Idle thoughts about how we replace keyboards

20.04, Friday 3 Jul 2020

Smartphone typing continues to be terrible. Maybe we look at possible futures to give us a way out?

Look, I’m terrible at boundaries and time-management with both a toddler and too much work is hard. So I’m spending a bunch of time typing with my thumbs, and reminded once again that it is sloooow. Keyboards are good because your fingers can prepare to hit the next key way faster than individual thumbs can move.

Alternatives to thumb-typing are: swiping keyboards and voice. Both suffer from the repair problem, which in a nutshell is: if you go wrong, how do you fix it? With swipe, you go word by word, and when you notice a problem you have to delete a whole word and just try again. You can’t “edit” without switching mode to typing. With voice, you go sentence by sentence, and repairing is even more of a context switch.

No, we need something better for smartphone typing.

So… we could ditch QWERTY?

The ideal smartphone keyboard would allow for the normal grip position with either one or two hands – and maximum efficiency of the available fingers, not just thumbs. Some avenues…

  • Maybe a keyboard on the back of the phone, or buttons running down both sides?
  • Why does a keyboard need to read taps? Could you micro-twitch your muscles instead and have that picked up somehow?

How about a chording keyboard, where you press a combination of keys? Last time I rambled about keyboards on Twitter, Tom Whitwell pointed me at the Microwriter keyboard invented in 1978 which is held in the hand and apparently faster than regular typing. With training.

Here’s a chorded keyboard for a smartphone as a hacker project (thanks Hans Gertwitz).

Way back in 2017, Ben Firshman said in this tweet what if typing was a conversation instead of a one-way thing? It could guide you towards your intent somehow – which I am super into.

What if you tapped a key, and around that key appeared the words that were most likely to be used next, and then based your movement towards those words, you saw the words that you might use next, maybe even appearing two or three words in advance, like gliding through super-intelligent autocomplete?

I remember there was work way back in 2010 about using vibration and electrostatic to create artificial textures on touchscreens. Here’s an open access paper from 2019 reviewing different methods. As you glide your finger, could the probability or otherwise of the following word be communicated as the resistance given to your finger?

The thing is, shifting from QWERTY to something different feels unlikely right now. We’re at the wrong end of the S-curve for paradigm shifts.


Instead of smartphones, we can imagine what comes after smartphones, and what the keyboard interface might be for that. And then evolve the smartphone keyboard to be training wheels for that future.

What I mean is: imagine Apple wants to get us all into augmented reality. That’s a huge shift. They might conceptually invent an input mode for that, then port components of it into the “today” to train us into using it (and to learn themselves, of course). So we get trained to use voice input to ear buds and swiping letters on smart watches – and the offer of augmented reality won’t feel quite as daunting when it comes.

So maybe augmented reality will be the next big thing. Smart glasses.

Or maybe the next big thing is - finally - ubiquitous computing. When I’m sitting on my sofa now, working, I am actively looking at five screens. No kidding. Laptop (for typing), tablet (for video calls). Watch (messages, notifications). E-ink screen (yeah, I don’t know either, but I built it and now it’s on the shelf telling me the time, and I look at it for that). Phone.

The phone is the interesting one. I use it to look up links when I’m writing a doc or on a call, then copy the link, and then the link magically gets transferred to my other devices and I paste it into the doc or into the chat. It is a super regular part of my workflow.

But what I mean is: five screens. Three keyboards. I use them as one device.

Seriously, I need one keyboard.

And if we imagine a world of smart glasses and then working backgrounds, what do we get? Perhaps…

  • Subvocalisation so I can use voice to type, but not dictation – something that assumes I have my hands on another kind of keyboard to do error repair.
  • Typing in the air, or maybe on any appropriate surface. I like using my wireless mouse for my iPad just next to me on the sofa… could I type by tapping my fingers on something nearby? Could I use a swiping keyboard by drawing in the air and having it picked up by a nearby camera?
  • Maybe blended with that gliding autocomplete idea from earlier?

Honestly I have no idea.

It feels like there’s a bunch of room to do some really interesting things here, and throwing away QWERTY and/or thumb-typing might open up some really interesting opportunities that we can’t yet imagine.

Anyway it’s Friday night which means it’s pizza night and I need to go stretch some dough. I’d be interested to dig around in this more if you know anyone thinking about it.

Follow-up posts:

If you enjoyed this post, please consider sharing it by email or on social media. Here’s the link. Thanks, —Matt.