From Bill Meisel’s Speech Strategy News (www.tmaa.com), December 2007

 

Editor’s Notes

Google’s open-source wireless phone platform

Bill Meisel, Publisher & Editor

On November 5, Google announced a software platform for mobile phones—in effect, the rumored “Google phone,” but in a form that it can be offered by multiple vendors. The software, called Android, is open-source, under a liberal license that allows developers to use and modify the code. Android is offered through a group Google initiated, the Open Handset Alliance. T-Mobile has committed to offering a handset using the software in the second half of 2008, probably built by Alliance member HTC Corp. The offering competes with other platforms, such as those from Microsoft and Symbian, and those competitors are suggesting that the open-source software creates security risks.

Nuance Communications is one of the more than thirty companies in the Alliance, and contributed a basic open-source speech recognition module. Peter Mahoney, Nuance vice president, worldwide marketing, said the module is speaker-independent, limited to English, and targeted largely at voice-activated dialing. Developers could also license from Nuance some of its more powerful solutions and integrate those into a phone using Android. Mahoney said that the Android Application Programming Interfaces to the speech recognition module can support the upgrade. Steve Chambers, President, Mobile and Consumer Services Division, Nuance, commented, “Nuance joined the Open Handset Alliance with other industry leaders to grow the entire mobile ecosystem. We’re committed to apply our strength and leadership in voice-based search and messaging to move the market forward…We believe deep collaboration with members of the Alliance will grow our core mobile business and fuel the proliferation of speech-enabled applications worldwide.”

The Android software supports phones with conventional scrolling controls as well as phones with touch screens. There was a veiled suggestion during the press conference that Apple might adopt some part of the Android software. Eric Schmidt, Chairman and CEO of Google, is on the Apple Board.

Despite the presence of the Nuance software, the Google solution focuses on the Graphical User Interface, with a key feature being a Web browser that Google said would allow the full Web experience on a smaller device. Demonstrations showed features such as easy movement around the Web page and quick zooming. In an online demo, Google highlighted text messaging, Internet browsing, Google mapping software, and an OpenGL 3-D graphics interface. Google is apparently counting third-party developers to create most applications for the platform.

Last month’s Editor’s Notes discussed some of the limitations of GUIs in general, claiming that the increased complexity of applications and web sites was making the familiar paradigm more hierarchical with too many levels and thus less effective. Smaller mobile devices emphasize this problem.

The open platform will, of course, allow speech-focused applications, just as it will support any other applications. The same is true for other wireless operating platforms. The phone operating system is largely irrelevant for speech applications that are entirely network-based, such as directory assistance service. For such services, tight coupling with the phone is not necessary.

For applications that interact with applications on the phone, such as web browsers or the user’s contact list, tighter integration can make speech applications more effective. With software in the phone, the application can, for example, take the text of a spoken information request and enter it into a Web search box or bring up a map of a specified location.

The operating system isn’t the main issue in putting such software on the phone. The wireless service provider controls the software on the phone. An analogy is the “on-deck” versus “off-deck” dichotomy that relates to any application on the phone. On-deck solutions come with the phone and one can navigate to them using the phone’s menu. If a solution is on-deck, it is much easier for a customer to find it and use it. Off-deck solutions can be accessed through a general browser, but have the disadvantage that they must be promoted outside of the phone, making it much more difficult to get adoption by consumers. One could argue that one of the most successful off-deck services is free directory assistance, accessed through the voice, rather than the data, channel.

The control of software on the phone by service providers is controversial. Some of the spectrum in the next FCC auction is designated for services that are not controlled by the service provider, and Verizon Wireless has just announced it will allow devices and software it doesn’t sell on its network. Google has a test network on its campus and prototype phones running Android on the network, and may bid for spectrum to ensure availability of an open network.

The importance of the wireless channel to companies such as Google is indicative of a clear trend toward making such devices central to most people’s activities—a personal assistant. A GUI-based approach will be at least partially successful, as evidenced by Apple’s iPhone, but speech applications may help overcome some of the limitations of the GUI on such devices. Ultimately, a “Voice Search” model—just-say-what-you-want or type-what-you-would-say—may drive software on the devices. More open networks will make it easier to show the power of Voice Search and a personal assistant model.

(For further business-oriented insights on Voice Search, attend the Voice Search Conference (www.voicesearchconference.com), organized by Bill Meisel and the Applied Voice Input Output Society.)