HN Explorer

Show HN: My one-hand keyboard software. Based on muscle memory, learn in ~5 min.

16 points by pkamb ↗ HN
It's based on two-hand touch typing muscle memory, which means if you're already a touch typist you'll be able to type -- fast -- with one hand in minutes.

http://www.staylimber.com

How it works: Examine how the keyboard is laid out. The motion you use to type "G" is the same motion you use to type "H". Index finger, home row, towards the center of the keyboard. Same with "E" and "I": middle finger, top row. It's the same motion, but with the other hand.

Let's say you broke a few fingers on your right hand. You can only use your left hand to type. Using Limber, all left-hand characters are typed normally. When you come to a right-hand character, type the left-hand equivalent instead. It's the same motion as the key you should be typing, the only difference is that you're using your good hand instead of the injured one.

Our brains are amazing. Your left hand already has the muscle memory to type right-hand characters. It's kinda like that thing where you can't draw a circle with one hand and a square with the other at the same time. The wires are crossing somewhere.

So, for example to type "this" you would hit the keys "tges". Left-hand characters are typed normally, and for right-hand characters type the left-hand equivalent instead.

So the raw key input "tges" is gibberish, made up of all left-hand keys. My software runs predictive text algorithms as the input is coming in. So it always appears as if you're typing the most likely word, "this".

Who it's for: For hackers. Stay productive and code while your finger is sprained or your arm is in a cast. Also for recent amputees and other permanent injuries. Or hold a baby in one hand, type with the other. Or mouse with one hand, type with the other. Etc.

Try it out! Free trial on the website. It's an unlimited trial, but makes you wait 60 seconds after you type too many words. Windows only for now, let me know if you want a Mac version in the comments.

I'm very excited to get this out there. Let me know what you think. Would you ever use it? Pay for it? Get your company to pay for it? I'm open to any and all criticism/suggestions/ideas, from the base concept to any typos on the website. Thanks!

12 comments

[ 98.8 ms ] story [ 659 ms ] thread
pedalpete
Though I get what you've done, and think it is a very cool solution. I'd really like to see a video of somebody actually using it.
dwwoelfel
This comment is typed with your program. It takes some practice, but it does work pretty will.

This is similar to http://blog.xkcd.com/2007/08/14/mirrorboard-a-one-handed-key...

You might read his post for more potential uses. If you mention it to him, perhaps he will feature you on his blogz

pkamb
"does work pretty will" - :) I see you hit one of the issues that's of course present in all predictive text programs. "will" has the same finger motions as "well" so Limber simply chooses the most frequent of the two words. Tap "Tab" to switch between the different word options.
dwwoelfel
One thing that you might consider changing: leave the right (or left) hand side of the keyboard unaffected. You'd have to choose a set typing hand, but it would make fixing errors much easier.
pkamb
That's a great idea. Added to the To Do list.
jjcm
How are you planning on dealing with collisions? As dwwoelfel's comment pointed out, "well" and "will" both have the same keystroke motions. I remember trying to implement something similar back in the day, and resorted to using the spacebar to essentially flip the keyboard. What are you plans for it?
pkamb
Most frequent word is inserted by default. Press "tab" to toggle between words. I'll post a list of the most common collision pairs in a bit, luckily they're fairly rare.
pavel_lishin
I did something similar to this, except I used the left CTRL to swap keyboard sides, using some Windows software that let you map custom hotkeys and events.

Any plans for an OS X version?

bigohms
Interested to know: Is the "predictive text algorithms" checking against a dictionary? What can someone do about conflicting entries (the entry from left hand matches an equivalent word on the right hand side)? Can you override the prediction?
hansy
Where did you get the inspiration for something like this? I think it's super sweet to run into hackers who implement concepts I learned in my boring neuroscience classes!

As I develop into a programmer in my own right, I am noticing with increasing frequency the tremendous overlap between computer science and regular ol' science. Fun little projects like staylimber can reveal a great deal about underlying neural circuitry without the hassle of messy frog guts.

Great work! Oh and Mac please!