A True Gesture Keyboard: Ending the Virtual QWERTY Tyranny
The familiar "qwerty" keyboard layout, originally designed for the mechanical typewriter, has been around since the later half of the 19th century. It survived the transition to the electronic computer keyboard and has become the standard way of entering text into a computer. However, when it comes to using it with a touchscreen, many problems arise. This article will discuss those problems and describe a new type of text entry that is fast and tailored specifically to work with the touchscreen.
Problems with QWERTY
Recently touchscreen versions of the "qwerty" keyboard have been used extensively in smart-phones and tablets. The "qwerty" layout, named after the top left row of letters on a standard keyboard, has worked well for typewriters and computer keyboards, but when implemented on touch-screens, it becomes much less efficient.
Any keyboard relying on pressing a separate key for each character, depends on the tactile feedback provided by the mechanical keys. When your eight typing fingers are placed above their "home" keys, you can feel the keys' contours and thus can tell when your fingers stray from their home position. This frees your eyes for other things, like the text you might be copying or reading, or the person you are with. Once the key positions have been memorized, i.e. once touch typing has been mastered, the typing can be quite fast.
This cannot be done with the touchscreen version of the "qwerty" keyboard because the screen is smooth and you can't feel the keys. Even if you could, your fingers must be off the screen between key presses. To compensate, you must use your eyes to monitor the location of your fingers. This is especially difficult with smartphones, where the key images are tiny and are blocked from view by your fingers!
As a consequence many tasks, normally done with a computer keyboard, are cumbersome with the touchscreen version. Copying text becomes particularly difficult in that your eyes must constantly jump between the text being copied and the keyboard. When in a meeting or classroom where notes are being taken, and a smartphone or tablet would really be handy, the same thing happens. The keyboard is also slow and error prone because of the eye-hand coordination needed to operate it.
There's Another Way
Although the touchscreen is clearly not well suited to emulating a "qwerty" keyboard, there are aspects of the touchscreen that lend themselves quite well to making a good text entry device. One of these is the ease of making swipes and other gestures due to the screen's smooth surface and the ability of the software to track the fingers' positions. This has been demonstrated by the success of the pinch gestures pioneered on the iPhone.
To take advantage of this feature, what if each key press were replaced by a unique gesture, like a tap or swipe?
This would eliminate the need to constantly look at the input area since the gesture can start anywhere. But how can we devise enough, easy to make, unique gestures to allow for the 26 letters of the alphabet as well as shifts and punctuations? What follows is a solution to that problem.
More Fingers: Suppose we use a small set of single finger gestures and combine them with multiple fingers to make possible a larger number of multi-finger gestures. Three fingers was chosen here to be compatible with the smaller screens found on smart-phones. To the right is a table showing 13 simple single finger gestures consisting of a tap and 12 different swipes. These consist of simple swipes in one of four directions and curved swipes where the the swipe starts in one direction and ends in another.
Four possible combinations of three fingers can be used: one finger by itself, two adjacent fingers, two non-adjacent fingers, and all three fingers together. When using more than one finger, all of the fingers simultaneously make one of the 13 single finger gestures.
The total number of gestures possible is thus (4 finger combinations) x (13 gestures) = 52, which is comparable to the number of keys on a typical "qwerty" keyboard. In the table to the right, these gestures have been assigned in a way that ensures that the most frequently used letters, shifts, and punctuations are paired with the simplest gestures. For example a "space", the most frequently used character is a simple one finger tap. The letter "e" is two adjacent fingers tapping. In the table a filled circle means a tap, an arrow designates a swipe, and a dash means the corresponding finger does not touch the screen.
To aid in learning the gestures, this table is displayed in the area of the screen, under the fingers, where the gestures are formed. It serves as a type of "crib" sheet for remembering the gestures.
The fact that the swipes can be as little as a few millimeters in length makes this technique potentially very fast. Thus far it has been demonstrated at over 40 words per minute and faster speeds are entirely possible.
Because of the lack of tactile feedback, the touchscreen version of the "qwerty" keyboard will always be slow and error prone requiring good eye-hand coordination to operate efficiently. A technique has been described which solves most of the problems encountered in doing text entry on touchscreen devices like smartphones and tablets. A unique gesture for each character is made possible using a combination of simple taps and swipes using up to three fingers. This makes possible eyes-free operation at high speed, making text entry comparable to what is possible with a conventional desktop keyboard.
This concept has been recently realized as an app for Android called Flicboard.