Triangular Keys for Your Touchscreen Virtual Keyboard
A post about a novel keyboard caught my attention a couple days ago. It’s got a cutey name that I like: Crocodile Keyboard.
Most discussion so far have been around whether Apple would adopt this design for its ubiquitous iPhone. However, little attention has been given to its real performance and I don’t think any discussion on adoption can be meaningful without answering the performance question first.
The inventor said this keyboard has a bigger “dead space” to prevent the users from pressing the neighboring keys unintentionally. This might be a favorable argument for a physical keyboard design, which is tactile, but not so much for a virtual keyboard.
Not registering a tap in the dead space can be more annoying than getting the wrong character. So ultimately the implementation has to honor a tap in the dead space and interpret it as the key closest to the coordinate.
Now the remaining magic lies in the visual presentation. You give users a smaller visual target to aim to enforce higher accuracy, but this is going to harm the performance based on the beloved Fitts’s Law.
There is an accuracy v. speed tradeoff here. Users will notice that they are slower when typing on the crocodile. When they try to speed up, they’ll start getting more errors like they used to get on the regular virtual keyboard. Assuming the user’s motor skill stays the same, how is typing on the crocodile different from typing on a regular keyboard carefully?
One key question is whether the visual presentation enhances the motor skill that’s guided by it. Perhaps a more plausible version would be, does a stricter target help train the user’s motor skill better over time?
If the result is false, or if the effect is only marginal, then the performance gain will not be substantial enough to expect any commercial interest.
Great analysis! I think as well that with triangular keys that the letters inside of them will be smaller, making them harder to see, and this, also more dificult to use, especially for non-letter keys (number, puncuation, symbols) that aren’t in the same locations as a full qwerty keyboard.
The white key on white background illustrated above does not represent what I believe will be the final version. If the light centre triangle key sits within a dark triangular background, I think that there will be tendency for more accurate key strikes for those that do not “touch type” (i.e. they look at each key). The background research I have done supports this conclusion. The way humans respond to triangular shapes allows them to find the centre of the key quickly and accurately.
The scaleability requires further testing but the prototype I have seen suggests that it works very well. The issue of seeing the letter will be largely determined by the font style used, and key colour constrast rather than the size. The one I saw went down to somewhere around 6pt in Arial and I could still see it quite will even though my eyesight is not that good.
While the handheld parket is likely to the first, I think that we will see its most effective application in industrial / defence / lowlight situations where users are not habitual users of keyboards.