Reinventing the keyboard, yet again: Kannuu

It feels like I write this headline every year or two: a new company has engineered a new and better way to enter text.

It really is a clever way to get text into a small device.

(Credit: Kannuu)

The latest: Kannuu, presenting at the DemoFall conference tomorrow. This one uses the four-way directional pad of your mobile phone or remote control as the input mechanism. For each letter you enter, you get four choices (one each for up, down, left, and right) for the next letter or the rest of the word--guaranteed by the system to be the most likely things you'll want to type. You can also get another set of selections if none of the first four are what you want.

I have seen dozens of alternative text-entry schemes, each more clever than the last. Of them, only T9 has achieved any traction on mobile devices. But most people still use the incredibly inefficient "multitap" scheme for entering text on mobile phones. And all smart-phone manufacturers still offer Qwerty keyboards of some sort.

Why is it then that new input methods take hold? I submit three reasons: first, keyboards are predictable. With every button you press, you know exactly what you are going to get. That means you can type with little or no attention paid to the screen. There are people, mostly young, who can actually touch type in multitap. I believe that having to monitor the screen carefully as you type is a killer.

Second, predictive keying technologies fall apart on proper names, even when they fall back to next-letter frequency as a predictive technology (as opposed to database lookup). So whenever you want to type a name or street, you either have to slog through a prediction engine that may appear to have suddenly become brain-damaged, or turn it off entirely.

Third, it's not end users you have to convince if you want to build a successful alternative keyboard. It's the device manufacturers. And they're chicken.

So I'm skeptical, and I come to this conclusion honestly. I have been utterly wrong in predicting the success of several previous keyboard replacement schemes, each of them brilliant. None of them are in wide use today.