Assuming that you have used a mobile phone, you must be familiar with the T9 predictive text technology used for text input on your mobile phone. Text input is ‘predicted’ based on a dictionary of words.
CleverTexting, an Indian startup out of Noida, claims to have developed predictive text technology, which is not based on a dictionary but rather on the statistical nature of a language. For example, using the dictionary approach will not work with the Polish word Grzegorz or the Swahili word Nyumbizi.
Since they base the probabilistic analysis on the nature of the language, the next predictions are generated from what actually belongs to the language.
The predictions are shown on the screen and you use the numeric keypad to make the choices, always with a single keystroke. No multitap, no mode change, no disambiguation. That user interface , with no characters fixed to the face of the numeric keypad and dynamic allocation of characters to the keys is also new and we have applied for a patent for it.
The prototype is current available as a downloadable Java app and a touch screen implementation is in the works.
The company also offers other apps like SMS compression and Panini – the Indic language keyboard app for typing in Indic languages. If you get a chance to try out their apps, let me know your experience.
