Not everyone wants to use the keyboard to input text into computers.
For people who have physical limitations such as repetitive stress injuries or who simply do not want to hunt around with their fingers when the weather turns cold and dry, voice recognition software is one possible solution.
IBM's Via Voice is a good product, although it is available only to English speakers. It is very easy to use and the accuracy of such programs is quite high as the English language pronunciation is based on roman letters.
In the Chinese realm, it is far more difficult to apply voice recognition because: There are many characters and meanings for the same pronunciation; Cantonese is more difficult to pronounce correctly; Chinese is composed of single-syllable characters, with numerous homonyms and complicated phrase combinations.
All of these make Chinese voice recognition more difficult. Lernout and Hauspie recently introduced a program called SP3, which the software firm hopes will solve this problem.
It uses PC-based, Cantonese continuous-speech recognition and also has a pen-based system for back-up input.