It’s awesome just how fast and accurate they can be, and most devs were of my mindset “wow can I learn to type like that - it woukd solve this problem and that”
Till we found out just how much work is needed to get good. It’s a true skill, and sadly undervalued but something that just has too little pro for the cons - in my opinion as a developer.
I already type at faster than I can code, and slightly slower than I can write English. A better keyboard, or the same keyboard at different workstations and laptops, or some typing tutorials woukd help me - but full on 100wpm is not going to help me debug Kerberos failures
Somewhat hilariously, having perused your first two links, and followed the link from Sweep to Ferris, I still find myself having no idea how any of this works. (But I know a lot about how the firmware and PCB layout is designed!).
IANACR, but I would think it's not practical, in that the stenotype keyboard (the official term for the keyboard used by court reporters) is used to record the phonetic _sounds_ of what is being spoken, rather than the actual words. These phonetic codes do not resemble anything approximating the actual words they represent.
Also, computer source code (whatever the language) typically contains variable names which often are (a) typically case-sensitive, and (b) abbreviated or even single characters. And even the basic syntax of the chosen language may not be easily capturable via phonetic sounds, what with open and closing parentheses, curly braces, square brackets, etc., and compound reserved words with prefixes (such as #foreach in Velocity template language).
Again, IANACR, but I don't see how it could possibly work...
I wonder how far programming-specific chorded keyboards could get. Like you usually only have a handful of variables in a function or method. I bet, at least, all the extra context provided by object oriented languages could be used to help the keyboard provide us meaningful suggestions.
I seem to remember reading once that what is typed by a stenographer is also only meaningful to them. And therefore must be transcribed into English later by the same stenographer.
That used to be true but it's much more automatic now that the process involves a computer translating into English in real-time (it's no longer a stenographer translating from their notes by hand).
This. Using QUERTY immediately feels uncomfortable when i have to use it. Learned NEO2 which has layers accessed with modifier keys. Having a numpad under your hand is one of its' many advantages.
A lot of the people customizing layouts for the Twiddler chording keyboard (https://www.tekgear.com/keyboards.html) are programmers, and are building custom layouts that reflect this.
Depends on your goal. Chording technique is superior when typing words contained in the dictionary. Meaning that typing some rarely used word required typing it multiple times to "confirm".
Writing code does not suite well for this, since coding with completion contains much more punctuation than plain text.
Instead, check out ergonomic mechanical keyboards: low-profile, split, with columnar stagger, preferrably with 36 or less keys. Uncommon keys are behind a modifier key that acts as a normal key when pressed, but as a layer when held (called modtap).
Also you can experiment with non-qwerty layouts, but IME it gives much less benefit than having a layered layout of physical keys.
Non-qwerty seems to me to be one of the biggest wastes of time we’ve come up with in search of productivity. The time spent learning it could be spent learning vi, learning Haskell, learning to shoot hoops or learning the guitar. Pretty much all of those would benefit you more.
I don't think that's true. You've probably had moments when you had the entire design of a program in your head, and it required typing out hundreds of lines, or even thousands, all of them mostly conforming to the original idea.
Sometimes, and more so in certain languages and the way they are used, massive amounts of boiler plate are required. Programmers use copy and paste techniques to do this, which are basically devices for massively amplifying typing speed and accuracy. When you copy 50 lines, and then change five places in the copy, that's faster than typing them from scratch. The editing commands are a kind of shorthand, which gets expanded in the editing buffer.
This doesn’t really answer your question but I was on Jury Duty recently and was disappointed to learn that the stenographer was using a normal QWERTY keyboard and boring Dell computer. It seems that they used special software however that was connected to an audio feed with some recall ability.
That being said there is some [support for stenography in the QMK programmable keyboard firmware](https://docs.qmk.fm/features/stenography). I’m not sure how widespread its use is.
Humor on HN is generally accepted if it's clever or makes you think in some way. But "haha some words have multiple meanings!" doesn't quite meet that bar.
It’s awesome just how fast and accurate they can be, and most devs were of my mindset “wow can I learn to type like that - it woukd solve this problem and that”
Till we found out just how much work is needed to get good. It’s a true skill, and sadly undervalued but something that just has too little pro for the cons - in my opinion as a developer.
I already type at faster than I can code, and slightly slower than I can write English. A better keyboard, or the same keyboard at different workstations and laptops, or some typing tutorials woukd help me - but full on 100wpm is not going to help me debug Kerberos failures
reply