more HTTP redirects (e.g. 302), the value of this session variable is set to the URI of the final target. session.uri includes the query string used when fetching the document, if one is present.

session.parentid

String that indicates the session identifier of the parent of the CCXML session that created this session. In the case the current session has no parent, the value of the variable will be ECMAScript undefined. Once a new CCXML session is created, the new session and its parent are completely independent.

session.connections

Associative Array which contains a list of the Connection objects that the session is currently using. The array is associative and each key in the array is the connection identifier for the Connection. For example, session.connections["1234"] would return the Connection object for connection id ID 1234. The following example demonstrates the use of the session.connections variable:

<transition state="in_vxml_session" event="connection.disconnected">
  <if cond="session.connections['1234'].connectionid==event$.connid">
    <--  if the disconnect is on the first connection, do something -->
  <else/>
    <exit/>
  </if>
</transition>

session.conferences

An Associative array which contains a list of the Conference objects for which the session is attached using <createconference> For example, session.conferences["1234"] would return the Conference object for conference id 1234.

session.dialogs

Associative Array which contains a list of the Dialog objects that the session is currently using. The array is associative and each key in the array is the dialog identifier for the Dialog. For example, session.dialogs["1234"] would return the Dialog object for dialog id 1234.

session.ioprocessors

Associative Array which contains a list of external event I/O access URIs which are available to the current session. The array is associative and each key in the array is the type of the event I/O processor. For example, session.ioprocessors["basichttp"] would return the access URI (e.g. http://www.example.com/ccxml/basichttp ) for the "basichttp" type event I/O processor.

session.values

Associative array which contains a list of session parameters passed on session creation. For sessions created using <createccxml> , session.values contains variables passed using the ' parameters ' attribute of <createccxml> . Other session creation mechanisms, such as the HTTP-based session creation I/O processor defined in Appendix L, define other mechanisms for populating this variable.

8.4: Application Variables

CCXML provides application variables which, like session variables, persists across the CCXML application . Application variables differ from session variables in that they can be modified by CCXML programs.

The application object is initialized by the CCXML implementation. Application variables that are properties of the application object are not explicitly declared. Application variables that are not properties, e.g. objects, must be declared. By default, application variables have the value ECMAScript undefined. Variables in the application scope are subject to the parent scope chain delegation model and have session as their parent scope. For example:

        <assign name="application.userid" expr="'user001'"/>
        <var name="application.obj" expr="new Object()"/>
        <assign name="application.obj.prop1" expr="'value of prop1'"/>
        <assign name="application.obj.prop2" expr="'value of prop2'"/>

      <assign name="application.userid" expr="'user001'"/>
      <assign name="application.obj" expr="new Object()"/>
      <assign name="application.obj.prop1" expr="'value of prop1'"/>
      <assign name="application.obj.prop2" expr="'value of prop2'"/>

Application variables are visible within documents which form the CCXML application . For example, a document in a CCXML application could assign an application variable a value using <assign name="application.userid" expr="'user001'"/> , and a later document in the CCXML application could reference application.userid to retrieve the value 'user001'.

Application developers should be careful in their use of application variables since they are visible to all CCXML documents within a CCXML application.

9: Event Handling

This section contains information on <eventprocessor> , <send> , <cancel> , <transition> and <move> .

9.1: Overview

Event Handling is one of the most powerful features of CCXML. CCXML events can be delivered at any time and from a variety of sources. This flexible event-handling mechanism is essential for many telephony applications.

Every CCXML session can send and receive events. These might be in response to a previous action by the CCXML session (e.g., an outbound-call request, which generates an event to indicate when the call goes off-hook), or on the initiative of some external source (e.g., an incoming call to be answered). Events can be generated by the telephony system (as in the two previous examples), other CCXML sessions (which emit events via <send> ), Dialogs, or external sources. CCXML sessions can also send events to themselves.

There is a core set of telephony-related events (derived from the JCC event model for connection objects; See see the JAIN Call Control API (JCC) [JSR021] for more information) that a browser MUST support. Implementers are otherwise free to MAY define and support any platform-specific events they like. In addition, users/programmers may use <send> to send arbitrary events to internal or external destinations, or may send arbitrary events to CCXML documents from internal or external sources and may specify transition handlers to handle these events.

The transmission and reception of events both external and internal is controlled by a logical component in the platform called the "Event I/O Processor". A platform may MAY support more than one type of Event I/O Processor and each of them may support a different format of external events (For example: SOAP, JMS, SIP, Simple HTTP or any other event transmission approaches). For incoming events the Event I/O Processor is responsible for accepting the incoming event and transforming it into an ECMAScript event object that can be accessed by CCXML. For outgoing events the Event I/O Processor is responsible for deciding the serialization and transport formats. The operation and behavior of Event I/O Processors is platform dependent; however, this specification defines a Basic HTTP Event I/O Processor (Appendix K) that defines event transmission over HTTP. External Event I/O processors and new events created using <send> will typically generate application-specific events with different names than those of the standard events in this specification; however, it is legal for external sources and for events created using <send> to generate standard events. For instance, it is useful to be able to generate a ccxml.kill event to attempt graceful termination of a session from an external context, or from another CCXML session. Platforms SHOULD reject any standard events that do not contain all of the mandatory properties defined in this specification, and SHOULD notify the sender of the rejection (for instance with an error.send event).

Each running CCXML interpreter has a queue, into which it places incoming events. Events are generally queued on a first-in, first-out (FIFO) basis; however, certain classes of events are queued differently, as described in the paragraphs below. A CCXML programmer can only gain access to these queued events by using the <eventprocessor> element with associated <transition> elements.

An event can be delivered to a CCXML session using a <send> element in which case an optional delay may be specified. If a delay is specified, the event is not delivered to the target CCXML session until the delay time has elapsed. When the delay has elapsed, the event is queued as if it had just occured, occurred, using standard queueing rules.

There are three types of events that, when delivered to a CCXML session, are handled differently. All ccxml.kill and ccxml.kill.* events are placed at the head of the event queue rather than the tail so that they are processed in preference to all other events. All error.* events are placed at the head of the queue but behind any error.* , ccxml.kill , or ccxml.kill.* events that are already on the queue. ccxml.loaded events are always placed at the head of the queue.

An <eventprocessor> is interpreted by an implicit Event Handler Interpretation Algorithm (EHIA). The EHIA's main loop removes the first event from the CCXML session's event queue, and then selects from the set of <transition> s contained in the <eventprocessor> . A <transition> always indicates a set of accepted patterns to match event types, names, MAY indicate a required current value of the ECMAScript state variable specified in the <eventprocessor> , and MAY indicate a further ECMAScript conditional expression to be evaluated. The <transition> that accepts the type for the just-removed event, is in the specified state, has a satisfied conditional expression, and appears first in the <eventprocessor> in document order, is the selected <transition> .

Once selected, an object representing the event being processed is created at transition scope, and the elements inside the <transition> are executed in document order. If a semantic an ECMAScript evaluation error occurs that prevents an element in the transition from being executed (such as during the 'cond' attribute execution of an element within a <if> <transition> , error.semantic being an invalid ECMAScript expression), then MUST be raised for that element, and successive elements within that transition will <transition> MUST NOT be executed; an error.semantic will be raised for the element that could not be executed. Note that elements that can be executed but while an element that generate errors (such as a <disconnect> on references an invalid connection ID) do not connection, dialog, or conference identifier also causes the platform to raise error.semantic , these scenarios MUST NOT terminate execution of the transition. <transition> in which that element is contained.

At most one <transition> will be chosen. If no <transition> meets all the criteria, none are selected and the event is simply dropped; the EHIA loop then starts over again, removing the event at the head of the queue. The only exception to this rule is when ccxml.kill , ccxml.kill.* and error.* events are received; in this case, the CCXML interpreter will end the session if there are no <transition> s that match the ccxml.kill , ccxml.kill.* and error.* events explicitly.

Any events that arrive while an event is already being processed are just placed on the queue for later. If the event queue is empty, and the EHIA wants to process an event, execution pauses until an event arrives.

During the processing of an event by the EHIA, the state of any ECMAScript objects exposed by a platform, such as the Connection object, must reflect the state of the CCXML session immediately following the occurrence of the event. For instance, if a 'connection.alerting' ' connection.alerting ' event is being processed against a connection with ID 1234, then session.connections['1234'].state would have a value of 'ALERTING'. Connection.ALERTING . This is true even if the actual connection has already been terminated, with a 'connection.disconnected' ' connection.disconnected ' event queued (but not yet processed) against the session. It is required that the ECMAScript context for the session is updated prior to the selection of a matching <transition>, <transition> , since the <transition> might contain an ECMAScript conditional expression, the value of which depends on the state changes caused by the event.

Code Elements inside an <eventprocessor> <transition> SHOULD run "instantaneously", "quickly", without blocking execution. This will allow events to be rapidly processed. CCXML applications should be aware of this and should keep calculations such as large ECMAScript functions within a transition to a minimum.

The only way for CCXML execution to leave an <eventprocessor> is via an explicit <goto> or <exit> inside a <transition> or an unhandled error or ccxml.kill event .

An <eventprocessor> MAY also declare a state variable attribute. An <eventprocessor> 's state variable must be declared in the ccxml scope using a <var> or a <script> . The <eventprocessor> can be considered, and programmed as, a finite-state-automaton, with the state variable indicating the automaton's current state or node, and the <transition> s, driven by incoming events, moving the machine to a new state and creating side effects along the way. If a state variable is not defined in the <eventprocessor> 's statevariable attribute a default variable named " state$ " defined at the ccxml scope will be used instead. The initial value will be ECMAScript ' undefined '.

9.2: Elements

9.2.1: <eventprocessor>

9.2.1.1: Overview

The <eventprocessor> acts as a container for <transition> s. A valid CCXML document MUST only have a single <eventprocessor> .

9.2.1.2: <eventprocessor> Attribute Details

<eventprocessor> can MUST contain only <transition> s . elements.

9.2.2: <transition>

9.2.2.1: Overview

The content of a <transition> specifies the actions to be taken when it is selected. The <transition> are examined by the EHIA in document order.

In order to be selected, a <transition> MUST satisfy three criteria:

1. If a state attribute is specified on the <transition> , the parent <eventprocessor> MUST specify a statevariable attribute, and the current value of the associated state variable (as specified in the parent <eventprocessor> ) MUST be equal to one of the values specified in the state attribute of the <transition> . If no state attribute is specified on the <transition> , this criteria is met regardless of whether or not the parent <eventprocessor> specifies a value of the statevariable attribute .
2. the expression specified by the <transition> 's cond attribute MUST evaluate to true, true , if that attribute is present
3. the current event's name property MUST match the pattern specified by the <transition> 's event attribute, if that attribute is present

A <transition> with none of the attributes, state , cond , or event , will always be selected when encountered by the EHIA.

The contents of the received event object will MUST be available via the transition scoped read-only event$ ECMAScript variable. This variable is accessible from the <transition> cond attribute to allow CCXML Applications to conditionally select <transition> 's based on the contents of the event.

9.2.2.2: <transition> Attribute Details

9.2.2.3: <transition> Event Matching

The event attribute of a <transition> specifies a pattern used to match event names. Each character in the pattern specified by event must match the corresponding character in the event name, except for the asterisk ('*') character, which acts as a wildcard that matches any substring of zero or more characters in the event name. Patterns are dot-separated strings of arbitrary length. Event names are case-insensitive.

Pattern match examples

9.2.3: <send>

9.2.3.1: Overview

<send> is used to send messages containing events or other information directly to another CCXML Interpreter or other external systems using an Event I/O Processor.

The event target of <send/> <send> is specified using the target and targettype attributes. These attributes control how the platform should dispatch the event to its final destination.

The target attribute specifies the unique identifier of the event target that the Event I/O Processor should send the event to. This can be the value of a CCXML Session ID or a Dialog ID if you wish to send an event to one of these respective targets. In the case where you are using some other Event I/O Processor this attribute should be able to describe how to connect to the event destination (For (for example a SIP URL URI for SIP-INFO messages or a HTTP URL URI for Web Services). If the value of the target attribute is not supported, invalid or unreachable by the Event I/O Processor the Platform MUST throw a error.send.targetunavailable event.

The targettype attribute controls what Event I/O Processor the event should be sent to. The default value of this attribute is 'ccxml'. ' ccxml '. If the event targettype specified is not supported the platform MUST throw a error.send.targettypeinvalid event.

A platform must MUST support the following values for the targettype attribute:

Platforms may MAY support other types of Event I/O Processors, for example: Web-services, SIP or basic HTTP GET. However, platforms SHOULD name the Event I/O Processor beginning with "x-" " x- " to signify that they are platform dependent.

<send> also specifies the content of the message to be sent. <send> may MUST specify message content in one of two ways (the following mechanisms are mutually exclusive):

• name attribute with an OPTIONAL namelist
  • The name attribute specifies an ECMAScript expression that returns the name of the event.
  • The namelist attribute specifies a space separated list of CCXML ECMAScript variables to be included with the message.

  <var name="target" expr="'tel:+18005551212'"/>
  <var name="content" expr="'http://www.example.com/mycontent.txt'"/>
  <send target="target" 
           targettype="'x-messaging'" 
           name="'fax.SEND'" 
           namelist="content"/>
  

• Inline XML content that specifies the contents of the message being sent.

  <send target="'csta://csta-server.example.com/'" targettype="'x-csta'"
  xmlns:csta="http://www.ecma-international.org/standards/ecma-323/csta/ed3">
   <csta:MakeCall>
    <csta:callingDevice>22343</callingDevice>
  
    <csta:callingDevice>22343</csta:callingDevice>
  
   
  <csta:calledDirectoryNumber>18005551212</csta:calledDirectoryNumber>
   </csta:MakeCall>
  </send>
  

When inline XML is specified, the content of the <send> is parsed but MUST be ignored by the sending CCXML Interpreter until the <send> is executed. XML namespace binding within the <send> element MUST be preserved in the message sent to the <send> target. This may result in XML namespace declarations being added to the message contents in order to resolve XML namespace prefixes whose associated namespace declarations were not contained in the message contents in the original CCXML document. It is the responsibility of the Event I/O Processor responsible for forwarding events to the <send> target to parse the incoming message and remove the namespace bindings, if required by the <send> target.

The sending CCXML Interpreter MUST NOT alter the content of the <send> other than by adding namespace declarations for unresolved namespace bindings. The data contained within a <send> MUST be sent to the destination specified in the target attribute of <send> using the Event I/O Processor specified by the targettype attribute.

The mechanism by which the content of the message is made available to the target depends on both the targettype and the way in which the content is specified:

• When <send> is used with the name and namelist attributes, and the targettype is "ccxml", " ccxml ", the mapping from namelist variables to event object properties is well defined. Each variable listed in the namelist attribute is added as a property of the event object. Qualified variables retain their qualification and are mapped into nested properties of the event object. If a variable is an ECMAScript Object, then all properties of the variable are mapped into nested properties of the event object.
• When <send> is used with inline XML content, and the targettype is "ccxml", " ccxml ", the mapping of that XML content to event object properties is implementation-specific, and outside the scope of this specification.
• When <send> is used with the name and namelist attributes, and the targettype is "basichttp", " basichttp ", each variable in the namelist which is not an ECMAScript Object is converted into an HTTP parameter. The HTTP parameter name is the variable name, with the same qualification as used in the namelist. The HTTP parameter value is the variable value after being converted into a string. If a variable in the namelist is an ECMAScript Object, the mechanism by which it must be submitted is not currently defined. Instead of submitting ECMAScript Objects directly, the application developer may explicitly submit the properties of an Object (e.g. "date.month date.year"). Using <send> namelist with inline XML content, and a targettype of "basichttp" is not permitted sent to the web server as noted defined in Appendix K. K
• When <send> is used with either name and namelist attributes or inline XML content, and the targettype is "dialog" " dialog " or a platform extension, the mechanism by which namelist variables or XML content are made available to the event target is implementation-specific, and outside the scope of this specification.

Although the full set of requirements for the Event I/O Processor is not within the scope of this specification, an event processor sending an event to a CCXML Interpreter is required to generate an event which can be processed in a CCXML Session. See Section 9.1 for details regarding the processing of incoming events by an CCXML Interpreter.

When a message is successfully sent to the target, a send.successful event will MUST be thrown. Note that this does not mean that the target processed the message successfully. It is up to the target to generate events specific to the message. These events are application specific.

If the send request fails, an event signifying the error will MUST be returned to the CCXML Session. The failure events are documented at the end of this section.

Events generated by <send> will have their eventid , eventsource and eventsourcetype attributes derived from the execution of the <send> element. eventid will match the value returned via sendid . eventsource will be the ccxml session id. eventsourcetype will be set to the string ' ccxml '.

9.2.3.2: <send> Attribute Details

9.2.3.3: <send> Examples

In this example we send the current CCXML session a hello.jack event that contains a single field. We catch the event, log the field and exit:

<?xml version="1.0" encoding="UTF-8"?>

  <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <eventprocessor>  
    <transition event="ccxml.loaded">
      <var name="jacksvar" 
              expr="'I am Jack\'s complete lack of surprise.'"/>
      <send target="session.id" targettype="'ccxml'"
       name="'hello.jack'" namelist="jacksvar"/>
    </transition>
    <transition event="hello.jack">
      <log expr="event$.jacksvar"/>
      <exit/>
    </transition>    
  </eventprocessor>
</ccxml>
  

In this example we send a event to our parent and then exit:

<?xml version="1.0" encoding="UTF-8"?>

  <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <eventprocessor>  
    <transition event="ccxml.loaded">
      <var name="jacksvar" 
              expr="'I am Jack\'s inflamed sense of rejection.'"/>
      <send target="session.parentid" name="'hello.jack'"
       targettype="'ccxml'" namelist="jacksvar"/>
      <exit/>
    </transition>
  </eventprocessor>
</ccxml>
  

In this example we catch a dialog.transfer request and just return a error event back to the dialog:

<?xml version="1.0" encoding="UTF-8"?>

  <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <eventprocessor>  
    <transition event="dialog.transfer">
      <var name="reason" expr="'I am a jack's unsupported transfer.'"/>
      <send target="event$.dialogid" name="error.unsupported.transfer"
       targettype="'dialog'" namelist="reason"/>
      <exit/>
    </transition>
  </eventprocessor>
</ccxml>
  

In this example we send the current CCXML session an event that contains an ECMAScript Object and a qualified variable. We catch the event, log some properties resulting in the string "1, " 1, 2, 3, 4", 4 ", and exit:

    <?xml version="1.0" encoding="UTF-8"?>

<ccxml version="1.0" xmlns="http://www.w3.org/2002/09/ccxml">
  <eventprocessor>  
    <transition event="ccxml.loaded">
      <var name="obj1">
      <assign name="obj1.prop1" expr="'1'">
      <assign name="obj1.prop2.a" expr="'2'">
      <assign name="obj1.prop2.b" expr="'3'">
      <var name="obj2">
      <assign name="obj2.prop1" expr="'4'">

      <var name="obj1"/>
      <assign name="obj1.prop1" expr="'1'"/>
      <assign name="obj1.prop2.a" expr="'2'"/>
      <assign name="obj1.prop2.b" expr="'3'"/>
      <var name="obj2"/>
      <assign name="obj2.prop1" expr="'4'"/>

      <send target="session.id" targettype="'ccxml'"
       name="'my.complex.event" namelist="obj1 obj2.prop1"/>

       name="'my.complex.event'" namelist="obj1 obj2.prop1"/>

    </transition>
    <transition event="my.complex.event">
      <log expr="event$.obj1.prop1 + ', ' + event$.obj1.prop2.a + ', ' +
                 event$.obj1.prop2.b + ', ' + event$.obj2.prop1"/>
      <exit/>
    </transition>    
  </eventprocessor>
</ccxml>
    

9.2.4: <move>

9.2.4.1: Overview

<move> is used to move an event source (such as a Connection object) to an executing CCXML session. When an event source is moved to a session, events originating from that source will MUST be delivered to that session's currently executing CCXML document. The event OR the source attribute MUST be specified. If neither attribute is specified or both attributes are specified, an error.fetch event will MUST be thrown.

Support for the <move> element is optional in CCXML platforms. Even if <move> is supported platforms may MAY only support it between a restricted set of CCXML sessions. If <move> is unsupported the CCXML platform will MUST raise an error.move event.

A <move> attempt may MUST fail for a number under any of reasons. These reasons include: the following conditions:

• The referenced event source does not exist, or does not belong to the session executing the <move> ;
• The referenced event source is immovable - the CCXML session itself and various external event I/O processors are examples of this;
• The referenced event source is currently bridged in a way that prevents it from being moved (as described below) ; to another object through an implicit or explicit join;
• The session to which the event source is being moved does not exist, or cannot accept the event source that is being moved;
• Other causes, which may be platform specific. The event attribute is used to identify the event source, but refers to multiple event sources (such as connection.merge.failed );

A <move> is not allowed once an event source is associated with another object through an implicit or explicit <join>. Moving event sources that are joined would disassociate the objects connected through the join. For example the result of moving connection that has an active dialog would be two CCXML sessions, one owning the connection, the attempt MAY fail for causes other owning than the dialog. above, which may or may not be platform specific

In the event of a <move> failure, an error.move event will MUST be generated to indicate that the <move> attempt was unsuccessful. Otherwise, a move.successful event will MUST be generated against the session that performed the <move> to indicate that the request has completed successfully.

Since each CCXML session has an event queue, it is possible that a session executing a <move> will already have events in its queue from the event source that is being moved. As an example, this situation could easily occur with connection.alerting and connection.disconnected if the incoming connection is abandoned before a CCXML platform initiates processing. Any such events will MUST be removed from the queue of the session performing the <move> , and placed into the queue of the session that the event source is being moved to. Each event moved in this manner will MUST be inserted into the queue of the target session as if it was a new event occuring occurring in the context of that session. The order in which events are queued is the order in which they appear in the event queue of the session performing the <move> . If the event attribute is specified, then the referenced event is MUST be placed into the queue of the target session before other queued events from the event source are inserted. Note that queueing rules governing the order in which the inserted events will be processed continue to apply.

Like all CCXML elements, <move> executes asynchronously. As such, there is a period of time during which events generated by the event source being moved cannot be processed either by the session performing the <move> , or by the target session. If the <move> fails, then these events remain in the queue of the session that attempted the <move> , and will MUST be processed normally. However, while the <move> is being performed, events from event sources other than the one being moved will MUST continue to be processed according to the EHIA. As such, a failed <move> request may result in events being processed in a different order than if no <move> operation was performed. Note that the relative ordering of events from the event source being moved is not changed even as a result of such a failure.

9.2.4.2: <move> Attribute Details

9.2.5: <cancel>

9.2.5.1: Overview

When a CCXML program uses <send> to send an event and includes a delay attribute, the <cancel> command will cancel the pending event, if possible.

The cancel operation cancels a pending event by removing it from the delayed event queue of the CCXML session that initiated the <send> , preventing it from ever being delivered. If the delay has expired and the event has already been removed from the delayed event queue, the <cancel> request will MUST fail and an error.notallowed event will MUST be delivered to the event queue of the CCXML session that executed the <cancel> . Otherwise, a cancel.successful event will MUST be delivered to indicate that the <cancel> was successful, and that the cancelled event was not delivered. Sessions are only permitted to cancel their own delayed events; they may not cancel the delayed events of other sessions.

9.2.5.2: <cancel> Attribute Details

9.3: Events

9.3.1: error.notallowed

This error event is MUST be thrown when the execution of an element causes an invalid operation to be performed on a session and/or connection. The fields available in this event are:

9.3.2: error.semantic

This error event is MUST be thrown when there is a semantic error in a CCXML element (ie. ( e.g. passing an incorrect value for an attribute, etc.).

The fields of this event are:

9.3.3: error.send.targetunavailable

This error event MUST be thrown when CCXML could not send the event to the target listed in <send> due to it currently being unavailable. The fields available in this event are:

9.3.4: error.send.targettypeinvalid

Could not connect to This error event MUST be thrown when the Event I/O Processor listed targettype attribute specified in a <send> . is not valid.

The fields available in this event are:

9.3.5: error.send.failed

Message in the This error event MUST be thrown when a <send> could not be sent completed for an unknown a reason . not covered by another error.send.* event.

The fields available in this event are:

9.3.6: send.successful

This event is MUST be thrown when an event is successfully delivered to the specified receiver. Receipt of the event does not imply the event has been processed by the receiver but simply that it has been sent without error. The fields available in this event are:

9.3.7: move.successful

This event is MUST be thrown when an event source is successfully moved to a CCXML session. The fields available in this event are:

9.3.8: cancel.successful

This event is MUST be thrown when the sending of an event has been successfully cancelled. The fields available in this event are:

9.3.9: error.move

This event is MUST be thrown when a move request performed by a session fails to complete successfully. The fields available in this event are:

9.4: Standard Events

9.4.1: Overview

CCXML can generate arbitrarily-named events. While any event name is possible, there is a small set of well-known events that are generated as a matter of course, and which any telephone application SHOULD should handle. There are three kinds of these events: connection events, language events and error events.

The first, and larger set, is present so a CCXML session can keep abreast of events happening with the telephone network. CCXML is designed to be neutral with respect to the telephony layer, so the event set MUST must be very generic and capable of describing the behavior of a wide variety of systems (e.g., [Q931] , [SS7] , VoIP, etc).

9.4.2: Standard Event Attributes

9.4.2.1: Overview

All events received in a CCXML session must have a number of standard fields. It is the responsibility of the Event I/O Processor that delivered the event to make sure that they are present.

9.4.2.2: Standard Event Attribute Table

9.4.3: Connection Events

9.4.3.1: Overview

CCXML applications are notified of Connection activities by events, which often reflect Connection state changes. Applications MAY may also take actions which change the state of a Connection and which cause other events to be generated.

Connection events and their properties are specified in sections section 10.6: Events , and 10.2.3: Connection Events

9.4.4: Language Events

Language Events are a general class of responses that occur as a result of the execution of elements within a CCXML document. These events may be further categorized as follows:

• Document Control: These events are detailed in section 6.3
• Dialog Control: These events are detailed in section 7.3

• Event Handling Control: These events are detailed in section 9.3

• Call Control: These events are detailed in section 10

9.4.5: Error Events

9.4.5.1: Overview

CCXML uses its event handling mechanism to notify a CCXML document of errors that occur during execution. An error notification takes the form of an error event that is added to the event queue of the CCXML document where the error occurred. All error events are named with the prefix "error." so that a properly defined <transition> can filter out error events.

Here is an example of a <transition> that can be used to filter out and report error events:

<transition event="error.*">
  <log expr="'an error has occurred (' + event$.reason+ ')'" />
  <exit/>
</transition>

All error events have a set of properties in common, shown in the following table:

9.5: Error Handling

A CCXML interpreter MAY provide the attributelist property on the error event. The attributelist object MAY have just the attribute and value in error, or MAY provide all the attributes and values. Due to the nature of CCXML's event handling mechanism, some error scenarios are treated differently. These scenarios are described below.

9.5.1: Fetching & compilation errors

Errors that occur in trying to load and compile a CCXML document, such as the inability to fetch the CCXML page or statically referenced ECMAScript content, XML parsing or validation errors, errors (Including checks for mutually exclusive attributes or any other constraints as defined in the this specification, it's attribute tables, schema or DTD), compilation errors or other errors that occur as a result of trying to fetch and run a CCXML page prior to document initialization. The handling of fetching errors is dependent on the context of the fetch; if the fetch was triggered by a running CCXML document using <createccxml> or <fetch> , then an error.createccxml or error.fetch (respectively) MUST be thrown to the initiating session. If an error occurs fetching the initial CCXML document for a session created to handle an incoming call, the CCXML session created for the incoming call MUST be terminated; the incoming call SHOULD be rejected. If the session is created through other means (such as the Session Creation Event I/O Processor), the session MUST be terminated; the creator of the session SHOULD be notified.

9.5.2: Document Initialization Errors

Errors that occur during documentation initialization (elements that occur in the CCXML document before <eventprocessor> ) occur outside of CCXML's event handling mechanism. These errors MUST cause the CCXML thread of execution to terminate and notify the platform of the document error.

9.5.3: Error In <eventprocessor> attributes

An <eventprocessor> contains <transition>s that comprise CCXML's event handling mechanism. Since errors in <eventprocessor> attribute evaluation could keep the EHIA from correctly processing an event, these errors MUST cause the CCXML thread of execution to terminate and notify the platform of the document error.

9.5.4: Error in <transition> attributes

<transition> attributes specify when the elements contained by the <transition> should be executed. Since errors in <transition> attribute evaluation could keep the <transition> from correctly handling the error event, these errors MUST cause the CCXML thread of execution to terminate and notify the platform of the document error.

9.5.5: Errors While Handling Error Events

If an error occurs during the handling of an error event another error event will be raised and posted to the front of the event queue. In many cases this may be perfectly acceptable and the CCXML application may be able to successfully recover from the error. In some situations however this can lead to an infinite loop in the CCXML session. Implementations MAY choose to include some form of loop detection and to terminate the CCXML session when a loop is detected.

9.5.6: Errors in hints attributes

Several elements within this specification define a hints attribute. Platforms that do not support hints on an element MAY ignore this attribute. If a platform supports the hints attribute on an element, the attribute is specified, and ECMAScript evaluation of the value of the hint fails or the ECMAScript expression does not resolve to an ECMAScript object, error.semantic MUST be raised for that element and successive elements within the <transition> containing that element MUST NOT be executed. Additional processing of hints by the implementing platform is platform-specific.

10: Telephony Operations and Resources

This section contains information on <accept> , <createcall> , <createconference> , <destroyconference> , <disconnect> , <join> , <merge> , <redirect> , <reject> , and <unjoin> .

10.1: Overview

The primary goal of CCXML is to provide call control throughout the duration of a call. Call control includes handling incoming calls, placing outgoing calls, bridging (or conferencing) multiple call legs, and ultimately disconnecting calls.

In CCXML call control occurs through three major concepts: Connections, Conferences and Bridges.

A Connection is an object modeling a resource by which two independent unidirectional media streams, and optionally any associated network signaling traffic, can be controlled by a CCXML session. This corresponds roughly to a "call leg" as the term is used informally. The picture below illustrates the media streams associated with the Connection c1 .

A Bridge occurs when the input and/or output media streams of Connections or Conferences are linked or "joined" together. The picture below depicts the result of a full duplex <join> between the connections c1 and c2 .

A Conference is an object that controls the mixing of media streams for two or more Connections through Bridges.In the picture below, the connections c1 and c2 are joined in a full duplex mode to the conference C1 .

These concepts are discussed in greater detail in the sections below.

10.1.1 Concepts Background (INFORMATIVE)

The goals of the CCXML call model are to focus on relatively simple types of call control and to be sufficiently abstract so that the call model can be implemented using all major telephony definitions such as JAIN Call Control(JCC) [JSR021] , [CSTA] , and [S.100] . The JCC call model meets these requirements by providing an event model for connections which abstracts away many of the differences between telephone networks (e.g., [Q931] , [SS7] , VoIP, etc). Additionally, this call model is small and easily-understood so that concrete example programs can be written.

JCC was designed to be a cross-platform high-level event set to describe as generic a phone model as possible. The JCC call model consists of Addresses, Calls, Connections, and Providers. In the context of CCXML, it was felt that the Address, Call, and Provider objects would add more complexity than value, so these were omitted as explicitly visible objects. Instead the behavior of Connections became the focus.

The CCXML call model therefore is based on the behavior of Connections. A call is received or initiated under control of a CCXML session through the properties of a Connection.

Note that the JCC model is designed for endpoint devices only.

10.2 Connections

The CCXML call model is based on the behavior of Connections.

Each Connection has one input by which it receives a media stream from another Connection or Conference.

Each Connection has one output media stream that can be directed to the inputs of multiple Connections and/or Conferences.

If a network call is active on a Connection, Connection (in the CONNECTED , PROGRESSING or ALERTING state), the media stream received from the network is the Connection output, and the Connection input media stream is transmitted to the network.

Dialogs are also modeled internally as a Connection for the purposes of media interaction. The Dialog object behaves as if it was a Connection object and can be joined to other resources in the same way as any other Connection. A Dialog will not however be listed in the session.connections session variable.

<dialogstart connectionid="c1" src="'example.vxml'" duplex="'full'">

For a Connection created by <dialogstart> , the Connection input media stream is available to a recognizer under control of the CCXML session, and the Connection output media stream can be sourced from a resource (such as a Text To Speech engine) under control of streams are available to the CCXML session. Dialog resource.

10.2.1: Connection State

The state of a Connection Object reflects events that are generated, either a result of actions that occur within the telephony network, or as a result of actions performed by the CCXML application. The following state diagram shows the major aspects of Connection behavior, but omits some detail in favor of clarity (e.g. The the ERROR state).

The list of valid states that a connection can be in is:

• ALERTING
• PROGRESSING
• CONNECTED
• FAILED
• DISCONNECTED
• ERROR
• CREATED

Connection Objects are created when a incoming call arrives to the platform via a connection.alerting event or via the execution of <createcall> .

Connection Objects are automatically destroyed by the platform after they have reached the DISCONNECTED , FAILED or ERROR states. Platforms are responsible for deciding when to remove unused Connection Objects, however platforms MUST maintain a Connection Object at least through the <transition> for the corresponding connection.disconnected , connection.failed , connection.redirected , connection.merged or error.connection event

The PROGRESSING and ALERTING states have reflexive transitions. This is intended to model protocols which have additional states at these points, and which MAY may exchange messages such as PROCEEDING , ALERTING , FACILITY , or NOTIFY . Platforms MAY choose to implement additional states which MAY may be reflected in the substate property of the Connection object. Additional messages can be implemented with <send> .

10.2.2: Connection Object

An instance of the Connection Object is associated with each telephony event source. Each instance is uniquely identified by its connection identifier. All Connection instances have a set of properties in common, shown in the following table. Properties that are not indicated as required only appear on an instance of the Connection object if they have a value. Other properties will always be present.

The Connection Object is an extension of the Media Endpoint Object as defined in section 10.4.3.

Platforms MAY choose to add properties to Connection instances. By convention, the properties MUST begin with an underscore, "_", to identify them as platform-dependent.

10.2.3: Connection Events

CCXML applications are notified of Connection activities by events, which often reflect Connection state changes. Applications may also take actions which change the state of a Connection and which cause other events to be generated.

In addition to the standard event attributes detailed in Section 9.4.2 , all Connection events have a set of properties in common, shown in the following table. Fields that are not indicated as required only appear on the event object if they have a value. Other fields will always be present.

Platforms MAY choose to add properties to events. By convention, the properties MUST begin with an underscore, "_", to identify them as platform-dependent.

10.2.4: Connection Operations

CCXML defines a number of elements, specified in detail in section 10.5, that are used to interact with Connection Objects. Execution of these elements does not MUST NOT immediately change the state of the Connection Object; Connection Objects change their state once the event corresponding to the event transition is handed. Elements, such as <accept> or <createcall> <disconnect> , do not actually change the state of the connection but instead issue requests to the platform to process the corresponding command that will in turn raise the events that chance the connection state. The sole exception to this is <createcall> , which creates a Connection Object immediately, with its ' state ' property set to CREATED , allowing the Connection Object to be accessed following the <createcall> , within the same transition. However, existence of the Connection Object does not guarantee anything about the success of the <createcall> request.

When elements are used to request an action against a Connection Object, a corresponding event is always generated to indicate the result of executing that request. For instance, when the <accept> element is used to answer an incoming call, a ‘connection.connected’ ' connection.connected ' event indicates that the call was accepted successfully, whereas a ‘connection.failed’ ' connection.failed ' indicates that the call could not be accepted. The events that can be generated by elements that act on connections, and the circumstances under which those events occur, are summarized in the table below. Note that the <join> and <unjoin> elements are not included, since these apply to both Connections and Conferences, and are discussed in detail in Section 10.4 on Bridges.

There are two common events that can be generated by any of the connection-affecting elements:

• If the element cannot be evaluated, for example if the referenced connectionid contained an invalid ECMAScript expression, then an ‘error.semantic’ ' error.semantic ' event is thrown – - as is the case for all CCXML elements. The error.semantic event MUST NOT change the state of the associated Connection Object(s).
• If the element cannot be executed because the current state of the referenced Connection Object (or Objects, in the case of <merge>) <merge> ) is not valid for the requested action, then an ‘error.connection.wrongstate’ ' error.connection.wrongstate ' event is thrown. Note that an ‘error.connection.wrongstate’ error.connection.wrongstate event is thrown based on the state of the Connection Object, which in turn reflects the events that have been processed by the CCXML application; it is not based on the actual real-time state of the underlying connection, which could change at any time.

It is important to note that because CCXML is fully asynchronous, and because events can occur on the underlying communications network at any time, that connection-related events can occur that are not related to application actions. The most common of these is connection.disconnected , which can be triggered by a user-initiated hangup, and which can occur in any state during which the cal call is active.

Finally, once an event is emitted by a Connection Object that transitions to a final state (such as DISCONNECTED), no DISCONNECTED ), further events will MUST NOT be generated by that Connection Object. Thus, even if an application has performed an <accept> and expects to receive one of ‘connection.connected’, ‘connection.accept.failed’, connection.connected , ' connection.accept.failed ', or ‘connection.failed’, 'connection.failed' , it could receive a ‘connection.disconnected’ 'connection.disconnected' instead. Since this event is associated with a transition to the DISCONNECTED state, no further events will be generated; thus the application will not receive any event as a result of having performed an <accept> .

The following table shows the events that reflect the outcome of the connection-oriented elements defined by CCXML, for this reason the table omits the event generated by the underling network (i.e. connection.alerting and connection.signal): connection.signal ):

10.2.5: Connection Class

All connection objects MUST be initiated via the Connection class. The Connection class has several MUST be read-only properties. and provide the properties defined below.

The following table lists the constants that correspond to the possible values of the state property of the connection object.

The following table lists the values of each index in the read-only array of Connection states. This array MUST be read-only. Each connection state constant can be used to index into this array to return the human-readable string corresponding to that constant.

The following example logs the current state of the connection object:

<transition event="connection.*">
   <log expr="Connection.states[event$.connection.state]"/>

</transition>

10.3: Conferences

CCXML applications can use <createconference> to create a new conference, or to attach to an existing conference. The CCXML application can connect or disconnect existing connections/conferences/dialogs to the new conference using <join> and <unjoin> (as described in Section 10.4 ). When a session no longer requires the use of a conference, it uses <destroyconference> to detach from the conference. Asynchronous events will be sent to the CCXML document upon completion of each of these operations.

Each Conference Object has one logical output and multiple inputs. The actual output streams of a Conference Object are derived by mixing all its input streams, less any contributed audio of an individual Connection, Conference Object who receives that output. The output of a Conference Object can be directed to the inputs of multiple Connections and/or Conference Object (as a result of bridging).

Some telephony call control definitions do not define a separate Conference Object, instead defining a conference simply as a call with more than two parties. In order to accommodate the widest range of underlying telephony API's, CCXML requires explicit use of a Conference Object whenever two or more audio streams are mixed.

Unlike connections and dialogs, which are local to a single session (but can be moved between sessions using <move> ), conferences are global across all sessions, and can be bridged with the connections/conferences/dialogs of any session. Conferences can be named in order to facilitate the use of that conference in other sessions. Assigning a name using the confname attribute allows other sessions to access the created conference by performing a <createconference> with the same value for confname . Conferences continue to exist so long as at least one session exists that has attached to the conference using <createconference> without a corresponding <destroyconference> . When a session terminates, it implicitly detaches from any conferences to which it is still attached. Note that it is not necessary for a session to share a conference by assigning it a name in order for other sessions to make use of that conference. Other sessions may still establish bridges to a conference using <join> and <unjoin> , regardless of whether or not they have accessed the conference using <createconference> ; only the conference ID is required for this. Because conferences are global, it is not legal to perform a <move> against a conference.

NOTE: A simple two-party call does not require the use of a conference object. This is discussed in Section 10.4 .

10.3.1: Conference Object Properties

An instance of the Conference class is associated with each Conference Object created by <createconference> and referenced in the session.conferences associative array.

10.3.2: Conference class

All conference objects MUST be initiated via the Conference class. The Conference class currently has no defined properties.

10.4: Bridges

A "bridge" is a relationship between the input and/or output streams of Connections or Connection, Dialog, and Conference Objects. The bridge concept and the details of its behavior are fundamental to CCXML. Bridges may also involve Dialogs, which CCXML Dialogs and Connections are equivalent to Connections with respect to establishing and terminating bridges. bridges; the term "Media Endpoint" is used as a common term to refer both to Connections and Dialogs are therefore not distinguished below. from the perspective of bridging operations.

There are two main ways in which CCXML applications can manipulate bridges:

• Implicitly, using the <createcall> (with 'joinid') ' joinid ') and <dialogprepare>/<dialogstart> (with 'connectionid'/'conferenceid') ' connectionid '/' conferenceid ') elements;
• Explicitly, via the <join> and <unjoin> elements.

Even in the simplest case of a network party interacting with a dialog, two Connections Media Endpoints are REQUIRED , and a bridge is established between them implicitly by the action of <dialogstart> . More complex situations, such as two-party calls, two-party calls with "hot" word "hotword" recognition, conference control, and "coaching" scenarios, all involve the use of multiple Connections Media Endpoint and explicit control of one or more bridges between them by using <join> and <unjoin> .

The nature of bridges, and the behavior of <join> and <unjoin> , is concerned with the mapping between the media stream inputs and outputs of Connections Media Endpoints and Conferences:

• Each Connection Media Endpoint has one input, and 0 or more outputs. If a network call is active on a Connection, Media Endpoint, the media stream received from the network is the Connection Media Endpoint input, and the Connection Media Endpoint output media stream is transmitted to the network. For a Connection Media Endpoint created by <dialogstart> , the Connection Media Endpoint input media stream is available to a recognizer under control of the CCXML session, and the Connection Media Endpoint output media stream can be sourced from a resource (such as a Text To Speech engine) under control of the CCXML session.
• The output of a Connection Media Endpoint or Conference can be directed to the inputs of one or more Connections Media Endpoints and/or Conferences. This is sometimes called "splitting" media streams. Not all environments/implementations will necessarily support "splitting" a media stream in this manner. In such environments, a request to establish a bridge that requires the presence of splitting may fail.
• A Connection or Conference Media Endpoint input can come from the output of only one Connection Media Endpoint or Conference.
• A Conference input can have multiple inputs from the outputs of Media Endpoints or Conferences.
• If a new bridge requires a Connection Media Endpoint to "listen", and the Connection Media Endpoint is already listening to a different source, this existing stream relationship is torn down automatically.

Bridges can be either one-way, in which the media stream flows only from party A to party B (such that B can hear A, but A cannot hear B), or two-way, in which the media stream flows in both directions between the parties invovled. involved. <join> has a duplex attribute to distinguish between two-way bridges and one-way bridges. For example, <join> ing Connection Media Endpoint A to Connection Media Endpoint B with duplex=full will direct the A output to the B input, and the B output to the A input, creating a simple two-party call. If instead the same <join> is done with duplex=half , it will direct the B output to the A input, and will not have any effect on the B input. Similarly, <dialogprepare>/<dialogstart> <dialogprepare> / <dialogstart> have the 'mediadirection' ' mediadirection ' attribute and <createcall> has the 'joindirection' ' joindirection ' attribute for controlling what kind of bridge is established.

For "hot word" "hotword" recognition on a two-party call, a two-way (full duplex) bridge MUST must be established between two network Connections, and a one-way (half duplex) bridge MUST must be established from one of the network Connections to a third Connection that is associated with a dialog. Dialog used to perform the recognition. There are several ways this arrangement can be achieved, depending on the initial states of the three Connections. Connections and Dialogs. For example, if the network party on Connection A is initially interacting with a dialog on Connection Dialog D (i.e., a full duplex bridge exists between them), all that is needed then is to do a <join> of Connection A to Connection B (the other network party) with duplex=full . This example highlights an important and subtle aspect of the behavior of <join> when one, or both, of the Connections being joined is already in one or more established bridges: bridges.

Note that <join> MUST NOT be used to add a Connection Media Endpoint to an existing two-party bridge in order to create a 3-way Conference Instead, this functionality can be achieved by first using <createconference> to create a Conference object and then <join> ing all three Connections Media Endpoints to this Conference. If a two-way bridge exists between A and B, and A is then <join> ed full duplex to C, the result will be a two-party bridge between A and C and a one-way bridge from A to B.

CCXML applications MUST only establish bridges on Connections (and Dialogs) Media Endpoints that are owned by the session running that application. As such, a session cannot bridge a Connection Media Endpoint that it owns to a Connection Media Endpoint that is owned by another session, nor can it establish a bridge between two Connections Media Endpoints owned by another session. Since Conferences are global, however, a session MAY bridge a Connection Media Endpoint that it owns to any Conference

Asynchronous events are used to notify the CCXML application upon the completion of bridging operations performed by that application. Such events are also used to notify the application of changes to existing bridges that were not requested by the application - such as a <destroyconference> terminating a conference that a local connection has a bridge to. However, the execution of bridging operations and completion notification varies depending on how the bridging operation was performed:

• Implicit bridges created using <dialogprepare>/<dialogstart> (by specifing 'connectionid' ' connectionid ' or 'conferenceid') ' conferenceid ') are established when the dialog is started. No bridging events are generated; the 'dialog.started' ' dialog.started ' event indicates that the dialog was started and the bridge is in place. If the specified bridge cannot be established, then the dialog MUST not be started, and an error.dialog.notstarted event is generated generated. Note that a session could still receive a conference.joined event when a dialog is implicitly joined to a conference; this event is generated by the conference.
• Implicit bridges created using <createcall> (with 'joinid') ' joinid ') are physically realized as soon as a platform is capable of doing so; however, the inability to establish the bridge immediately upon execution of a <createcall> does not constitute an error. Prior to the point in time at which the call enters the CONNECTED state, it may not be possible to establish the requested bridge, or it may only be possible to establish the bridge one direction. As a result, when using <createcall> with ' joinid ', an event MUST be generated when the requested bridge operation is completed completed. This event is independent of connection-related events such as 'connection.connected', ' connection.connected ', and may occur before or after such events depending on when the bridge is actually realized. If the bridge is established successfully, the 'conference.joined' ' conference.joined ' event is generated. If the requested bridge cannot be established even after the connection reaches the CONNECTED state, or if some other issue prevents the bridge from being established, an 'error.conference.join' ' error.conference.join ' event will be generated. Referencing a dialog that has been prepared but not started in 'joinid' ' joinid ' always result in an error, and thus an 'error.conference.join' ' error.conference.join ' event. Note that the call itself MUST proceed independently of whether or not the bridge can be established. for For instance, it is possible to receive both an 'error.conference.join' ' error.conference.join ' event and a 'connection.connected' ' connection.connected ' event for a single call initiated using <createcall> .
• Explicit bridges creation requests using <join> are performed immediately independently, of the state of the underlying resource. Requests to terminate existing bridges using <unjoin> are also performed immediately. CCXML applications can use <join> and <unjoin> at any time, except in the case of dialogs, where <join> and <unjoin> can only be used on dialogs that have been started. For instance, a CCXML application can join two connections where one (or even both) of the connections is in the ALERTING state, providing early media on the ALERTING connection. If the CCXML implementation does not support a requested <join> due to the current state of an underlying entity, then this will result in an error.conference.join event with the 'reason' ' reason ' field set appropriately.

10.4.1: <join> Outcomes

As an aid to understanding, the outcomes of all possible <join> operations are shown diagrammatically below for three different initial conditions:

1. a single Connection A, that is not currently in a bridge, represented graphically as: (A)
2. two Connections A and B, with a half duplex bridge (A output to B input), represented graphically as: A -----> B
3. two Connections A and B, with a full duplex bridge (A output to B input & B output to A input), represented graphically as: A <====> B

In summary, <join> behavior always respects three invariants:

1. The media stream relationship specified in the <join> is established between the two Connections/Conferences referenced in the <join> . In particular, any existing stream relationship between these two Connections/Conferences is torn down automatically if it conflicts with the specified relationship.
2. If the relationship specified in the <join> requires a Connection to "listen" and the Connection is already listening to a different source, this existing stream relationship is torn down automatically.
3. Any existing stream relationship that does not present a conflict according to invariant #1 or invariant #2 is preserved.

10.4.2: Invariant Examples

To illustrate some typical invocations of <join> invariants a few example scenarios are presented below. In the first scenario, connection c1 is bridged to a conference C1 , via a <join> where the duplex mode is full. full .

After <join id1="'C1'" id2="'c1'" duplex="'full'">

Session object properties
session.connections['c1'].input == 'C1'
session.connections['c1'].outputs.length == 1
session.connections['c1'].outputs[0] == 'C1'
session.conferences['C1'].bridges['c1'] == 'c1'

If c1 then became a participant in a <dialogstart> where d1 represents a connection to a dialog and the mediadirection is both, the original picture would change as follows:

After <dialogstart connectionid="'c1'" src="'example.vxml'" mediadirection="'both'">

session.connections['c1'].input == 'd1'
session.connections['c1'].outputs.length == 2
session.connections['c1'].outputs[0] == 'C1'
session.connections['c1'].outputs[1] == 'd1'
session.connections['c1'].dialogid == 'd1'
session.conferences['C1'].bridges['c1'] == 'c1'
session.dialogs['d1'].input == 'c1'
session.dialogs['d1'].outputs[0] == 'c1'
session.dialogs['d1'].outputs.length == 1

The <dialogstart> required c1 to "listen" to d1 , however, c1 was already in an established bridge listening to C1 . Consequently, the full duplex bridge between c1 and C1 is changed to a half duplex, where c1 is not allowed to "listen" to C1 and a full duplex bridge is established between c1 and d1 .

In this second scenario, c1 and c2 have been joined into a conference, C1 .

After <join id1="'C1'" id2="'c1'" duplex="'full'"> 
and:  <join id1="'C1'" id2="'c2'" duplex="'full'">

session.connections['c1'].input == 'C1'
session.connections['c1'].outputs.length == 1
session.connections['c1'].outputs[0] == 'C1'
session.connections['c2'].input == 'C1'
session.connections['c2'].outputs.length == 1
session.connections['c2'].outputs[0] == 'C1'
session.conferences['C1'].bridges['c1'] == 'c1'
session.conferences['C1'].bridges['c2'] == 'c2'

If a <join> is then executed that specifies c2 and C1 as participants and the duplex mode is half, the bridge between c2 and C1 will be re-established with C1 able to "listen" to c2 , but c2 no longer able to "listen" to C1 .

After <join id1="C1" id2="c2" duplex="'half'">

session.connections['c1'].input == 'C1'
session.connections['c1'].outputs.length == 1
session.connections['c1'].outputs[0] == 'C1'
session.connections['c2'].input == undefned
session.connections['c2'].outputs.length == 1
session.connections['c2'].outputs[0] == 'C1'
session.conferences['C1'].bridges['c1'] == 'c1'
session.conferences['C1'].bridges['c2'] == 'c2'

In this third scenario, there are three connections c1, c2 and c3. Connections c1 and c2 are connected with a half duplex bridge, connection 3 is not bridged at this point.

After <join id1="'c1'" id2="'c2'" duplex="'half'">
Session object properties

session.connections['c1'].input == 'c2'
session.connections['c1'].outputs.length == 0
session.connections['c2'].input == undefined
session.connections['c2'].outputs.length == 1
session.connections['c2'].outputs[0] == 'c1'
session.connections['c3'].input == undefined
session.connections['c3'].outputs.length == 0

A full bridge is established between connections c2 and c3, connection c1 is still receiving from c2.

After <join id1="'c2'" id2="'c3'" duplex="'full'">

session.connections['c1'].input == 'c2'
session.connections['c1'].outputs.length == 0
session.connections['c2'].input == 'c3'
session.connections['c2'].outputs.length == 2
session.connections['c2'].outputs[0] == 'c1'
session.connections['c2'].outputs[1] == 'c3'
session.connections['c3'].input == 'c2'
session.connections['c3'].outputs.length == 1
session.connections['c3'].outputs[0] == 'c2'

10.4.3: Media Endpoint Properties

Connection and Dialog objects all have a common set of base properties relating to media input/output. The standard set of media endpoint properties are defined below:

10.5: Elements

10.5.1: <accept>

10.5.1.1: Overview

MUST The execution of an <accept> MUST cause the underlying platform to signal the telephony system to connect the specified Connection to the CCXML platform. The CCXML document MAY then initiate interactive dialog sessions with the incoming caller, or perform other telephony operations (e.g., place outgoing calls, join calls, etc).

10.5.1.2: <accept> Attribute Details

10.5.2: <redirect>

10.5.2.1: Overview

When a CCXML document executes a <redirect> within the <transition> block, this MUST cause the underlying platform to signal the telephony system to send the call to a specified destination. The use of <redirect> is only valid when a call is in the ALERTING and CONNECTED states.

If the call is currently joined, it MUST be unjoined before the redirect is performed and a conference.unjoined event MUST be generated. Note the platform is not required to generate the conference.unjoined or connection.redirected in any particular order.

10.5.2.2: <redirect> Attribute Details

If the platform is unable to redirect the call this MUST result in the generation of an connection.redirect.failed event.

10.5.3: <reject>

10.5.3.1: Overview

The execution of an <reject> MUST cause the underlying platform to signal the telephony system to reject the incoming Connection from the platform.

10.5.3.2: <reject> Attribute Details

10.5.4: <createcall>

10.5.4.1: Overview

A CCXML document is able to instruct the platform to attempt to place an outgoing call with <createcall> . This element MUST instruct the platform to allocate a Connection and attempt to place an outgoing call to a specified address. The CCXML interpreter MUST receive an asynchronous event when the call attempt is completed. An <eventprocessor> <transition> block can handle this event and perform further call control, such as conferencing. If the call was successfully placed, the transition block can also initiate a dialog interaction with the called party.

The execution of <createcall> MUST result in the generation of one or more connection.progressing events (depending on platform support for call progress) followed by a connection.connected event on success, or zero or more connection.progressing events followed by a connection.failed event on failure.

10.5.4.2: <createcall> Attribute Details

10.5.4.3: <createcall> examples

The following example illustrates the simplest use of <createcall> .

<createcall dest="'tel:1235551234'"/>

This example illustrates the use of several attributes of <createcall> . A SIP URI is provided as the originators caller id, a selection of protocol specific parameters are provided (callingDevice ( callingDevice and callCharacteristics) callCharacteristics ) and a string of application specific data is provided to be presented to the remote endpoint. The connection id ID for the new connection is returned in the variable "myConidVar". " myConidVar ".

<var name="myConidVar"/>
<createcall
 dest="'sip:+1-212-555-1212:1234@gateway.com;'"
 callerid="'sip:j.doe@big.com'"
 connectionid="myConidVar"
 aai="'This is application specific data'"
 hints="{callingDevice: 'notSpecified', callCharacteristics: 'voiceUnitCall'}" />

 hints="var tmp = {callingDevice: 'notSpecified', callCharacteristics: 'voiceUnitCall'}; tmp"/>

10.5.5: <createconference>

10.5.5.1: Overview

A CCXML document can attempt to create or attach to a Conference Object using <createconference> . This element instructs the implementation to allocate a Conference Object using the specified options. The successful execution of <createconference> MUST result in the generation of a conference.created event. If for any reason the implementation is unable to create the Conference Object using the specified options it MUST fail with a error.conference.create event.

Since conferences are global in scope, it is possible that other sessions will establish or terminate bridges to a Conference created by an application. Sessions that have created or attached to a Conference using <createconference> receive notifications of any bridges that are created or destroyed, through the 'conference.joined' ' conference.joined ' and 'conference.unjoined' ' conference.unjoined ' events. However, if a session establishes or terminates a bridge between a Connection that it owns and a Conference that it has created/attached to, it will only receive one such event - not one event each for the Connection and for the Conference.

It is legal for a session to perform a <createconference> multiple times with the same value for the 'confname' ' confname ' parameter. The same conference ID will be returned in each case, allowing this to be used as a mechanism for looking up conference IDs.

10.5.5.2: <createconference> Attribute Details

10.5.6: <destroyconference>

10.5.6.1: Overview

A CCXML document is able to instruct the platform to attempt to detach from an existing Conference Object using <destroyconference> . This destroys the conference if no other sessions are attached to it. The target Conference Object is identified using the conferenceid attribute. The successful execution of <destroyconference> MUST result in the generation of a conference.destroyed event. If for any reason the implementation is unable to deallocate the Conference Object it MUST fail with a error.conference.destroy event.

Since other sessions may have created bridges to a conference using the conference's ID, but without performing a <createconference>, <createconference> , destroying a conference MAY affect the bridges established by other sessions. If any bridges are terminated in this fashion, a 'conference.unjoined' ' conference.unjoined ' event MUST be posted to indicate to the session that its Connection is no longer bridged to the Conference.

The platform MUST implicitly tear down any existing bridges to the dialog and send a conference.unjoined to the CCXML document once the media paths have been freed.

10.5.6.2: <destroyconference> Attribute Details

10.5.7: <join>

10.5.7.1: Overview

A CCXML document can attempt to create a bridge between two connections, conferences, or dialogs using <join> . This element instructs the implementation to bridge the connections, conferences, or dialogs specified using the id1 and id2 attributes in accordance with media options specified by the other attributes of <join> . The successful execution of <join> MUST result in the generation of a conference.joined event. If either of the id's are invalid an error.semantic event MUST be thrown. If for any reason the implementation is unable to create the bridge using the specified options it MUST fail with a error.conference.join event.

Any Connections or Dialogs referenced by the 'id1' ' id1 ' and 'id2' ' id2 ' attributes of <join> MUST be owned by the session performing the <join>. <join> . If 'id1' id1 or 'id2' id2 refer to Conferences, then it is not necessary that the session has performed a <createconference> to create/attach to that conference; it is sufficient that it has a valid conference ID.

When joining a Connection or Dialog to a Conference, or when joining two Conferences, the 'conference.joined' ' conference.joined ' event MUST be posted to all sessions that are attached to the affected Conference(s). This MUST NOT result in multiple 'conference.joined' ' conference.joined ' events if the session performing the <join> is attached to the conference, nor if any session owns both conferences when two conferences are joined together. If the implementation is unable to join the objects an error.conference.join MUST only be sent to the session that issued the <join>. <join> .

Implementations MAY disallow two dialogs from being joined together. If both id1 and id2 specify dialogs and the platform does not support joining together two dialogs, then an error.conference.join event is thrown.

10.5.7.2: <join> Attribute Details

10.5.8: <unjoin>

10.5.8.1: Overview

A CCXML document is able to instruct the platform to attempt to tear down a bridge between two existing connections, conferences, or dialogs using <unjoin> . This element instructs the implementation to tear down the bridge between two connections/conferences/dialogs specified using the id1 and id2 attributes. The successful execution of <unjoin> MUST result in the generation of a conference.unjoined event. If either of the id's are invalid an error.semantic event MUST be thrown. If for any reason the implementation is unable to terminate the bridge between the specified connections/conferences/dialogs, or if no such bridge exists, it MUST fail with a error.conference.unjoin event.

Any Connections or Dialogs referenced by the 'id1' ' id1 ' and 'id2' ' id2 ' attributes of <unjoin> MUST be owned by the session performing the <unjoin>. <unjoin> .

When 'id1', 'id2', ' id1 ', ' id2 ', or both reference to a Conference, the 'conference.unjoined' ' conference.unjoined ' event MUST be posted to all session that are attached to the affected Conference(s). This MUST NOT result in multiple 'conference.unjoined' ' conference.unjoined ' events if the session performing the <unjoin> is attached to the conference, or if any session owns both conferences when two conferences are being unjoined. If the implementation is unable to unjoin the objects an error.conference.unjoin MUST only be sent to the session that issued the <unjoin>. <unjoin> .

10.5.8.2: <unjoin> Attribute Details

10.5.9: <disconnect>

10.5.9.1: Overview

A CCXML document is able to instruct the platform to disconnect a Connection by using <disconnect> . The underlying platform MUST send the appropriate protocol messages to perform the disconnect, and send an asynchronous event to the CCXML document when the disconnect operation completes. A CCXML document may use <disconnect> to abandon an outbound connection created using <createcall> which has not yet entered the CONNECTED state. If <disconnect> is used to abandon an outbound call, it results in the generation of a 'connection.failed' ' connection.disconnected ' event.

If the connection had been bridged when the <disconnect> request was made, the platform MUST tear down all bridges to the connection and send a conference.unjoined to the CCXML document once the media paths have been freed.

Note the platform is not required to generate the connection.disconnected / connection.failed or conference.unjoined in any particular order, unless an outbound call was abandoned in which case a connection.failed connection.disconnected event MUST be generated.

10.5.9.2: <disconnect> Attribute Details

10.5.10: <merge>

10.5.10.1: Background

Many of the network environments in which a CCXML implementation may be expected to operate provide facilities by which two existing calls can be merged into a single call at the network level. The following diagram illustrates this for the case where user A has two independent calls with users B and C, and utilizes this functionality to merge the calls together:

Figure 1: Initial Call State

Figure 2: State Following Merge

The diagram illustrates the call control connections, or sessions, that exist between users A, B, C and the network that connects them; the media streams between users is not shown on the above diagram and may differ from the path for call control that is shown. The media path between users may or may not be affected as a result of the merge, depending on the properties of the underlying network; typically any media streams to user A would be terminated since call control sessions between user A and the network are terminated.

There are many different implementations of the merging capabilities described above, across both PSTN and Voice-over-IP networks. Known implementations include the following:

• Release Link Trunk (RLT) [ISDN, (ISDN [ Q.931/DSS1 ], SS7 ISUP] ISUP [ SS7 ])
• Two B-Channel Transfer (TBCT) [ISDN, (ISDN [ Q.931/DSS1 ], SS7 ISUP] ISUP [ SS7 ])
• Explicit Call Transfer (ECT) [ISDN, (ISDN [ Q.931/DSS1 ], SS7 ISUP] ISUP [ SS7 ]ISUP)
• AT&T Transfer Connect (some variants) [ISDN] (ISDN [ Q.931/DSS1 ])
• Q.SIG Path Replacement [ISDN] (ISDN [ Q.931/DSS1 ])
• SIP REFER with Replaces [SIP] (SIP [ RFC3261 ])
• H.450.2 Consultative Transfer [H.323] (H.323 [ H.323 ])
• CSTA and other interfaces to PSTN switches
• Hook-flash Transfers [Analog, (Analog, T-1 RBS, E-1 CAS] CAS)

Different implementations may also have different restrictions on when and how merge functionality can be used. Some implementations might allow calls that are alerting to be merged, whereas others might only operate on calls that are already in the connected state. In addition, some implementations might only be able to merge calls in which one of the calls is an outbound call that specifically identifies the associated inbound call when that outbound call is placed (via hints on <createcall> ).

10.5.10.2: Overview

The <merge> element allows two calls being handled by a particular CCXML session to be merged together at the network level, if supported by the underlying network and CCXML platform.

If successful, the two referenced calls MUST be merged at the network level, and the connections to the CCXML platform associated with those calls MUST be terminated. A connection.merged event MUST be generated on each of the two calls affected by a merge. If the merge fails, then a single connection.merge.failed event MUST be thrown which identifies both of the connections against which the merge was performed.

The platform MUST implicitly tear down any existing bridges to the connections and send a conference.unjoined to the CCXML document once the media paths have been freed, except for the media path between the two connections being merged.

10.5.10.3: <merge> Attribute Details

10.6: Events

10.6.1: Overview

This section defines the events related to telephony operations including events related to the call state, success and failure events for the various telephony operations.

Several of the events defined in this section are associated with the change in state of a Connection Object. For instance, the 'connection.connected' ' connection.connected ' event causes a transition to the CONNECTED state. Platforms MUST perform updates to Connection Object state when events are dequeued by the EHIA before the <transition> selection process (and thus before execution of <transition> content). This is necessary such that conditional expressions on transitions, as well as executable content, can reference Connection Object state that is consistent with the event being processed

10.6.2: connection.alerting

This event MUST be emitted when the underlying telephony connection transitions to an ALERTING state or sends notification of call progress while the Connection Object is in the ALERTING state. This event is a transition to state ALERTING .

The fields of this This event are: includes the standard event and connection event properties along with the following additional event specific properties:

Information provided by the protocol prior to connection is accumulated and stored with the identified Connection object. This information MAY be available when the connection.alerting event is delivered to the application. Any further information provided by the protocol prior to connection MAY be provided in subsequent connection.alerting events, and made available in the updated Connection object. Alternatively, this data MAY be delivered in connection.signal events, and made available in the updated Connection object. This behavior is platform dependent. Call related information provided after connection MAY result in connection.signal events.

10.6.3: connection.progressing

This event MUST be emitted when the underlying telephony connection sends a notification of call progress while the Connection Object is in the PROGRESSING state. This event is a transition to state PROGRESSING .

The fields of this event are: Attribute Name Required Type Details name true string connection.progressing connectionid true string The ID of the Connection associated with this event. protocol false string The platform protocol ID of the Connection protocol. info false ECMAScript Object An object which provides additional platform or protocol dependent information. connection true ECMAScript Object An ECMAScript object reference to the Connection object identified by the connectionid property of this event. eventid true string The unique identifier for the event. This must match the sendid attribute of send , if the event was generated by a CCXML document. eventsource true string The unique identifier of the event source. This identifies the Platform Event I/O Processor that sent this event. eventsourcetype true string The name of the Event I/O Processor that sent this event. The value is platform dependent as this is an event generated by the Platform Event I/O processor. Subsequent connection.progressing events MAY be generated to support protocols which exchange multiple messages during the PROGRESSING state.

This event includes the standard event and connection event properties along with the following additional event specific properties:

10.6.4: connection.connected

This event must be emitted when a call an incoming connection is accepted successfully, or as when an outgoing connection is answered. This event is a transition to state CONNECTED .

The fields of this This event are: includes the standard event and connection event properties along with the following additional event specific properties: