Virtual keyboard design for mobile devices

by admin on September 1st, 2009

Everyone, or at least most of us, knows how easy it is to make mistakes when typing with on-screen keyboard in todays handhelds: missing tactile feedback, cramped keyboards and so forth. There are countless of problems, but one of them isn’t actually forced by the dimensions and feel of the device, rather than just being an ancient millstone that the industry drags around: the keyboard layout – the QWERTY.

QWERTY

QWERTY -layout for keyboards was designed during mechanical typewriter era, when it was problem with individual characters to clash and jam together when typing rapidly. In 1870’s fellow named Christopher Sholes designed a layout which would position the characters such way that it would minimize the chances of two sticking together, thus making the machine more reliable.

QWERTY keyboard layout

Wikipedia – QWERTY

DVORAK

Later on along came professor named August Dvorak, with his DVORAK keyboard layout. This layout was inspired with efficiency and fatigue problems of QWERTY keyboard and aimed to fix all these issues by studying letter frequencies and hand physiology. Still, as 99.99% of the computers in the world use QWERTY derived layouts, DVORAK has been unsuccessful to gain actual ground on market shares. There are two main reasons for this A) because people are reluctant of learning a completely new keyboard layout and B) because practically DVORAK layout only applies to cases when you type on English language, other languages have different letter frequencies. Also, when you are typing with most of your digits you are using muscle memory, and this increases the effort needed for learning a completely new keyboard layout.

DVORAK keyboard layout

Wikipedia – DVORAK

Virtual keyboards

Either of these designs is valid for touchscreen virtual keyboards (on mobile devices), where the user scenario usually is about user who is using only one or two fingers for typing with QWERTY layout, with small individual buttons and hardly no tactile feedback. All this accumulates to error prone and slow typing as the layout just isn’t designed for such use cases.

Possible solution

For starters, the keyboard layout would need complete overhaul with design process close to the idea Sholes had, over century ago: Keys would need to be laid out in such way that the most common characters, in a language X, would be farther away from each other (like most common vowels), preferably equally on left and right side of the keyboard (this minimizing fatigue of either hand). Also, keys around these would need to be set in a such way that neighboring keys would have different probabilities of coming after a character from another group (like how Z hardly ever comes after X), thus making it easier for the software to make educated guesses of which key the user actually meant to press. This, though, would need extensive study of the language that the keyboard is designed for. This would also mean that, as the keyboard layout is designed for specific language, the virtual keyboard itself would need to provide easy way to change the layout to different language.

Even though DVORAK hasn’t still gained any substantial market share, I don’t think on-screen keyboards would necessarily be suspect to the same treatment as you use just one or two fingers and hardly utilize any muscle memory when typing with virtual keyboard. Because of this it is easier to learn new layouts since brains are doing most of the work of finding keys for you to hit.
QWERTY
QWERTY -layout for keyboards was designed during mechanical typewriter era, when it was problem with individual characters to clash and jam together when typing rapidly. In 1870’s fellow named Christopher Sholes deigned a layout which would position the characters such way that it would minimize the chances of two sticking together, thus making the machine more reliable.
http://en.wikipedia.org/wiki/QWERTY
DVORAK
Later on along came professor named August Dvorak, with his DVORAK keyboard layout. This layout was inspired with efficiency and fatigue problems of QWERTY keyboard and aimed to fix all these issues by studying letter frequencies in English language and hand physiology. Still, as 99.99% of the computers in the world use QWERTY derived layouts, DVORAK has been unsuccessful to gain actual ground on market shares. There are two main reasons for this A) because people are reluctant of learning a completely new keyboard layout and B) because DVORAK layout only applies to cases when you type on English language. Also, as typing with all of your 10-digits, which makes use of muscle memory, increases the effort needed for learning a completely new keyboard layout.
http://en.wikipedia.org/wiki/Dvorak_Simplified_Keyboard
Virtual keyboards
Either of these designs is valid for touchscreen virtual keyboards (on mobile devices), where the user scenario usually is about user who is using only one or two fingers for typing with QWERTY layout, with small individual buttons and hardly no tactile feedback. All this accumulates to error prone and slow typing as the layout just isn’t designed for such use cases.
Here is one solution to these problems:
For starters, the keyboard layout would need complete overhaul with design process close to the idea Sholes had, over century ago: the keys would need to be laid out in such way that the most common characters, in a language X, would be farther away from each other (like most common vowels), preferably equally on left and right side of the keyboard (this minimizing fatigue of either hand). Also, keys around these would need to be set in a such way that neighboring keys would have different probabilities of coming after a character from another group (like how Z hardly ever comes after X), thus making it easier for the software to make educated guesses of which key the user actually meant to press. This, though, would need extensive study of the language that the keyboard is designed for. This would also mean that, as the keyboard layout is designed for specific language, the virtual keyboard itself would need to provide easy way to change the layout to different languag
Share and Enjoy:
  • email
  • Digg
  • Facebook
  • Twitter
  • del.icio.us
  • Reddit
  • MySpace
  • StumbleUpon
  • Slashdot

Random Posts

From Articles

3 Comments
  1. John permalink

    Besides the Dvorak, have you seen any research on optimal keyboard layouts for the smartphone form factor? It would seem that the paradigm of using two thumbs in a more vertical layout, as opposed to ten fingers in a horizontal one, would change the optimal key placement. I’m actually interested in writing a virtual keyboard along these principles for smartphones.

  2. Thanks for the comment.

    There isn’t too much of a research for smartphones, everyone just seems to settle on all too familiar QWERTY. Also, there is at least one research that I’m aware of, made some years ago, about character frequencies and a keyboard design based on the results. Though, if I remember right, the keyboard design was based on the principle that most frequent characters were on the middle, rarest being at the borders, whereas the concept that the article talks about needs more in-depth language research to define the specific layout for reducing “misclicks” (letters which are most least to appear side by side in a word are placed side by side in the keyboard)

    Although, one could make a basic keyboard design based on frequencies, where the most frequent characters are placed well apart from each other. Also I would think of placing them in a brick-wall like pattern, which would help to determine some of the misclicks with rows. (ie. if two keys on a same row are pressed with a key between the two, on a row above, one could guestimate that the actual target was the key on the row above, the two keys below just happened to be pressed at the same time by accident. Of course this works both ways.)

    I would also lean more towards the horizontal layout since, when you are grabbing something with both of your hands, it’s easier to move your thumbs sideways (just by using 2nd and 3rd joint of your thumb), rather than up and down, specially if you need to move farther up/down, forcing you to ever so slightly change your grip on the thing you’re holding. Also, horizontal layout would leave space for the text on top, making navigating and/or reading the text, you are writing, easier.

    Vertical layout however, would fit nicely for one handed input, since both sides of the keyboard would be reachable by single thumb, and the effortless thumb movement (2nd and 3rd joint) would be vertical when you’re holding a phone vertically. Although, since this is much less accurate posture I would go for larger buttons and maybe even double-tap feature, something like the old phone keyboard have where you have number and three characters per key, thus making it possible to have even larger buttons and hopefully less error-prone setup.

    Also, there are few concepts around, RETYPE being the most recent one (http://www.mintusability.com/2008/11/29/retype-the-text-in-your-iphone/).

  3. Comfortabl y, the article is really the greatest on this notable topic. I agree with your conclusions and will eagerly look forward to your approaching updates. Saying thanks will not just be adequate, for the fantasti c lucidity in your writing. I will directly grab your rss feed to stay abreast of any updates. Gratifying work and much success in your business dealings!

Leave a Reply

Note: XHTML is allowed. Your email address will never be published.

Subscribe to this comment feed via RSS