Background & Motivation

JSAPI2
Supported by Conversay

The explosion of the embedded world:

  • Billions of Java ME devices
  • Network in your pocket
  • Lots of content
  • Smaller form factors
  • Reduced screen sizes
  • Limited number or size of buttons
  • Can be harder to use

Server Side Speech Approaches:

  • Speech technology on a server
  • Can be used with standard telephones
  • Uses the W3C VoiceXML markup language

Client-Side Speech Approaches

  • Speech technology on the device
  • Fits well with Java ME devices
    • Enough CPU and memory
  • Increases usability of the device
  • Minimizes latency for responses
  • Easier access to local resources
    • Screen, user data and local applications
  • Supports multi-modal interaction
  • No telephony channel noise

Many applications for speech recognition and synthesis

  • Games – “fire missiles at the red team in sector 7”
  • Data entry – “39 widgets left in bin 27”
  • E-mail – “read the message from Steve”
  • Calendar – “Is next Thursday open?”
  • Learning – “What is 8 times 7?”
  • Accessibility – “Read chapter 3”
  • Car navigation – “Take exit 34 in ½ mile”
  • System Alerts – “Your fuel level is low”

We need a standard

  • Java Speech API 2 (JSAPI2) – JSR 113
  • Based on JSAPI1 for Java SE
  • Aimed at Java ME
  • Covers both recognition and synthesis
  • Makes speech technologies easy to use

Over the past five years, Conversay as the Specification Lead has led the effort to define JSAPI2 (Java Speech Application Program Interface Version 2), the first standard speech interface for Java-based mobile devices. Listed as JSR 113, JSAPI2 is part of the Java Community Process (JCP), which oversees development of the Java language and all Java APIs (Application Program Interfaces).

In 2001, Sun Microsystems—the designer and primary proponent of the Java, the most popular programming language for applications in cell phones—selected Conversay to lead the design of JSAPI2. Subsequently, Conversay assembled a group of experts from the leading Java industry participants, including Nokia, Motorola, IBM, Intel and TI, to form an expert group to assist in the design process and to speed adoption.

As the Specification Lead, Conversay maintains ownership of the Specification and is additionally responsible for creating the Reference Implementation (“RI”) and the Technology Compatibility Kit (“TCK”), two products necessary to establish a standard within the JCP.

In June of 2004, the Community Draft of JSAPI2 was released by the Expert Group to the Java Community for review. Conversay then led the Expert Group in the preparation of the Public Draft. This task was completed and the Public Draft was released by the Expert Group on March 2005. The public recommendations were collected, reviewed and, where appropriate, incorporated into the Proposed Final Draft.

Conversay’s Early Release RI has been completed and will be launched at JavaOne, May 8, 2007; the TCK is scheduled for completion shortly thereafter. The JSAPI2 standard is now about to go public, the world’s first industry standard for the use of speech on embedded devices. In addition to publishing the finalized JSAPI2 documentation (i.e., API’s), the Final Release will also include the RI, the TCK, and licensing terms.

Following the approval process by the Executive Committee, Conversay will officially, publicly release JSAPI2. Coincident with this release, Conversay will commence a communications campaign to over 100 wireless carriers, dozens of cell phone manufacturers, and the global Java programming community, estimated to include over 4 million developers. Conversay has completed partnership agreements with Symbian and Qualcomm. Discussions and negotiations with a number of JVM (Java Virtual Machine) and OEM vendors are on-going.

Conversay believes the opportunity represented by its Java leadership is very significant. 160 Telecommunications Carriers have licensed J2EE (Java Servers), and Java is the leading transport language for downloading applications to cell phones, 80% of which already support Java (over 800 million). Java developers worldwide are estimated to grow beyond 10 million, and the number of devices with Java to over 3.5 billion by the end of the decade.

Most importantly, the difficult task of establishing industry leadership consensus on a new standard has already been completed.

In addition to engineering work undertaken internally on JSAPI2, Conversay has been working with key players in the industry in position to facilitate application integration. Among these is our partner Skelmir, a premier JVM (Java Virtual Machine) vendor in the Pocket PC and set-top box markets. Skelmir counts Motorola among its major set-top box customers. Our partnership with Skelmir has resulted in porting JSAPI2 onto Pocket PC and Java-enabled set-top boxes using the Skelmir JVM. This enables JSAPI2 application developers to go after the Java SmartPhone market, as well as the set-top box market.

Computer Conversation is two-way speech between man and machine. As an interface medium, Conversation is faster, easier, more intuitive and in many cases safer than point-and-click (or other tactile interfaces) in combination with GUIs (Graphic User Interfaces). In fact, the best conversations combine the best features of VUIs (Voice User Interfaces) and GUIs in a dynamic, multi-model interaction with information systems.

An important element in the development of Computer Conversations is standardization of the hardware/software platforms. With the establishment of standards, developers can dramatically expand their market reach. Standards overcome the barriers to commercialization created by the profusion of operating systems, processors and technologies in the global device market.

Conversay is dedicated to the creation and adoption of standards for speech. We have led one such the development effort, JSAPI2 (Java Speech Application Program Interface Version 2), the first standard speech interface for Java-based mobile devices. Listed as JSR 113, JSAPI2 is part of the Java Community Process (JCP).

Java is the most popular programming language in the world for small devices, especially portable devices such as cell phones, cameras, and music players. In 2006 more than 80% of all the cell phones shipped in the world – more than 800 million cell phones - contained Java. All modern desktop computers also contain Java. In total more than three billion devices use Java worldwide supported by more than 4 million programmers, a number expected to triple by 2011 becoming the largest body of developers in the world, servicing the software and application needs of the largest body of computers in the world.