I have been considering an ergonomic keyboard for a while and came pretty close to getting a sofle [0]. But I realised that I actually like having all the symbol keys. So now thinking about trying a keychron with alice layout [1]
Interesting. Back in the '80s, the ergonomic keyboard [0] designed as part of the Japanese TRON project was originally supposed to be produced in three different key pitches for differently sized hands, based on measurements of Japanese people [1].
Another aspect was that some key columns were splayed, being more apart further away.
It has inspired many follow-ups in the DIY ergo-mech scene in more recent years.
Years ago a chip supplier suggested using their RTOS. We asked what it was called. They nearly shouted, "Industrial standard Micro Industrial The Real-time Operating system Nucleus, haven't you heard of it?" This would begin a year of Google translation of Japanese RTOS docs.
Neat keyboard. Also pretty cool how far the TRON project went and how many parts of computing it influenced.
I designed my own for similar reasons, and it's interesting to see a somewhat similar position for the thumb keys. I optimised for hand/wrist position with minimal movement. However, I messed up the column stagger (off by one). Here's the writeup:
Interesting to see the thumb key positions are very similar to mine! Can you reach the outermost (outer regarding one half) thumb key without any issues?
Is your hand position like two columns for the middle finger instead of the pointer finger?
I've yet to wrap my mind around two things that are extremely common on "ergonomic" keyboards: ortholinear keys, and the positions of the C, X, and Z keys in the columns.
I type those with my index, middle, and ring fingers, respectively, because when I pull my fingers back toward my wrists with my hands sitting at a neutral angle, those are the keys they pass. Moving from A to Z on a regular keyboard would be a disaster of a hand position!
Who is it who's teaching people to type that way on standard keyboards? I'm an elder Millennial, so the first class in my school that ever practiced typing in kindergarten, in the very early '90s. When and why did anyone switch to a horrible hand position?
It's only from that horrible hand position that I can imagine "keys in a straight line" being an improvement. My index fingers cover more than one column of keys, so staggering them makes all of them easier to reach, rather than one set really easy and the other set much more awkward. My fingers do not move in straight lines from coiled to uncoiled, and I doubt other people's do either. They splay as they extend. They should be able to cover more keys with less movement extended than they do coiled, so putting keys in straight lines makes it worse, not better.
> I type those with my index, middle, and ring fingers
I have never heard of anyone typing like this. I type those keys with my middle, ring, and pinky finger respectively.
A quick google search of “qwerty finger map” shows that middle-ring-pinky seems to be the standard. I don’t see any disagreement across any of the results. I suspect that your typing class was the exception, or you just picked up doing it differently than what was taught. I don’t think there was a large scale “switch” on how things are taught.
But yes, if you type differently than that then ortholinear keys would probably require you to make adjustments.
I was taught touch typing with pinky on the Z and index on the V for the bottom row. I think that's how it's designed to work, and that in this context stagger is a historical accident that ergonomic keyboards fix.
I agree, but would suggest that the solution is splayed columns rather than staggered rows. Lots of keyboards do this, even more “normal” keyboards that are far from being Dactyl-like: https://github.com/GEIGEIGEIST/TOTEM
I'm currently designing a board from actual traces of my finger movements on paper (I was mildly amused when I found that they are not symmetrical, like all human features) in combination with these: https://fkcaps.com/keycaps/hex. It's also my first attempt at using a MCU directly (kinda, MDB50Q).
Probably not possible to make a product from my anthropometrics, but I could try average out a few hands.
Thanks for the interest, it's on the back burner at the moment because I'm currently rage-coding a text editor :). Here's the repo regardless https://codeberg.org/jcdickinson/hexkbd
Interesting to me you say this, as my main criticism of most ergo / split keyboards is they don't have enough keys on the thumb. I started with a Maltron [1] with 8 keys on the thumb and that seems a good arrangement to me.
I have the moon lander and have settled to almost never using anything on the thumb cluster. It just didn’t feel very natural to me to have to reach for the thumb. I have larger hands too. I think what I do have mapped is only the first two keys too.
My Mitosis layout has 8-button thumb clusters but only three on each hand are used frequently as thumb keys. On the other hand, they're set up in my layout so two on each side can be held simultaneously for shift+num layer.
The rest I just use for miscellaneous other uses, like pgup/pgdn and inverted-T arrow keys. It's not for use with the thumb, but it's just a handy place to put keys you'd have to move your hand for anyway.
I started my ergo-keyboard journey with a corne. I ended up having issues with the thumb cluster positioning and built my own prototype on new years eve this/last year. That board features four thumb keys, six columns, with four rows for index and middle finger and five for ring finger and pinky (essentially a 6x5 board with four of the bottom keys moved into a thumb cluster. A lot of keys ended up unused on that design (the lowest pinkey and ring finger keys were completely useless).
A couple of months back I built another prototype, this time with a 6x4 + 4 layout. I'm still not loving the layout, four keys is too much for the thumb cluster and I'm looking at dropping down to three. Ironically I've already setup my layout so it almost only uses 6x3 + 3 (I use the top row for non-typing keys like F1-F12), and in effect reinvented the Corne that I started out with...
MacBook keyboards have a similar modifier layout to a Space Cadet in some ways: On your left hand, there’s {Cmd, Opt, Ctrl, Fn} in a row. Emacs on MacOS can use all four modifiers, e.g. mapping them to the historical {Ctrl, Meta, Super, Hyper} keys that were in the same place on a Space Cadet.
Note that this is slightly different from PC keyboards. On Mac, Fn as a modifier is handled by the OS and can be remapped (in System Settings or by individual apps like Emacs). On most PC keyboards, Fn causes different signals to be sent to the OS, making it very hard to use it for other purposes than intended.
Given the thumbs limited movement, I came to a conclusion that 2 keys are comfortably reachable, but to reach the third thumb key, the thumb needs to get into a weird and uncomfortable position. That was my problem with the corne I had and I just stopped using the innermost thumb key because of that. Do you have a similar experience?
That is why I have those for keys I rarely use, such as 'esc' and 'del'. Relatively rare - as they are obviously used frequently enough but not really when typing.
And then I hit them with my middle finger instead, or rather smash them as it requires the whole hand to move slightly.
So in my case does not affect speed or comfort (not a frequent VIM user etc)
But if I only had two thumb buttons I would be ok.
Thumb tuck is awful, and is one of my biggest complaints with Corne-style keyboards. I’m building a Dasbob in the hopes that the thumb arc is more comfortable, but we’ll see.
Have you tried or were inspired by The Voyager? My friend has a Moonlander and we were discussing that the Voyager might be the ideal for the setup we were looking for.
I haven't tried and wasn't aware of the Voyager at the time I started this project. Looking at it now, I think I would come to the same issue of the switches being too close to each other.
My first split was a corne (I have a feeling that many people started with a corne here) and then I wanted a draculad[0], but it wasn't available for almost a year. These two keyboards had given the most of the inspiration for the YetiS.
Neat! I’m surprised to see no splay, is it just subtle or have you decided against it? The thumbs look comfy and aren’t tucked under the palm, which is my biggest complaint with most keyboards, but I’m using Miryoku (36 key layout) and I’m not sure that I’d know what to do with a 34-key layout.
(Miryoku has a kludge to let you press two thumb keys at once to emulate the third, for 34-key keyboards, but that works best with low-travel switches very close together.)
How do you like the lateral pinky position? I’m currently building a Dasbob and love the pinky splay/stagger, but I’m curious about moving the top pinky key to the right.
Good point for the splay. For some reason I didn't consider it, but looking at it rn I could put a some of the keys in an angle which would feel better. I'll consider it for the next release (or a different version)!
I have nothing against 36 keys layout, the only reason for eliminating 2 thumb keys is that I can't find a proper place for the third thumb key where it would be in a nice position. I could add one key at the inner side, but the thumb just goes below the palm and it is not a comfortable position (for me at least). If you have any suggestions, please let me know! There is always space for improvement.
The lateral pinky position wasn't on the first design, but then someone with a shorter pinky who couldn't comfortably reach the top key tried it and suggested to place it there. I tried it and actually felt a lot better. (to reach the top pinky key I need to make a wrist movement, but not like this). I love it.
Interesting, I have found the standard key spacing too far apart, I cannot fathom it being too close for some. I guess hand size variance is bigger than I think.
Could you explain what you did to achieve "diodeless" design and how do you prevent masking?
> Could you explain what you did to achieve "diodeless" design and how do you prevent masking?
I was not precise with the "diodeless" definition - the keys are directly connected to the processor pins without using a matrix.
(Therefore there are no side effects of the matrix design - ghosting, masking, key rollover. The key presses are more precise and
for me they seem more responsive - less latency, but unfortunately I have no data to prove this statement.)
Thanks for the question, I will fix the definition!
Also eliminating the diodes reduces the cost and the time it takes to build the keyboard, which is always nice IMO!
Having implemented a custom matrix in qmk, I can say the timescale involved is not something you'd notice a difference between diodes vs diodeless. On ARM boards there's a setup delay to let the gpio propagate before reading the columns, which by default is set to 0.25us, which means reading 4 rows would take about 1us.
I built a macropad that has some regular keys and a SNES controller on the "same matrix", total scan time is 335us, which is dead slow as far as scanning goes, but nothing you'd be able to notice.
Other points are fair though, fewer parts in the BOM. Though routing is a bit trickier with diodeless, especially if you also want to adressable RGB.
I could believe you feel less latency, but suspect that’s more a symptom of matrix scanning working for the general case rather than being optimised for your layout. Scan with an RP2040’s PIO [1] and after debouncing I’d challenge you to measure the difference, much less notice it.
(Of course if you’ve got the pins, you should use ‘em!)
34 keys are usually diodeless as many microcontrollers (promicro, elite-c and etc.) have enough connections to connect everything directly on one side of split keyboard and when you need one connection to communicate between sides. RP2040 can accommodate even more keys (even RP2040-zero). You need two microcontrollers of course.
[0] https://josefadamcik.github.io/SofleKeyboard/
[1] https://www.keychron.uk/products/keychron-k15-pro-alice-layo...
Another aspect was that some key columns were splayed, being more apart further away. It has inspired many follow-ups in the DIY ergo-mech scene in more recent years.
[0] http://xahlee.info/kbd/TRON_keyboard.html
[1] https://www.semanticscholar.org/paper/An-Implementation-of-t...
Neat keyboard. Also pretty cool how far the TRON project went and how many parts of computing it influenced.
https://qubyte.codes/blog/why-did-i-create-a-keyboard
Interesting to see the thumb key positions are very similar to mine! Can you reach the outermost (outer regarding one half) thumb key without any issues?
Is your hand position like two columns for the middle finger instead of the pointer finger?
I type those with my index, middle, and ring fingers, respectively, because when I pull my fingers back toward my wrists with my hands sitting at a neutral angle, those are the keys they pass. Moving from A to Z on a regular keyboard would be a disaster of a hand position!
Who is it who's teaching people to type that way on standard keyboards? I'm an elder Millennial, so the first class in my school that ever practiced typing in kindergarten, in the very early '90s. When and why did anyone switch to a horrible hand position?
It's only from that horrible hand position that I can imagine "keys in a straight line" being an improvement. My index fingers cover more than one column of keys, so staggering them makes all of them easier to reach, rather than one set really easy and the other set much more awkward. My fingers do not move in straight lines from coiled to uncoiled, and I doubt other people's do either. They splay as they extend. They should be able to cover more keys with less movement extended than they do coiled, so putting keys in straight lines makes it worse, not better.
I have never heard of anyone typing like this. I type those keys with my middle, ring, and pinky finger respectively.
A quick google search of “qwerty finger map” shows that middle-ring-pinky seems to be the standard. I don’t see any disagreement across any of the results. I suspect that your typing class was the exception, or you just picked up doing it differently than what was taught. I don’t think there was a large scale “switch” on how things are taught.
But yes, if you type differently than that then ortholinear keys would probably require you to make adjustments.
Also relevant, but for improving the ergonomics of “normal” keyboards with staggered rows: angle mod. https://colemakmods.github.io/ergonomic-mods/angle.html
Probably not possible to make a product from my anthropometrics, but I could try average out a few hands.
Do you have the project somewhere online to share - git, blog, reddit, etc.? I would like to have a follow up on your progress!
(Note: this breaks down a bit once you move to concave key wells)
[1]: https://www.maltron.com/store/p11/Maltron_L89_dual_hand_full...
The rest I just use for miscellaneous other uses, like pgup/pgdn and inverted-T arrow keys. It's not for use with the thumb, but it's just a handy place to put keys you'd have to move your hand for anyway.
A couple of months back I built another prototype, this time with a 6x4 + 4 layout. I'm still not loving the layout, four keys is too much for the thumb cluster and I'm looking at dropping down to three. Ironically I've already setup my layout so it almost only uses 6x3 + 3 (I use the top row for non-typing keys like F1-F12), and in effect reinvented the Corne that I started out with...
And then I hit them with my middle finger instead, or rather smash them as it requires the whole hand to move slightly.
So in my case does not affect speed or comfort (not a frequent VIM user etc)
But if I only had two thumb buttons I would be ok.
https://www.zsa.io/voyager
My first split was a corne (I have a feeling that many people started with a corne here) and then I wanted a draculad[0], but it wasn't available for almost a year. These two keyboards had given the most of the inspiration for the YetiS.
[0] - https://keebd.com/products/draculad-keyboard-kit
(Miryoku has a kludge to let you press two thumb keys at once to emulate the third, for 34-key keyboards, but that works best with low-travel switches very close together.)
How do you like the lateral pinky position? I’m currently building a Dasbob and love the pinky splay/stagger, but I’m curious about moving the top pinky key to the right.
Good point for the splay. For some reason I didn't consider it, but looking at it rn I could put a some of the keys in an angle which would feel better. I'll consider it for the next release (or a different version)!
I have nothing against 36 keys layout, the only reason for eliminating 2 thumb keys is that I can't find a proper place for the third thumb key where it would be in a nice position. I could add one key at the inner side, but the thumb just goes below the palm and it is not a comfortable position (for me at least). If you have any suggestions, please let me know! There is always space for improvement.
The lateral pinky position wasn't on the first design, but then someone with a shorter pinky who couldn't comfortably reach the top key tried it and suggested to place it there. I tried it and actually felt a lot better. (to reach the top pinky key I need to make a wrist movement, but not like this). I love it.
Could you explain what you did to achieve "diodeless" design and how do you prevent masking?
I was not precise with the "diodeless" definition - the keys are directly connected to the processor pins without using a matrix. (Therefore there are no side effects of the matrix design - ghosting, masking, key rollover. The key presses are more precise and for me they seem more responsive - less latency, but unfortunately I have no data to prove this statement.) Thanks for the question, I will fix the definition!
Also eliminating the diodes reduces the cost and the time it takes to build the keyboard, which is always nice IMO!
I built a macropad that has some regular keys and a SNES controller on the "same matrix", total scan time is 335us, which is dead slow as far as scanning goes, but nothing you'd be able to notice.
Other points are fair though, fewer parts in the BOM. Though routing is a bit trickier with diodeless, especially if you also want to adressable RGB.
(Of course if you’ve got the pins, you should use ‘em!)
1. https://github.com/isoxliis/firmware-micropython/blob/72687e...
https://www.moergo.com
Switched from Kinesis Advantage after 15 years. Love the Pro Red config that's available out of box without expensive customization.