用于移动设备的新型输入法——XNav

微软的XNav 技术是一种用于移动设备的新型输入法，它只需要几个按键甚至不需要按键，从而提高了输入速度并降低了输入疲劳度。XNav 拥有特殊的潜能，可以为使用移动设备在 Internet 冲浪、以及发送即时或文本消息的非西方语言用户改进文本输入。

 有兴趣的朋友可以在此网址获得更多详细内容：
http://www.microsoft.com/about/legal/intellectualproperty/search/details.mspx?ip_id=IDAEI4AE&techType=&ipCat=Any&feeStructure=Any&keywords=&ipVenture=true

微软的对这项技术的预览原文如下，有时间再翻译吧：

XNav

Technical Overview

Overview

Microsoft Research has developed XNav, an alternative text input and device navigation technology, which is designed to define a consistent set of gestures for direct command and control and text input for a wide variety of devices. It is generic enough for use on mobile devices, set-top boxes, remote controls, phones, mini-computers, etc. It allows rapid, reliable, and consistent text input on devices with a variety of input hardware, such as numeric keypads or touch-screens.

 

XNav is based on a technology called Virtual Serial Chording, which has the following properties:

 

• The input device is divided into nine zones, which correspond to the number keys on a standard telephone numeric keypad.
• With the exception of numbers, all text input strings begin and end in the Hub, zone 5.
• Each of the following steps is executed without the user lifting his or her finger/stylus. In fact, even spaces and backspaces are doable without lifting the finger, so the user does not need to lift the finger or the stylus until all text input is done.
• The first stroke is from the Hub to the petal zone where the representation of the desired character is found on the software user interface.
• Without lifting the finger, the second stroke travels from the current petal to the zone relative to the character’s location on that petal.
• To execute a letter or character, the user returns to the Hub, indicating this selection.

 

Examples:

a zones 5 – 1 – 5   b zones 5 – 1 – 2 – 5   c zones 5 – 1 – 2 – 3 – 5   space (“sp”) zones 5 – 6 – 5

E zones 5 – 2 – 5   f zones 5 – 2 – 3 – 5   k zones 5 – 3 – 6 – 9 – 5   backspace (“bksp”) zones 5 – 4 – 5

 

Because each stroke begins and ends in the Hub zone, each letter flows into the next, and the user quickly encodes shapes that represent words:

 

“the”	“cool”	“cat”
5 – 7 – 5 – 3 – 2 – 5 – 2 – 5	5 – 1 – 2 – 3 – 5 – 7 – 5 – 6 – 7 – 5	5 – 1 – 2 – 3 – 5 – 1 – 5 – 7 – 5

Gestures that begin and/or end on a petal, rather than the Hub, represent non-text string inputs, such as navigation, text selection, word completion, dial controls, or predefined commands. The core system even defines gestures for commonly used emoticons.

Components

The core of this technology consists of the software algorithms that translate keyed input into text. These algorithms are hardware-independent, and can interface with text services such as Input Method Editors (IME).

 

In addition to the software algorithms, Microsoft Research has prototyped a unique capacitance touch pad/wheel input device that improves text input further.

Core Scenarios

• As an input methodology for any mobile device, regardless of input hardware.
• For entering characters from non-Roman alphabets. The large number of possible gesture combinations maps well to alphabets with a large number of characters.
• Text input for users who need to input text without looking at the device screen. This input technology allows users to effectively “touch type” on a mobile device once they have learned the gestures.

Related Work

Portions of the core technology were acquired through a collaborative research relationship between Microsoft Research and the Center for Advanced Technology at New York University.

Differentiating Features

A number of text input technologies exist for cell phones and other mobile devices. Examples include multi-tap, T9, Palm Graffiti, etc. These technologies are frequently hardware specific (e.g., Graffiti requires a touch screen and stylus).

 

This approach works with any device that has dual axes (x,y), e.g., game controller, cell phone, set-top box remote, etc. The primary advantages of this technology over other competing technologies are:

 

1. Text input is continuous (no pausing for cursor advance).
2. Text input can be done with one hand while holding the device.
3. Designed for raw speed, with the least possible fatigue. The system requires fewer keystrokes than other methodologies.
4. Once the letter pattern is learned, a user can input text without looking.
5. Text input gestures also match gestures frequently used by video games.

 

The technology also utilizes a predictive matching dictionary to perform word completion.

Demo/Prototype

Development Status

Microsoft Research has created prototypes of this text input technology that work on Smartphone devices via a numeric keypad, Windows CE devices, and Windows XP machines via mouse input. Researchers have used these prototypes as their primary input tools for several months.