TouchType Fluency™ Prediction
TouchType Fluency™ Prediction
The prediction engine supporting TouchType Express relies on a range of modules as shown in the diagram below. The key is the ability to provide contextual predictions, based upon multiple language domains, and adapt predictions to specific users’ styles:
This technology can be provided as a standalone prediction engine and can be implemented on a wide range of current devices (touch-screen or conventional), offering significant improvements over existing prediction and word-completion technologies.
Keystrokes Per Character (KSPC)
This is a standard means to measure the efficiency of a text input device. This relationship describes the number of keys that must be pressed (on average) in order to enter a character of text. QWERTY keyboards generally give performance of around 1 KSPC, since each button on the keyboard corresponds directly to a character. On the other hand, without using text prediction, 9-button keypads need more than one keystroke to enter a character as each button represents several characters. Devices, such as TouchType Express, which offer word completion or prediction of words, are able to achieve less than 1 keystroke per character. This enables faster text entry for a user as significantly less buttons need to be pressed in order to enter a section of text.
The graph below shows how TouchType's products compare with conventional prediction software:
- T9 the current standard for 9-button keypad mobile phones
- iTap a disambiguation/prediction technology for 9-button keypads developed by Motorola
- QWERTY the current standard 26-button keyboard with no prediction
- A standard mobile device using a 9-button keypad and TouchType Fluency prediction
- A device using a 26-button QWERTY keyboard and TouchType Fluency prediction
- A touch-screen device using TouchType Express
It has not been possible to obtain keystrokes per character data for the latest text prediction applications including XT9, eZiText, and SureType. Preliminary analysis suggests that these technologies have a KSPC measurement of between 0.8 and 1.
| Text entry device | KSPC |
|---|---|
| T9 | 1.19 |
| iTap (Motorola) | 1.01 |
| QWERTY | 1 |
| TouchType Fluency single word prediction 9-key | 0.67 |
| TouchType Fluency single word prediction 26-key | 0.64 |
| TouchType Express | 0.47 |
Indicative list of devices
The following list is indicative of the range of devices into which the Fluency prediction engine could be integrated:
| Manufacturer | Device |
|---|---|
| Apple | iPod/iPhone |
| RIM | Blackberry Storm 9500 |
| Nokia | 5800 XpressMusic, N770 Internet Tablet, N800, N810 |
| Samsung | Omnia SGH-i900, SPH-i700, Pixon, Soul U900 Midnight, Tocco F480, Giorgio Armani P520 |
| LG | Cookie, Renoir, Viewty, Venus, Prada, Nero, Secret |
| HTC | Touch, Touch Dual, Touch Cruise, Touch Diamond, Touch Pro, Touch HD, Touch Viva, Touch 3G, MAX 4G, TyTN, Titan, G2 |
| Sony Ericsson | G900, Xperia X1, W960i |
| T-mobile | G1 |
| O2 | Xda Zset, Xda Serra, Xda Ignito, Xda Mantle, Xda Stellar, Xda Orbit 2 |
| Motorola | Q9c |
| Palm Treo | 750, 700wx, 800w, Centro |
| Asus | P535 |
| Mio | A700 |
| Hewlett-Packard | iPAQ 110, 210, 614c |
| Acer | n311 |