Once seen as science fiction, computers that can speak, recognise voice commands and capture handwriting are getting a little closer to everyday reality. Companies ranging from Logitech to start-ups like Canesta and Tyzx, to big names like IBM, Intel and Microsoft, all have various research projects in the works that explore the field of natural interfaces and data input. Intel, for example, is investigating computer vision, which could someday allow people to operate computers using gestures. All of the major companies also have handwriting-recognition and -activation projects underway. IBM used this week's DemoMobile show in La Jolla, California, to unveil its ViaVoice Translator, new language translation software that lets handhelds translate English into French, German, Italian and Spanish. Phrases translated by the software can be rendered as text or spoken by the handheld. Also at DemoMobile, Pen&Internet demonstrated its Advanced Notes Recognition program, which can recognise shapes and convert handwriting to text. In addition, Canesta introduced a set of chips that will give computers the ability to recognise gestures. Using the chips, handheld devices could project images of a full-size keyboard and mouse and then use a three-dimensional sensor to translate finger and arm movements into keystrokes and mouse manoeuvre. The concept is not unlike that used by Virtual Devices for its Virtual Keyboard gadget. The company plans to sell the chips, which will be used first to create a virtual keyboard and mouse for handheld devices, to makers of PDAs, cellular phones and tablet computers. It has already begun shipping small numbers of the keyboard chips to customers. Microsoft is also chiming in; the company is taking another swing at improving the pen-driven user interface and handwriting recognition for its Windows XP Tablet PC software, which will be available on new notebook computers later in the year.





