Tablet computers offer a unique opportunity to rethink how we type, mostly because traditional touch typing doesn’t work well on them unless the tablet is resting on a flat surface, like a desk or a lap. But most of the time, we’re holding tablets, which turns typing into a process of hunt-and-peck with one hand while the other hand tries to hold the tablet steady. Keyboard solutions from mobile phones don’t work well either; they just aren’t designed for the larger screen space, and spaced out keys make for a Fitts’ Law nightmare.
In an earlier post, I noted that typically, unless you are using a stylus, a tablet is held with two hands, with thumbs on the front or on the side of the device, and your four remaining fingers on the back. It’s a natural, grasping gesture for humans. This leads me to think there there could be two possible new typing methods: one for typing with just your thumbs using the screen on the front of the device (which I outlined as the 3x2x2 Method), or by typing with the remaining four fingers on the back of the device, using ghost fingers.
Ghost Fingers are a pattern I found when researching my book Designing Gestural Interfaces. Ghost Fingers are when a device becomes seemingly transparent so the users can “see” (onscreen) their fingers on the back of or inside of the device. (Some examples.) Ghost fingers could be employed with tablets by putting a touchscreen (it could even be just a wide capacitive strip) on the back of the tablet that would only turn on when the keyboard was deployed. It could even be a physical keyboard, although that could get annoying when simply holding the tablet. As the user pressed the (digital or analog) buttons on the back, the corresponding key on the front of the screen would be highlighted.
The fingers of your hand (aside from your thumb) have four positions: straight (extended), half curled, curled, and tucked under (into the palm). Tucked under doesn’t really help much for typing, so there are basically three positions that fingers can be used to press buttons/keys. This means at most, three columns of keys. Additionally, the index and pinky fingers can stretch away from the hand (the index finger more widely), which allowed for two extra rows of keys. Index fingers are far more flexible than the pinky, allowing for an extra row. However, stretching away from the hand limits flexibility, so the further the stretch, the less bend the finger has, and so less columns of keys. Those extra rows only gained eight keys, but it was enough to make the keyboard viable.
Depending on weight, tablets can actually be held without any fingers touching the back, with the device held between the two palms occasionally with a slight assist from the bottom of the index finger. This allows for easier finger movement and for less accidental touches of the keyboard.
With those ergonomics in mind, I designed the following keyboard:
The keyboard roughly maps the existing QWERTY keyboard to a new split configuration, much like ergonomic keyboards currently do. Stretching the index and pinky fingers allow access to more keys. Here’s what it would look like on an iPad:
I prototyped it with simple paper taped to the back of an iPad. I had to adjust the keyboard layout and key size several times until it felt right. Done incorrectly, it felt like too much of a stretch to reach keys, or the curl of the finger was too much. Note that the “keys” on the back would likely not need visuals like I’ve shown here. After all, they are on the back of the device, unseen by the user.
There are some limitations to Ghost Fingers typing: namely that you could never lay the device down flat to type; unless you have a USB/bluetooth keyboard or a more traditional touchscreen keyboard, you’d always have to hold the device to type. There would obviously be a learning curve, and in all honesty, I wonder how long it would take to become comfortable not seeing your fingers type. Then again, people text message with their mobile phones are in their pockets, and aren’t focusing on their fingers as much as the screen with regular typing anyway, so I suspect it’s not a huge barrier to overcome.
Like the 3x2x2 method, the next step for this design would be to prototype it working on the iPad (or similar device), check its time against regular typing, and perhaps adjust the alphabet to key matchup. I don’t have any idea how this would translate for other, larger, alphabets.
Related Post: 3x2x2: A new method of thumb typing for tablet computers