Typing, RSI, and what I do differently

Or, "Any excuse to talk about folk music"

by Robert May on Afternoon Robot

Background

A few years ago I had quite bad RSI (Repetitive Strain Injury) caused by working entirely on a small laptop for 6 months. It was a mistake for many reasons, but the major one was that I had inadvertantly changed my typing style to work with it. I was resting my wrists on the laptop case, I was using the touchpad, and I had it quite close to my body because it was small. I had to make a fairly dramatic change to halt the onset of the RSI, but at the time I didn't understand what I had done to cause it - after all, I've been typing a lot, at a reasonable speed, since I was ~10 years old.

It took 20 years, but finally I had suffered the consequences of my constant computer use.

Changes

Immediately I set out to spend money. The first change I made was to buy an "ergonomic" keyboard, and I went for an Ergodox EZ, which is a nice bit of kit. I also bought a wireless trackball; a Logitech MX Ergo. I'm now on my second of these and I really like it.

At the same time I started trying to learn the Workman layout, which I think is largely rather smart, and which focuses on home-row typing. I initially tried Colemak but that seems oddly flawed in applying too much frequency to keys under the little fingers.

After some time, rather wonderfully, my RSI was substantially improved.

However, it wasn't actually for the reasons I originally thought. In fact, the RSI started disappearing because I essentially took a 3 month break from typing a lot, due to the high learning curve of swapping to a split board and new layout at the same time. Once I was back up to speed, the RSI started coming back.

Why does it still hurt?

Had I broken my body irreparably? No, as it turns out. In fact, all I had to do was just go back to everything I was doing before the laptop incident, and give myself time.

As in all things, folk music is the cure

Discovering the real cause of the problem came about by comparing typing on a computer to my musical instruments. I never get RSI on my diatonic accordions, or on the piano. Why are they different? Surely they use the same muscles, more or less. In fact their key weight is notably heavier. Even more weirdly, my RSI was non-existent when I played on them, even when it was at its worst on the computer.

There are a variety of ways to play the diatonic accordion but I tend to use a single strap and use the weight of the instrument to allow me to shift my hands whilst still holding it relatively still; I prefer this to strapping it tightly to my body and having entirely free hands. Here's an example video from a few years back:

There are better tunes and examples I could make but that would involve filming something new and I had this lying about. This video is a fairly good example in that I regularly slide my hand quickly downwards instead of stretching. It's more tempting to stretch when first learning the instrument as you don't want to lose your reference point, but learning to shift up and down the keys makes more difficult passages playable and actually quite comfortable. It's generally more comfortable to keep your fingers straight-ish and shift your hand up and down in that position. I now do this even more than in this video and regularly play tunes that use the full length of the keyboard. In fact, they're often my favourite ones to play! I now usually play with my thumb behind the keyboard, which I find makes it very easy to slide up and down. Here's another, rougher, old example, sure to give you motion sickness, from when I was selling an unwanted instrument:

Yes, accordions are a very quick way to ruin a motion-stabilisation algorithm-writer's day.

The hands aren't as free-floating as on a piano, as you have to hold the instrument at the same time. The key weight is actually fairly comparable to a computer keyboard, though at the heavier end. The piano is different, in that the key weight is actually pretty immense, particularly at the lower-end on a grand piano keyboard, and therefore requires a somewhat different method of striking the keys to a computer keyboard - the strength doesn't come from moving your finger up and down (but I won't go into that here as there are a million piano ergonomics articles).

Problems with ergonomic or split keyboards

The primary goal of many ergonomic keyboard enthusiasts is to reduce movement entirely and to move everything onto modifier layers. The Ergodox is known to be flawed in this regard as the thumb cluster is a bit of a stretch for most people, as well as the larger keys around the outside being a bit of an odd stretch. It's almost like it has two competing designs: one to reduce movement entirely, and the other to require movement in order to hit these extra buttons.

But the major flaw for me is that it ensures you can only ever use each hand on one half of the keyboard, and this means that keys which could often be a shared workload, such as T,Y,G,H,B are now dedicated to a hand. This is also a flaw I see with the "ortho-linear" layouts which remove key staggering. The argument for these optimisations is to remove stretching, which I agree with, but they do it by reducing movement, which I don't.

Instead of forcibly splitting the workload, I share it even more between my hands. But not by stretching! Instead my hands move freely over the keyboard, picking up letters or symbols in awkward places without having to perform difficult stretches with my fingers.

The problem with home-row typing

In my personal experience I found home-row typing to actually be unergonomic mostly because of the additional bend required from moving both hands closer together. This produces more strain on your wrist in order for your hand to be straighter on the keyboard. This is one aspect the split-keyboard design does solve, but it's also solvable on a normal keyboard.

It can be alleviated somewhat by moving the keyboard further away from you, allowing your wrists to be straighter. This however often means applying more pressure to the under-side of the wrist or arm as holding the weight of your hands further from your body is more tiring.

Instead I looked back at how I used to type, and realised that although my left hand sat more-or-less in the home-row position, i.e. with the index finger on F, my right hand sat further to the right, with the index finger somewhere between K and L. This is a wider "stance" and results in straighter wrists, as well as less stretching required for the Enter and Right Shift keys.

This is the Way

Looking at both instruments and reflecting on how I had typed before The Laptop, First-born of Satan, I settled on some ideas that are completely contrary to the accepted ergonomics of typing:

  1. The wrists themselves should not be under strain or pressure, including wrist-rests
  2. The hands should move freely over the keyboard
  3. The idea of "less movement" is incorrect, and instead I should aim for "less stretching"

So I eBay'd the Ergodox, and used the money to buy a RealForce Topre keyboard. Weirdly, I bought a Japanese layout keyboard, which has some very slight advantages specific to my work that I'll explain shortly. I also ditched the Workman layout - it's nice, and I would recommend it, but most likely for those writing more prose than those programming. It flows well for words and becomes more awkward when using, say, vim, but it was particularly awkward in combination with the flaws of the Ergodox. Ultimately I dropped it because I realised my issues weren't down to QWERTY.

One aspect of the Ergodox which is nice is the ability to cant the keyboard over or away from you - I recommend doing this. If you raise a keyboard up, say with the feet usually attached at the back, you apply extra strain to the back of your hands to accommodate for the extra angle. Having a flatter keyboard is better in this regard.

So this weird choice to use a Japanese layout keyboard huh?

RealForce Topre R2

So I bought a Japanese layout RealForce Topre R2 keyboard with variably weighted springs. I love it, it's by far the best keyboard I've ever used. It was also hideously expensive, so if anyone from RealForce is out there and wants to supply me with keyboards to shill I am 100% on-board.

The Japanese layout both is and isn't odd. The extra keys for kana writing can be ignored, I just rebind them in .Xmodmap and use them for things like push-to-talk on Discord. The rest of the layout isn't too far from the UK ISO layout. Where it gets interesting is in the symbols.

  1. Double-quotes are above 2, which I like as it splits them between the hands. I typically use Right Shift+2 to type it
  2. Single-quotes are above 7, which is hella weird at first, but actually I now really like it. I use it 50/50 with left or right shift, always using both hands
  3. Both semi-colon and colon are first-level keys. This is awesome for programming
  4. @ is a first-level key. This is particularly nice for Ruby
  5. Underscore is a bit weird, and sits directly next to a shortened shift. This was the hardest key to get used to. I tend to move my right hand over and use two fingers to press the combination
  6. Everything else is slightly shifted from the UK ISO layout but not weirdly so

Because I like it so much now, I also have a Bluetooth Japanese-layout HHKB for a travel keyboard, which is very nice. I just stick it on top of any laptop I might have to use, to show who's in charge.

Demonstration

This was surprisingly easier to film than I expected. You'll have to forgive me the dusty desk, I currently don't have a carpet in my office so it's almost entirely pointless dusting as it's back to this state within 2 days.

I've noticed my left hand moves less than my right, but you can see that it often covers over to Y and I also move frequently to 7 for the single-quote on this keyboard. I rarely use my pinky fingers on either hand for letters. The right hand moves very freely across Enter, modifiers, punctuation, and the remaining letters. My wrists are mostly straight and any pressure on the desk is further up the forearm, if I do rest my arms on the desk.

Conclusion

Ultimately I discovered that I had been doing it fine my whole life, and I just needed to take time after realising the laptop was the problem to allow my body time to heal. That can be challenging when you type every day. My suggestion on noticing the onset of RSI would be to pause, take some time off work (or schedule a lot of meetings so you look busy but no actual work gets done), and assess your setup. My personal conclusion was that using a laptop for a prolonged period was the actual issue, and the combination of a condensed layout and trackpad usage caused my RSI. Once I reduced my typing for a few months via the Ergodox experiment, and then reverted back to my original desktop setup, all symptoms disappeared entirely. If I use a laptop for more than about 4 hours the symptoms start to return.

There's nothing inherently wrong with movement when typing, QWERTY, or non-ortho-linear keyboards. What spurred this article was seeing repeated comments online saying "movement is bad", when really, there are many factors in the ergonomics of typing, and no universal solution for everybody.