Workman Keyboard Layout (2010)

I hear you. I did it. But am I supposed to carry it around with me in a special case for all occasions?

The short of it is that in the end, I stepped down from my horse and conceded that it was more sensible just to make peace with QWERTY and get good at QWERTY'ing. This won't be the case for everyone, I'm just trying to shed light on my experience as "the keyboard guy". Colemak worked for me. The ubiquity of QWERTY worked against me. In the end I capitulated because going against the grain wasn't worth it for me all things considered.

> What do you mean, that Apple should sell Colemak keyboards? I agree there, but it's a bit of a chicken and egg situation.

You're right, it's chicken and egg. But Apple have a history of making a bold stand around product features.

Phones without keypads, phones without audio jacks, laptops without CDROM, laptops without USB-A.

If someone is able to ween the world off it's QWERTY habit - it's Apple. It doesn't need to be Workman or nothing. It could be as simple as offering Workman as an alternative replacement, signalling a shift toward Workman as the future. Then one day, probably before everyone is ready, no more QWERTY MacBook. Baptism by fire is the only way, in the end.

I think this is an example of exceptions looming large in our mind, precisely because they are exceptions, and taking an outsized proportion of our attention. Before you go crazy remapping your keyboard to make symbols easy, take a scan of your real code. I fired a script to count the incidence of all symbols across a couple codebases large enough to be fairly representative here:

    Perl                   Java            Go
    1 . spc - %26.40       spc - %17.98    spc - %10.43
    2 .   e - %6.15        e - %7.93       e - %8.92
    3 .   a - %4.41        t - %6.76       r - %6.83
    4 .   t - %4.34        r - %5.68       t - %6.72
    5 .   r - %3.85        s - %5.34       s - %4.99
    6 .   s - %3.84        a - %5.19       i - %4.83
    7 .   i - %3.34        i - %4.96       n - %4.67
    8 .   o - %3.03        o - %3.95       o - %4.57
    9 .  \n - %2.99        n - %3.79       a - %4.41
    10.   n - %2.97        p - %3.37      \t - %3.60
    11.   l - %2.67        u - %2.91      \n - %3.44
    12.   d - %2.32        c - %2.79       c - %3.05
    13.   c - %2.25        d - %2.55       l - %2.73
    14.   u - %2.07        l - %2.50       u - %2.55
    15.   _ - %2.02       \n - %2.44       d - %2.32
    16.   $ - %1.70        . - %1.99       p - %2.08
    17.   m - %1.70        m - %1.94       g - %1.98
    18.   p - %1.55        g - %1.62       . - %1.81
    19.   f - %1.41        ) - %1.29       m - %1.65
    20.   h - %1.31        ( - %1.28       b - %1.62

scp is space. Even Perl, a notoriously symbol-heavy language, has its first "symbol" show up only slot 15. This codebase prefers separation_by_underscores, so if you have camelCase you might not even get that. Java's first shows up as the period in 16, go's slots in at 18. (Go programmers typically aren't using the TAB key that much, it's autoformatted into existence with gofmt.)

I'm not saying there's no room for a bit of optimization, but I suspect a lot of programmers would mentally place symbols way higher than they actually belong in the precedence scale.

The other problem that emerges if you get too serious with this is that the languages are different enough that if you do swap out some symbols for some more privileged spaces on your keyboard, they aren't the same ones across languages. In Perl, $ shows up in slot 16 at 1.7%. In Java, it's slot 66 at .01%. Java favors the period first in slot 16 at 2%, in Perl that's slot 40 at .37%. I once did a reasonable attempt at a multi-language survey with roughly equal amounts of each language and what I found was if you're multilingual you have only marginal improvements available with the symbols.

Whether it's worth it is up to you, but the fruit here is much less low-hanging than you might think.

(I don't have a J/K codebase here to check.)

I had the same experience. After trying Dvorak and Colemak and even stenography for a while I decided it was time to explore more hardware.

Went from mouse → trackpad → trackball. I tried fingerballs and thumbballs and settled on the Logitech MX Ergo (thumbball). Using the digit with the strongest muscles makes a lot of sense to me, although it helps to have bigger hands.

I wouldn't go back to a trackpad, even with the loss of gestures (I just use keyboard/trackball shortcuts instead), and I wouldn't go back to a mouse — pushing a rock repeatedly to move a pointer around the screen just feels weird to me now, and I notice wrist cramp within an hour.

Similar story with keyboards. I went Apple Magic Wireless → Ergodox EZ → Microsoft Surface Ergo (the default angle works great for me so I don't have to adjust anything and I like the single piece of hardware over the Ergodox; it's easier to move between rooms). I have the Keyboardio Atreus on order and am looking forward to comparing that. It seems like the perfect balance between size, features, hackability, and ergonomics to me.

https://www.kickstarter.com/projects/keyboardio/atreus

A good example is the space cadet shift. If I tap left shift, I get an open parenthesis. If I hold it and use any other key, I get a capital letter. It's a lot nicer than hitting shift+9/0 and I got used to it in about 10 minutes.

Any keyboard using the QMK keyboard firmware can do this.