After 150 years, we should finally redesign the computer keyboard | Hacker News Reader

After 150 years, we should finally redesign the computer keyboard

blog.prototypr.io

63 pointsdanny00118 comments

I worked a lot in human-computer-interaction a couple of decades ago, and people always try to make a better keyboard, and outside of niche uses, they have failed, except for small, incremental tweaks. In fact, it is an example I often use of how, most often, when people try to reinvent something, they instead end up learning why the existing thing is so great. I know I've done it multiple times with various technologies. It's a great learning experience, but my money is betting on the keyboard staying the keyboard :)

I think there is already better keyboard designs but it's hard to compete with the one that everybody is already familiar with, fits easily on laptops and costs dirt cheap to buy and make because the volume is huge. Most people aren't going to spend $300+ on a keyboard that they have to learn how to use and businesses aren't going to want to spend the money and time training employees how to use it. The inertia of the standard keyboard is far too high.

I suppose if one has the issues mentioned in the article, a few hundred dollars isn’t much

Aren't keyboards with control, shift, escape, etc... where the thumbs should be objectively better in terms of ergonomics, though? At least if you're hitting those keys all the time, since the thumb's the strongest finger. I'm seriously considering getting a split keyboard after realising just how much I'm going to type in the following years and how much Emacs and/or Vim I'm using.

Something something Emacs pinky.

Binding Caps Lock to Control helps with that. It's still far from ideal, but it is at least no longer actively terrible.

s/Control/Escape perhaps?

That's what I do. It definitely helps. Heavy vi(m) users tend to map it to Escape, but I find that:

- I use Control a lot more than Escape (especially in the shell, where I spend a lot of time); and

- Escape is already not in a horrible position, and it's not a modifier key so I don't need to twist and curl my wrist and fingers in order to type a combination of keys.

I have capslock configured to act as control when held down, and as escape when tapped. Best thing since sliced bread.

I don't understand why one would set Caps Lock to Escape. As a vim user, I set it ton Control. Control+[ is effective for my Escape needs.

Spacemacs leader key setup is very nice (I'm sure you could achieve it without needing Spacemacs) as nearly every command you do begins from the spacebar, and thus your thumb. Haven't gotten a sore pinky since. Well, actually, I never used the normal Emacs keybindings. I learned my Emacs knowledge starting with Spacemacs, but I never cared to learn the default bindings anyways.

> when people try to reinvent something, they instead end up learning why the existing thing is so great

As someone who has actively tried to switch to Dvorak and Colemak keyboard layouts and failed, I'd argue it's less because that Qwerty is so "great" and more that it's just so entrenched everywhere, that unlearning it for keeps is difficult.

Like I'd have to set up Colemak everywhere I go, if I sat at someone else's computer or a public computer it would be back to Qwerty, if you ever look down at your keys for reference (something you want to do a lot when you're learning a new layout) they're telling you Qwerty, etc.

Also the unlearning process can be slower than I'd like, and my typing speed plummets while I'm learning these things, so I'm less useful at work or school or working on personal projects. I've been using Qwerty for over 20 years at this point, whereas my Colemak experiment was only a few months.

It did seem like I was straining my fingers less when I was using Colemak, though, and could potentially type faster eventually if I could ever fully commit to it.

> Like I'd have to set up Colemak everywhere I go, if I sat at someone else's computer or a public computer it would be back to Qwerty, if you ever look down at your keys for reference (something you want to do a lot when you're learning a new layout) they're telling you Qwerty, etc.

I use Dvorak. One of the first things I did when switching to Dvorak was buying a keyboard with hardware Dvorak layout and a black skin. The hardware layout meant that to use it on any computer all I had to do was set the computer I wanted to use to US English. The black skin meant I couldn’t cheat and look – it was painful for a short while and it was extremely effective in forcing me to learn the layout.

The fact that other computers are in QWERTY don’t matter. In fact on my laptop I use QWERTY when using the builtin keyboard. On my phone I use QWERTY. But with my keyboard I use Dvorak. You don’t need to unlearn QWERTY to learn Dvorak. In fact it’s better to keep using QWERTY on the side like I do, because you still reap the benefits of Dvorak when you use Dvorak, without becoming incapable of using “normal” computers and devices.

The original keyboard I bought when learning Dvorak was the TypeMatrix 2030 USB ortholinear keyboard with an all black skin.

My current keyboard, that I’ve been using for the past few years, is the programmable mechanical ortholinear keyboard ErgoDox EZ Shine. All black.

For my ErgoDox EZ Shine I created a custom Dvorak-based layout that I use: https://github.com/ctsrc/ergodox-ez-shine-dvorak

I'm glad you figured out how to make it work for you. But in a way you're kind of proving my point when you're saying you use QWERTY on your laptop and your phone. You may be able to context switch fine but I was having a hell of a time. Also last time I tried was probably 15+ years ago.

I've used Dvorak for over 15 years.

I keep Qwerty on my phone, since the advantages of Dvorak are disadvantages on a phone keyboard -- I want the letters to be spread all over the place, so the swipe-to-type thing works.

Jobs differ, but I very rarely type more than a word on computer keyboards other than my own. If a colleague asks me to help with something, I almost always think it's more polite to ask them to type.

You need to learn qwerty. Then you need to learn Dvorak (or in my opinion the far superior Colemak). Then you need to learn to switch between them. It can be done but it’s three separate stages.

Agree. Thinking back, I remember that for after I learned to use Dvorak I did actually struggle with QWERTY for a while. So for some time I stuck to only Dvorak. And then I learned to switch between both.

I ran into these issues myself.

I've been typing in dvorak at home and qwerty at work for about 20 years now. I was having pain in my fingers typing all day on qwerty, and sought out another way of typing hoping it would help. Switching to dvorak + qwerty permanently solved the pain problem for me.

I was unaware at the time how difficult it would be. There is no incremental approach to switching keyboard layouts, at least for me. As soon as I started learning dvorak, my performance in qwerty dropped drastically. I'd blank on what key to type or I'd spend more time correcting dvorak keys in qwerty or qwerty keys in dvorak errors than I was actually typing. I was at a severely degraded typing ability on any keyboard layout for a week or two and it was a couple of months before I felt comfortable in either again. It was several months after that before I could easily context switch between them without errors. Occasionally now I'll start typing gibberish and realize that I need to switch my brain over because I'm at work now and need to type in qwerty, but once I've made that context switch I only have normal typos for the layout I'm in instead of the cross-layout problems I used to have.

I don't know if I'm just terrible at learning keyboard layouts or if this is something everyone will encounter, but I wouldn't recommend switching over to anyone that doesn't have a good couple of weeks to effectively be at the hunt-and-peck typing level again and a long while until they can expect to type long documents without multiple errors.

I don't use a hardware dvorak keyboard, I still use the qwerty layout keyboards with a different keyboard layout in software. This may have been part of the problem I had learning dvorak. I learned to touch type without looking at the keys in high school on IBM Selectric typewriters, so that helped a lot.

> Like I'd have to set up Colemak everywhere I go, if I sat at someone else's computer or a public computer it would be back to Qwerty

When I switched to Colemak (in Grad School) I bought a cheap laptop so I could make a clean break and never need to use a "public" computer again. (That was just before my first iPhone, and I actually imagine that should be even easier now that so much of what the need for borrowing "public" computers as a student in those days has moved to our phones. The rest was already mostly replaced at that point with tools like SSH and RDP remoting. One useful discovery at the time was that if I was RDPing into my desktop machine back in my dorm on a "public" computer in a school lab, I'd need to hunt and peck my address and password in in Qwerty, but as soon as I connected to my own machine, I could touch type in Colemak because it was my machine and set to use Colemak as default.)

I also made peace with looking like a slow idiot and hunting and pecking out QWERTY when trying to help someone on someone else's computer. When I need typing speed I use a computer I control. I learned to stop worrying about my typing speed on other people's computers.

> if you ever look down at your keys for reference (something you want to do a lot when you're learning a new layout)

This was a feature for me rather than a bug. Proper touch typing is "never glance at the keyboard" (and I was taught that over and over from childhood with Qwerty lessons, though I was bad at it with Qwerty; including teachers that insisted lessons needed to be done with a cover over your hands and the keyboard so that you couldn't look) and if you can't trust the key caps you have more reason to never look.

> Also the unlearning process can be slower than I'd like, and my typing speed plummets while I'm learning these things, so I'm less useful at work or school or working on personal projects.

For what it is worth, what worked well for me was taking a long holiday weekend where I didn't have work or school or personal project deadlines to do just about nothing but Colemak typing lessons. Going entirely cold turkey on not using anything but Colemak when touch typing from that point onward. (I also make a distinction between touch typing and "swipe typing" as entirely separate skills despite both involving "a keyboard" and don't have a problem touch typing in Colemak and swipe typing in Qwerty.)

That long holiday weekend was enough that I felt comfortably fast with Colemak for work/school/personal projects as I started to get back into them that I never felt like falling back on my old bad Qwerty habits, even if it was still maybe a small bit slower than where I had felt like I'd been before the switch, and as with anything once committed to it and using it every day speed only improved from there until I didn't feel any speed difference at all (and presumably am faster than I ever was with Qwerty though I never trusted speed metrics enough to say for sure).

I was highly motivated though to unlearn Qwerty, so obviously your mileage will vary. My Qwerty touch typing form was wrong from years of self-taught programming (my left hand was shifted off the home row left by a key to spend more time on the accelerator keys and my right hand had to move a lot more to compensate). Late in grad school with a mountain of assignments both school and otherwise I was dealing with an immense amount of hand pain that I was sure likely meant I was deep into the early stages of carpal tunnel and buying a cheap laptop and switching to Colemak seemed like a great cost savings over surgery. More than a decade later, almost closer to a decade and a half now, it still seems like a great deal and I'm very glad that I did it.

To ease into colemak, you can yous TARMAK, transitional colemak layout (2-3 keys switched at a time).

fastest way to ramp up.

https://forum.colemak.com/topic/1858-learn-colemak-in-steps-...

I think moving j just to move it again is annoying. I don't see how this was considered worth it. Cool to see someone trying this concept of transitional layouts.

There are numerous versions of this. I’ve constructed my own with a generator and it was actually quite helpful to maintain speed.

You don't have to stop using qwerty. You basically stay fluent in both layouts, enjoying the advantages of Colemak on your personal machines and still typing quickly on other machines. For me, it's definitely been worth the effort in terms of managing RSI.

The only complaint I have had is bindings in vim. I found there was no point in trying to have a custom vimrc since I would be using vim in all sorts of situations where I couldn't have a vimrc. My hjkl was awkward to get used to, but now I don't even realize I'm pressing random buttons for up down left right.

The only thing that prevents some ergonomic keyboards from becoming mainstream is the price. Low price is the only great thing about standard keyboards.

Very skeptical of this take. If it were the case you would expect most or all rich people to be using ergonomic keyboards, since even an expensive keyboard would easily be affordable to them. Yet when I look at the folks who work at my office (none of whom make less than $200k/y), fewer than 15% have non-standard keyboards.

I think the real issue is that for most people, ergonomic keyboards do not offer a real benefit.

> I think the real issue is that for most people, ergonomic keyboards do not offer a real benefit.

Or they are not aware of the benefit until RSI strikes, which is arguably too late.

They will as soon as they get RSI. This is the other side of the problem, people is reactive, not preventive, towards RSI problems. IT needs to get the into the mindset that not using an ergonomic keyboard is a problem.

For the fraction of people who end up developing RSI, and for whom the RSI is treatable or preventable merely by switching to an ergo keyboard, it's great. However,

- not all RSI is treatable by using an ergonomic keyboard, and

- many people do not develop RSI no matter what keyboard they use.

It's not clear to me the exact fraction of people who end up developing RSI and what fraction of those cases would be prevented by ergonomic keyboards. However, I can say that most people I work with are not newbs. It seems likely that if they were ever going to get RSI, at least some of them would have done already.

That was a very long road that took the reader to right where they started.

Laptops don't fit ergonomic keyboards, but nothing else about them is ergonomic either. Otherwise, for anyone who cares, ergonomic keyboards are readily available (yes, low volume of sales plus high design expense will yield expensive products) and I've never been on an enterprise network so locked down that you couldn't swap keyboard layouts.

One tip: if you put an alternate key mapping on the shared top secret submarine laptop and leave it readily accessible, don't be surprised when someone wakes you up in the middle of the night because your time-saving macro "broke".

I've witness that. A coworker brought his ergo keyboard to a Redhat class/lab/exam. The instructor said if that thing was still on the computer when he had to configure them for the lab exams that he would fail. I imagine by now they have fully automated the lab setup so the keyboard doesn't matter but it did at one point.

How is that not an ADA violation?

Only if you can prove you have a disability.

Having an ergo keyboard doesn't mean you have a disability. For many people, it is just personal preference.

Also the instructor didn't prevent him from using his ergo keyboard, he just asked him to put back the regular keyboard after he is done with it.

This was exactly the case. The ergo keyboard was his personal preference. He did have wrist issues, but the keyboard was not prescribed. He put back the keyboard that came with the lab computer and he passed the exam.

The whole argument that something needs to be replaced because it has been around for such a long time has always struck me as a bit peculiar. Odds are that if something has stuck around for a long time, it is exactly because it works pretty well and whatever perceived benefits of doing it differently haven't panned out, or have been small relative to the hassle of changing the paradigm.

I think you're describing a fallacy known as the Appeal to Novelty:

[1] https://en.wikipedia.org/wiki/Appeal_to_novelty

I don't think this is a case of that. I honestly think it's the author just looking for battle-tested things that have been around for so long, erroneously concluding "they've been around so long, there must be potential for disruption here!", and then scouring for evidence to support this. Almost mindlessly searching for something to improve.

Also, the Lindy effect:

https://en.wikipedia.org/wiki/Lindy_effect

This, exactly.

"Ooohh ... Shiny!" is quite common, these days. It has resulted in some ... entertaining trainwrecks.

Every now and then, though, something new does work, and it then becomes ... something old, as it is mainstreamed and adopted.

Circular steering wheels are an example. At one time, they were new and radical.

We didn't need keyboards, until we needed them. Maybe, one day, we won't need them.

Until then, there's absolutely nothing stopping people from designing new types of text input devices. There's no "QWERTY Cabal," killing off alternative keyboards. If someone designs something that really works (not just in the technical realm; but also in the social, educational, manufacturing, cost-relevance, and distribution realm), then, it is likely to do well.

I think it's easy to overlook that it isn't enough that there is some benefit to changing things around, it needs to be a significant benefit to outweigh the hassle.

Dvorak probably is the better keyboard layout compared to QWERTY, but by a fairly small degree that doesn't even materialize for most typists. If it made the average person type twice as fast, nobody would be using QWERTY.

Another example is NEMA plugs. Compared to Schuko, their safety features are quite lackluster, but the cost of changing the power plug standard is extremely high and safety features of NEMA are probably good enough.

Are we talking about people, or keyboard layouts?

> Imagine a Mercedes (...) Would you drive such a car on the narrow, unsuspended wheels of the first automobile without the comfort offered by today’s technology?

I think this analogy is not accurate. Keyboards of today have been greatly revised for comfort and rely on different materials and components; it's just that the general design didn't change. I think it would be more accurate to ask whether we'd still use cars with seats, engine and four wheels; which we do.

Exactly. Sit down at an old typewriter, with its 45-degree-sloped keyboard, 2" key travel, and tendency to jam any time you press two keys at once, and tell me that's not a Model T to a modern mechanical keyboard's Mercedes.

I still occasionally use an Olivetti to write letters to family members, loved ones, and the IRS. It’s more of a 1973 Porsche Carrera compared to the Kia Soul that is my computer.

The analogy is not accurate because old mechanical keyboards are awesome.

Over hundred year old typewriter keys suck.

True, but certain 60-year-old typewriter keys are unbeatable. [Selectric, 1961]

Yes we should. I have a couple Kinesis advantages. Having the ability to have my thumbs do backspace, control and enter are fantastic. I also put symbol keys like {},[],(),`,',~ on the keypad layer that I can easily shift into (I remapped the giant delete key to be keypad shift) so my weak ass pinky fingers don't have to do as much work. For me that's something that gave me the most pain during typing. I'm not really sure what to do for laptop typing though. There's nothing like the Kinesis advantage in a laptop. I'm thinking maybe I can get a laptop with a JIS keyboard and remap the extra keys near space to shift into a layer instead. I wish manufacturers were interested in improving keyboard design and not just how many keys can they remove to save money.

After just reading about those kinesis keyboards I'm still reconciling with the fact the my middle finger is supposed to curl to press c rather than my index finger.

Ha, yeah it takes a little bit of getting used to. I pressed v instead of c a lot for the first few weeks.

Having switched to a properly split keyboard with 12 inches between the parts, yes yes and yes.

I'd like to see a reprogrammable keyboard where the letter shown on the keycap can change with reprogramming. Or with whichever layer is selected.

Shape and size wise the Kyria [1] is the best I've found for me.

1. https://splitkb.com/products/kyria-pcb-kit

> a reprogrammable keyboard where the letter shown on the keycap can change with reprogramming.

Indeed. I imagine e-ink keycaps would work well for this. Get them integrated with QMK and you've got the keyboard of the future.

Art. Lebedev Studio had a line of keyboards with OLED displays on the keys, which were reprogrammable. It's more of an design project rather than a mass produced product, but they did sell a number of these really expensive keyboards[1].

[1] https://www.artlebedev.com/optimus/

The cheaper solution would be to use an on-screen-popup. Programmable keyboards have started to implement APIs, allowing tinkering live from the OS-side. Displaying a popup with he active layout would be good enough, as people should watch to much on their keyboards anyway.

> A possible solution would be the use of small displays integrated into the keys. […] However, this idea has not been implemented yet.

Optimus sells a line of keyboards with integrated OLED keys: https://www.artlebedev.com/optimus/maximus/

Although the keyboard version of that was very elusive. almost 2000$ afaik, and I'm not sure I've ever seen any test or report of someone actually having one.

Here’s a review from 13 years ago: https://m.youtube.com/watch?v=IJ5rX6WpxTk

Apple tried to kill the function keys, a pretty modest attempt at a change, and was hit so hard they rolled back, which is very rare for them.

It's really hard to beat the current design, as obviously flawed as it is. The next thing has to be a whole lot better.

Because they didn't do it better. Even if you ignore that the software support never really got there since it only existed as a feature on expensive MacBook Pros, it just wasn't better. The lack of any feedback makes it hard to use without having to look down at it and made it really easy to trigger something when reaching for a key like delete. I've literally brought up Siri on my MBP hundreds of times by accident before I gave up and made it a space that does nothing.

Am I the only one who keeps my wrists straight while typing on a regular keyboard? Always have. You don't have to make the mistakes depicted in the article, regardless of which direction the keys are facing. And I hate "ergonomic" split keyboards, since there is an overlap between the keys I press with each hand, depending on how they're sequenced, meaning such keyboards decrease my typing performance.

Mouse distance while using keyboards with numpad is a real problem, though, and I've suffered from tennis elbow associated with mouse use.

While the shape of keyboards is not ergonomic by itself, my experience is that cheap keyboards (especially the ubiquitious rubber dome/membrane variety) are much, much, much harder on the body than a semi-decent "mechanical" board.

One reason is that rubber dome/membrane keyboards require bottoming out to register keystrokes. Most mechanical mechanisms/switches can register before the key is fully depressed.

I think the most obvious redesign would be to switch to a stenography system instead of Querty/Dvorak/etc. By chording, we could type faster with less movements. Steno takes longer to learn, but it's easier to type over 200WPM.

On the other hand, I tried using Plover for a couple months and found it was much better for writing English than it was for programming in an IDE. I missed the function keys. Plover was also a pain to use...I've been meaning to program a Raspberry Pi to run Plover but act like a normal USB or Bluetooth keyboard. I bought the Pi and a small screen, but never got further with the project. The Pi is now the piano synthesizer for my 25 year old music keyboard.

I was surprised to see the article didn't touch on chording. I have never made the leap myself, but I can see there's a lot of potential there if you're willing to commit. The whole QMK / DIY keyboard scene seems to open up possibilities, like[1]

1: https://www.paulfioravanti.com/blog/chording-qwerty-qmk-comb...

I'm waiting on a soon-to-be-shipped Kickstarter product called CharaChorder which is a regular qwerty keyboard with chords. Very excited to try it out

Anything as long as it's not a touchscreen.

> In orthogonal layouts, for example, the rows of keys are no longer offset, but are arranged in a linear grid.

The word is ortholinear

My next keyboard will be columnar/ortholinear. It just makes so much more sense compared to the staggered layout on a standard keyboard and it doesn't require you having to learn a whole new layout like dvorak.

Wow, lot of pessimism in this thread about a problem that has been solved pretty well more than once. Here's how I see it:

Computer keyboards have one major problem: they cause carpal tunnel because your fingers have to fly all around the keyboard, especially to hit the modifier keys. Changing the order of the keys is a bit like rearranging the deck chairs on the titanic. Any solution solution has to move the modifier keys onto the thumbs and palm, and move the keys closer to the fingers in 3 dimensions so that finger travel is reduced. This is a solved problem now. You have to pay 300 bucks to get the keyboard and it's limited release, but you know how economies of scale work; if enough nerds want it, poor nerds will be able to get it eventually. I use this perfect keyboard, and it gives me instant wrist relief even if my fingers are screaming already from laptop keyboard use.

This is my second of its type actually, the first was stolen :( and I had to deal with wrist pain until I could afford another one.

That said, there is probably no way to fix the laptop keyboard, due to it lying in a plane (so the laptop is flat) and being tiny (must be smaller than the screen). So that sucks.

it's a keyboardio if anyone is curious. They're doing another run soon if you want to back it. Just realized if you google the words I used in this post you get the kinesis keyboard, which is the same price and has several fewer features.

I would also be curious, I've got giant forearms from daily exercise because I never wanted carple tunnel, anyone else who exercises their hands regularly here who can testify to the efficacy of maintaining your body even with a cush job like programming. In my experience I have only gotten body aches and pain when eating too much sugar and not exercising daily. In fact when I start to feel old I do a fast and realize how much of this pain is diet.

The mouse is just as big of a problem. I honestly don't understand why there are (almost) no keyboards with a built-in mouse. Being able to perform at least some mouse movements without moving your hand(s) off the keyboard is great.

Yes, I'm aware of the Lenovo keyboard that comes with a trackpoint. I had one for several years before accidentally killing it. The mouse part is great but it's not ergonomic in any other regard which is a deal breaker.

Rollermouse and other similar bar mice solves this for me perfectly. Trackpoint gives me too much hand strain if used +7 hours a day.

Glad to see rollermouse named. Is my answer for this case, as well.

+1, I use it in front of a Maltron L90 keyboard, plus a trackball on the side.

oh snap that sounds like a nice setup.

Thanks for the suggestion. Never heard of the Rollermouse.

It doesn't seem to be a good fit for ergonomic keyboards, which is a deal breaker. But still, an interesting product.

Yeah they don't really jive with oddly shaped keyboards. Only thing I can think of that might suit your needs would be a trackball, placed in front of your keyboard between your arms.

On my ergodox ez, I mapped some keys on a layer to move my mouse using vim keys, to do a mouse scroll with the keys just under and to do left/middle/click with the keys above. it is not great, but it is functional.

I am not 100% sure, but I think that's supported by any keyboard using QMK

Keyboards and mice are nice because they require relatively small, efficient movements.

The only thing that would improve mice is to make them disappear. Replace the physical puck with some kind of image capture or accurate position sensing system that tracks all fingers for pressure and position, and you'll be able to make even smaller movements on any surface while keeping your wrist rested.

Keyboards need some physical travel, and modern designs have reduced that as far as it can go. Some people prefer Cherry-type clacker keyboards, some people prefer Apple-style micromovements, and some prefer something in between. Pick whatever works for you. There are choices.

I suspect there are flat touch keyboards in Apple and MS labs somewhere, and I also suspect they'll stay there. It's a perennial idea, but it's not happening until a technology is invented that can emulate key travel in any location(s) on a lightweight programmable display surface. That would be amazing - like the app-dependent touch bar, but both comfortable and useful.

Gestural interfaces look sexy in movies, but I really don't believe days of waving your hands and arms around and talking are going to be easier on your body than a modern keyboard+mouse mix.

I wonder is it possible to implement trackpad on keyboard keycaps.

Later BlackBerry phones had this. First on the BlackBerry Passport that ran BB10, and later on their Android phones (Priv, KeyOne, Key2). It works really well. You can throw predictions into place with an upward swipe, swipe down to open the symbols page, swipe left to delete a word, tap and pan to move the cursor, etc.

The article is kinda weird. Author is most probably suffering from mouse but blames keyboard. As an argument article against modern keyboards describes physical parameters (like distance between rows of keys) of typewriters. And the schematics of hand positioning while using keyboard - they're just unrealistic. Mostly because it's really hard to even place your hands that way without training.

"we should finally redesign the computer keyboard"

Bullshit. If you want to try, go ahead and try. But it's not like it's a bad interface. Keyboard is still the best interface for speed we have come up with. Until someone comes up with something else that is objectively better, it's not time to redesign anything.

In my personal experience, it doesn't take much to make the difference between a computer setup that causes RSI versus one that doesn't. When I was using my Amiga 500 for a number of years, I started developing carpal tunnel symptoms in my teens. Switching to a detached keyboard that could actually be placed at the right height for my posture was thankfully enough to allow my body to recover.

In the past decade the best change I made is switching to a combination of a standing desk + kneeling chair + regular office chair. By switching between different positions, my body feels a heck of a lot better. The kneeling chair is particularly effective at forcing my body into a much better posture position, and is good from recovering when the other options start to irritate things.

> And so the introduction of the mouse in the mid-1980s was the last major innovation in the computer industry.

I think he missed the whole touchscreen thing, particularly multi-touch, not to mention compass, accelerometer, etc. Plenty of software makes use of these, and is either unusable, or just cumbersome without them.

It seems like this mostly focuses on ergonomics as the next dimension of improvement, but there are several other factors that come: Manufacturability improvements could reduce the cost and make customized solutions cheaper and more accessible. Repairability improvements could extend the life of keyboards and add options for improvement over time. Recycling would be nice to consider since most keyboards quickly become a lump of mostly plastic e-waste that would be good to systematically melt down and reuse or at least dispose of safely. And perhaps most obvious with the pandemic is sanitation. Most existing keyboards are not easy to clean, may or may not survive a trip through a dishwasher, and protective covers tend to mess up the ergonomics and are not necessarily built for cleaning or long life themselves.

This is one of numerous societal examples where the relevant detail is the ubiquity of the standard rather than the optimality. Once a given set of parameters is broadly-applied, any improvement requires too high a short-term cost to justify. To wit, consider the colors of a traffic light. Perhaps [red, yellow, green] is not optimal. However, we've all generally grown to understand the pattern, and if somewhere, that pattern changed, there would be numerous accidents for a non-trivial period until folks adjusted.

The goal shouldn't be a better keyboard, but rather, a better paradigm for conveying linguistic intention to a machine (e.g. BCI), which can then be constructed in a manner more reflective of what is CURRENTLY optimal.

Microsoft Natural and its derivatives is an excellent ergonomic keyboard.

Replacing QWERTY is simply hard.

However, adjusting to a Kinesis Advantage also took me several weeks. Although the layout is the same, the keyboard is feels different enough that learning a different key layout might be possible.

Posture matters a lot in ergonomics, and in particular screen height. Raising the screen to eye level reduces neck pain. I adopted piano posture and have less strain on my limbs.

That is why I will never understand typing on a laptop open at 120-degree angle where the keyboard is angled upward and the screen is at keyboard level. That’s the worst of both worlds.

Merely raising the laptop screen to eye level and using an external keyboard makes it far more comfortable.

Like many readers of HN I work in an office (well, not at the moment), and quite often I'll end up working with someone else, and perhaps using their keyboard or vice versa. I don't think I have the mental muscle memory to switch between two keyboard layouts, from the 99% of users' flat-keyboard QWERTY layout to two isolated Dvorak keypads, or whatever.

And I don't expect other developers to carry their keyboard over to my desk, switch keyboard settings on my PC in order to pair-program either.

I suspect this article will exist in another 150 years, unless we're in Minority Report territory (with unfeasibly fatigue resistant arms...).

This is somewhat fair, but changing layouts is a simple command space in Linux and Mac. So not that hard to do.

I'm talking about two keyboards, one qwerty (theirs) and one dvorak (mine) both plugged in.

Back in the day, it's a (Windows) OS setting per session, so both keyboards would need to be the same. Is it now USB driven? Does the Mac or Windows support two different layouts simultaneously?

Fiddling around with keyboards isn't something I do (nor running Windows tbh).

This feels alien to me. Makes some sense, but sharing sessions is pretty niche. Collaboration usually involves networked access, or passing the keyboard.

Yeah, and the "passing the keyboard" was my original comment as to why I can't personally waste the time to learn a different keyboard layout to QWERTY, or even a split keyboard, when everyone else (99%) use flat rectangular qwerty keyboards. I doubt I have the muscle memory to switch between the two in an office situation, so I may as well just remain one of the 99%.

Also, I can touch type, so I don't really see how squeezing out ten or so words per minute using dvorak is going to improve my productivity.

I fully cede that it is marginal benefit at best. I did it really just to see if I could. Same reason I flip left she right hand for mouse.

Switching between keyboard layouts is very easy, and is done every day by millions of bilingual people, especially if they type languages with different scripts (English, Russian, Chinese, Arabic etc).

On Windows, pressing Ctrl+Shift (nothing else) flips between the configured layouts. Other systems have similar shortcuts.

~10 years ago I ran the gamut on keyboard layouts during a slow period at work. I discovered a few things:

- Mentally switching between layouts is (surprisingly) far easier for me than e.g. switching between mouse, trackpad, and touchscreen inputs between devices. Most of the time I don't even realize I've done it I just start typing in QWERTY on other people's computers. I don't think I'm super human I think it has more to do with the next item

- About 95% of originally "learning to type" seems to have been unrelated to the learning the keyboard layout or shape of the device. If I had to guess most of it must have been learning the motor skills of how to position your fingers over a dense input device without having to actually think, look, or focus on doing so.

- Corollary to the above, if you learned incorrectly (e.g. you're reaching between sides of the keyboard and you go to use a split keyboard where you can't do that) relearning that alone might take as much effort as learning a new layout.

- My hands stopped hurting even if I spent the day typing on a laptop keyboard. This was definitely the biggest gain for me.

- The quickest way to pick up a new layout was to just cut over cold turkey having a layout reference handy for only the first ~30 minutes (otherwise you spend more time looking at the reference because you're only 90% sure instead of pressing the key and becoming more sure). If you honestly do a hard cut and make an effort to type as long as you normally would have you can be quite proficient again within 1 week and just as good as you were QWERTY within 2 weeks. For reference when I started I thought it might be a multi month kind of process to get to parity.

- Spending a month or two on a non QWERTY layout I was a better typist (both accuracy and speed) on the new layout but the gains in this regard were not much more impressive than spending a month or two practicing typing. I.e. getting to the same speed will happen quick but not because eventually you'll type at twice the speed just from a layout change or something.

- You will never "forget" the commonplace layout (in my case QWERTY). Apart from other people's computers your BIOS, VMs not configured for you, some types of remote sessions that use the server's layout not the client's, certain poorly coded games, and probably some others I'm forgetting will assume you are on QWERTY. Thankfully because of the above this is not a problem, the hard part is typing 5 characters you have to delete once you realize it's QWERTY.

- Corollary to the above: there are keyboards that allow you to set the layout in hardware fixing things like BIOS or remote sessions. It's great outside the game use case where poorly coded games scan for letter but it's the physical positions of the keys that matter for the user resulting in e.g. your wasd being all over the keyboard. Some have a switch but it's usually not easily flipped like a software layout hotkey.

- Other layouts were a bit like a bike, I forgot them after a few months but going back to them was a matter of days not weeks.

- As far as noticing benefits it didn't really matter which layout I used as long as it isn't QWERTY.

These days I really only use Carpalx and QWERTY and I honestly would have been happy with <anything I tried> and <QWERTY>. I also stopped caring about special keyboard hardware as it just seemed like unless I was going to carry my keyboard around with my laptop bag I wasn't really spending much time with a keyboard I could pick anyways (and ones that fit in laptop bags weren't all that great).

I can touch type at a fairly high WPM, so I don't really think I'd benefit from dvorak or any other type of keyboard.

I am interested in a split keyboard, but again, I don't have any joint pain or similar that would possibly benefit from that.

Typing with the "wrong" finger because the key is supposed to be typed with the other hand is precisely why I have no desire to use a split keyboard. Let me double up the border keys, please!

I don't see anything particularly broken with the keyboard. Much as the design of a knife is a convergent, functional end-design of a utility, I'd say that keyboards are the same.

As with knives you could make variants for specific purposes (fruit knife, chopping knife, serrated knives) but that doesn't mean that we have to 'redesign' knives altogether, or in the same way, keyboards.

Luckily there is an easy fix. But It's nice to see someone who cares about others being upvoted on hn, a rare thing. Get dumbell, hold in your hand at your side and let it roll down your fingers, now curl your fingers up and lift the dumbell with your fingers. Do this until it burns. You will never have carple tunnel now. I've done this since I was a teen, and never developed hand issues. Been typing daily since 8 years old.

>Get dumbell, hold in your hand at your side and let it roll down your fingers, now curl your fingers up and lift the dumbell with your fingers.

at your side facing which direction? may be it is a perfect description and it seems it is but for more exact interpretation is there some video demonstrating it?

Or may be you can describe it with more-then-required details just to make sure that one who will try to repeat it would not do any harm while thinking he is helping himself.

I found using a Theraband Flexbar helpful for my medial epicondilitis (golfer's elbow). It's a thick column of rubbery material which you twist in various ways for resistance exercises. You can also do exercises which mimic the effect of the dumbbell exercise which the parent commenter described.

https://www.physioparts.co.uk/thera-band-flexbar

I have the green one seems to be about right for my needs.

I get all kinds of misspellings, only one mobile, due to uio keys right next to each other.

I get one variety of misspellings on Android using swipe (the word is wrong w.g. I type due and it guesses sure) And another variety of wrong vowel. And autocorrect messes these up in different ways too.

So while the keyboard layout is a factor, another is language, and input method (type or swipe) and probably user idiosyncrasies, etc

To attenuate my hand pain issues, I am using a mix of a tenkeyless 60% keyboard, AutoHotkey, espanso and dictation. Also, instead of focusing on using key combinations and typing fast, I am trying to practice 'typing mindfulness'; that is, really to be present and type consciously, accurately and slowly, instead of fast and erratically.

Still amazes me how terribly un-ergonomic most keyboards are.

I had RSI, but recovered, in part because of a split keyboard (the other two components were a pen-like mouse and a pause program).

https://henrikwarne.com/2012/02/18/how-i-beat-rsi/

I guess it is kinda similar as:

"After 150 years, we should finally redesign the transportation system". I mean the writing is on the wall - public transportation is the future. Electric, sustainable public transportation.

“Without deviation from the norm, progress is not possible.” (Frank Zappa, musician)

Now let's communicate the idea to all the public starting with car manufacturers.

While speech is mentioned, it's role of dominant input method in the future is understated.

I've considered going to ergonomics and/or dvorak and such but is it really going to help? I need a better desk, better chair, better posture.

You know what we need to do instead? Brain computer interface. Remove the whole keyboard and mouse altogether.

Can't believe no one posted that: http://kanji.zinbun.kyoto-u.ac.jp/~yasuoka/publications/PreQ...

QWERTY keyboard is widely used for information processing nowadays in Japan, United States, and other countries. And the most frequently asked question about the keyboard is: "Why are the letters of the keyboard arranged the way they are?" Several papers in the field of information processing answer the question like this: "To slow down the operator." It's nonsense. In this paper we reveal the prehistory of QWERTY keyboard along the history of telegraph apparatus: Morse, Hughes-Phelps, and Teletype. The early keyboard of Type-Writer was derived from Hughes-Phelps Printing Telegraph, and it was developed for Morse receivers. The keyboard arrangement very often changed during the development, and accidentally grew into QWERTY among the different requirements. QWERTY was adopted by Teletype in the 1910's, and Teletype was widely used as a computer terminal later.

All we really need is to eliminate Caps Lock as a key and as a function.

Failing that, could all Linux distributions please default to mapping that key to Ctrl, instead? Most especially, in "live OS" images?

Thank you. The whole wide world thanks you.

Why? I use Caps Lock all the time when typing with one hand, although the location could realistically be somewhere else with no ill effect. But the location should be left up to the hardware maker, not the OS.

Instead of yet again change the same low-influence gears, people would better start learning proper usage and positioning. How you type and sit has far more influence on your health than the layout ones uses.

> One of the main problems of today’s keyboard designs — including alternative layouts — is the fixed key labeling.

In linux you can redefine the keys.

Well it's not about just redefining the key input mapping, it's about the visual physical labeling on keys. There are (or were) some keyboards with small LCDs under each key on the market where you can dynamically remap the visual presentation on the keys themselves as well (or even show small images) without moving to a virtual keyboard.

I've been using keyboard for many decades and I still occasionally look down and visually map out a key or fingering position. The visual mapping needs to be correct, I don't want to play a mental mapping game that 'q' should be 'z' or whatever.

Or, just get a set of keycaps with no labels.

There are parallels here between railroad gauge and Roman cart wheels.

We also need to get rid of caps lock.

Why?

We need to get rid of caps lock

My first task with a new system is to remap capslock. I like having it replace ijkl with arrows, but my muscle memory is now so dependent on this convenience I am impeded when I switch back to a vanilla keyboard.

I picked up remapping Caps Lock to Backspace as a far more convenient Backspace. We all make plenty of mistakes when typing and no need to walk a mile in the snow to the Backspace we have. Plus, it's not terribly annoying when it doesn't work as remapped because you still have the other backspace.

(On Windows, I'd been using SharpKeys for years to do the remap, but I appreciate the Keyboard Manager built into the PowerToys these days since I'm going to use FancyZones anyway.)

Reply on news.ycombinator.com