Hacker News new | past | comments | ask | show | jobs | submit login
Keyboard lets people type so fast it’s banned from typing competitions (vice.com)
630 points by zdw on Jan 6, 2022 | hide | past | favorite | 467 comments



The "500 WPM" video is a little suspect. He types literally one memorized sentence and extrapolates a WPM from that one 10 word sentence. I'd like to see him actually type 500 meaningful words in under 60 seconds.

Anyway, this is cool but basically just a modern stenography device. Steno has a learning curve problem that makes it so most people won't use it.


> Monkeytpe won’t let Keen run up the leaderboard, but he did win another typing competition using the CharaChorder to hit 267 WPM with 76% accuracy.

IMHO, 76% accuracy is honestly low. If the person has to sacrifice so much accuracy to hit 267 I'm not sure it's worth it. For most typing, making a mistake actually interrupts your thought process. This speed is only achievable/useful when there are text on the screen that you blindly retype.

This reads like a PR article.


I type for fun and 76% is awful. If I dip below 85% I instantly stop. 76% means I have super cold fingers while having butter slathered all over the keyboard!


That's why you should never watch Bear Grylls, Gordon Ramsay, and a Sandi Metz video at the same time.


Possibly best HN comment ever


“CharaChorder’s internal processor arranges the letters on-screen in real time faster than the human eye can perceive.”

So, 76% may also be an artifact due to the way the keyboard interprets inputs rather than something a user would need to correct.


I own a CharaChorder and this is exactly right. Using the CC, you can't get 100% on monkeytype because when I type `helo` it then backspaces that and types `hello` using the CharaChorder's onboard firmware.

The final text is 100% accurate--it's just what was typed by the keyboard wasn't 100% accurate.


What if you're debugging SMTP and you actually meant HELO?


Our keyboard gives users a sense of pride and accomplishment when they type at 200 WPM. User testing has revealed that most of our customers can't spell. We are proud to have made autocomplete an integrated hardware solution on our entire line of products.


As explained downthread it's not misspelling, they're using a chord.


The device allows character entry--you just type h, e, l, o just like you would on a keyboard.

You would know if it was a chord or not--not every word you type will be one.


Outlook, word and I think even google keyboard does the same thing. You have to basically type it again in them and they take it to mean you don't want it autocorrected.


It is the first thing you disable though.


Exactly, I disabled autocorrect early on and never looked back. For convenience I MAY sometimes use the suggested words.

But I prefer a regular keyboard. This doesn't make me a luddite or anything, mobile keyboards are objectively bad compared to regular keyboards.


So is it autocorrecting helo to hello?

Predictive tweaks aren't good in a keyboard, you want everything to be intentional, or you'll lose time and patience correcting things. Spellcheck should be post hoc imho.


>Predictive tweaks aren't good in a keyboard, you want everything to be intentional, or you'll lose time and patience correcting things. Spellcheck should be post hoc imho.

Just depends on what kind of keyboard you're using and why. This is a very specialized keyboard for specialized tasks. I could see lots of different kinds of activities and roles it could be extremely useful for.

For programming, you probably wouldn't want it... though maybe some percentage of programmers who feel their brain moves a lot faster than their fingers might like it. Combined with stuff like Tabnine and Copilot, you hypothetically might eventually be able to "predictive-program" at an incredibly high speed. You'd inevitably have to spend some time revising bad predictions afterwards, but the total amount of time spent developing could decrease by a lot.

I don't even care if this is an ad. If it is, it worked and is the first and only ad I've appreciated. I've never impulse bought anything, but I'm buying this thing right now* after spending 10 seconds looking at it. (I used to spend a lot of time on online typing competition sites, so I'm closer to the target demographic.) My only concern is how much it might affect my normal typing muscle memory.

*As soon as it becomes available. Seems to be sold out, unsurprisingly.


No--the chord for hello is h+e+l+o--if you mean to type helo, you'd use a different chord you have setup for that or type it letter by letter


Ahh, having the option for single entry in combination with word chording makes a difference. I think I'm going to give it a go, this product looks interesting.


I tested that by turning off Apple's autocorrect on my Mac. It turns out I'm a less accurate typist than I thought. Now, I'm happy to let autocorrect put right all "teh" mistakes I make. It's annoying when I want to type something unusual, but that happens less often than I make mistakes, so the balance is in autocorrect's favour.


My goodness, I turned that off under a minute after I discovered that the "feature" existed. Autocorrect is for onscreen keyboards. I am a far more accurate typist than my computer gives me credit for.


It's a nightmare if you're writing in two different languages at the same time, eg. French sentence with English words in it. I often have to disable one or the other and live with most words being red underlined, because autocorrects are just not great at guessing what language I'm using.

BTW, if anyone from Alexa, Apple or Google is reading this, please have someone work on this. It's been 10 years since I've been using it, and neither speech recognition, nor autocorrect have made any kind of progress on the dual language stack.


Google has multiple language mode which works with Google keyboard and Chrome. I type in English and French on the regular at it auto corrects each properly even when mixed (Google keyboard on Android will even complete suggestions properly when mixing languages).


I'm using Google's keyboard on Android in dual language mode, it works pretty decent.


I use the same, it's mostly okay except when I mistype an English word that happens to be a word in French and the keyboard decides to correct the next few words to French.

Also amusing is that autocorrect, even while in English mode, suggests the bear emoji when I type "ours" or the cat when I type "chat".


The 76% was from his competition score. It's a measure of the keyboards actual output, not the error rate of the typist. Any corrections the keyboard makes are already accounted for.


Not if multi-step words are handled by writing the wrong thing first, then deleting and replacing it when the second set of inputs comes in. Which is a very common strategy for implementing chorded keyboards.


I actively worked to remove my oh I typed that wrong reflex. Because I need to ready anything important again and I'll click the squiggles. Everything else dosent mattr enughf for me to kare. Or I use with Google voice dictation and just deal with the craziness.

Dictated not red


I can't tell if this is poorly typed as a joke.


Perfect


I know someone who makes tons of typos in their regular text, they don't seem to use autocorrect, and it's fucking annoying to try and read what they post. It's common typos like jsut instead of just. They just seem too lazy or indifferent to fix their shit.


You bring it up to them?

My main issue is when I type on the phone the text area is so small it scrolls out of view and the autocorrect puts the wrong stuff all the time. So I actually use voice to text a lot which gives you crazy phonetic sentences that end up being like weird poetry or GPT3 sometimes.

Some people said it was hard to read. So I now set the expectation that hey I'm out for a hike, so I'm not going to type. We can either do a phone call, we can do this later, or you can deal with the phonetic oddities. Usually people opt for the oddities.

That being said I notice my own mistakes much less when I'm in a slack type back and forth convo, I'm optimizing for speed not correctness. Whereas here I'm on a desktop on a 4k monitor with a text box the size of a google doc and taking my time.


When it comes to natural language (as opposed to code) I wonder what the optimal error rate is. Autocorrect fixes some issues, so it's not 0 errors.


It also introduces some, so it just might even out.


I'd say it's a lot suspect. Couple this with the fact that he scored 267 WPM with 76% accuracy in another competition. Until we see more and it gets more accurate I'm not sure typing this fast is very meaningful if I'm going to have to fix 24% of what I just typed.


I bet the 76% accuracy is coming from the CharaChorder's input method. From its website (https://www.charachorder.com/product-page/charachorder):

> CharaChorder’s internal processor arranges the letters on-screen in real time faster than the human eye can perceive.

So if it inputs one word, deletes it, and writes a new word every time you press a new letter, that would result in some decreased accuracy, even if the user doesn't ever correct anything.


Yeah, this thread is full of people who didn't take enough time to read anything. We're all too quick nowadays to get angry and type without thinking, sadly.


> We're all too quick nowadays to get angry and type without thinking, sadly.

Well, CharaChorder won't exactly help with that...


Spam out epic rants in record time!


They're typing too quick with their CharaChorder. :-)


could they not do that on device and only spit out the correct characters when it feels certain?


Probably. But then it's harder to learn because you have no feedback until you finish something. And for chorded keyboards especially, you're kinda always learning, because the dictionaries are huge and you often keep tweaking them to fit your personal needs.


I don't see why not. The UX would not be as good, though.


That sounds even more like cheating.


This article reads as a PR piece to promote that specific product.

No mention of Plover or http://www.openstenoproject.org


Yeah, I don't buy it either. I can type very quickly, but what brings my WPN down substantially is loading what I want to type next into my mind. Eg on typing speed tests, I can write the first sentence (if I can see it before the timer starts) much, much faster than the rest, because I can pre-memorize it. Later ones require me to multitask: read ahead while my fingers are still typing previous sentences, and this is much slower for me. If I memorize and practice a short sentence, I bet I could reach an extrapolated 500 WPM on my Kinesis Advantage with the Colemak layout with a few days of practice.


I look forward to hearing your results. I think you'll reach the 140-190 wpm range that way, not 250 or 350, and definitely not 500.


If you watch some of his videos linked by others here, you can see that to get those high rates he has chords mapped to entire words. I can do that with the Kinesis Advantage too.


Oh, indeed, with keyboard macros you could surely reach 500 words per minute. I thought you meant typing on Colemak.


QWERTY has a steep learning curve as well, it’s just that it’s so ubiquitous that most people don’t question or remember the work that was put into.


A significant difference between a conventional keyboard and steno is that a conventional keyboard is capable of producing all the of the text we normally type, and a steno machine is not. A necessary part of stenotype is a combination of pre-work (dictionary development for e.g. proper names) and post-work (editing the steno transcript to produce a "real English" document instead of one that may only have the generally correct sounds). Modern steno software helps a lot by partially automating these steps but steno is still inherently not capable of producing correct spelling without manual assistance - the basic architecture of steno is that you type the phonemes and (in modern usage) software guesses the correct spelling based on a dictionary. Much of the speed advantage of steno comes from the basic fact that it is a "lossy" process in the information-theoretic sense, that is, the "text" that you enter does not contain spelling information, only pronunciation... and even pronunciation is sometimes a simplified or substitute form as the American steno machine can't represent all of the phonemes that see use in English (mostly due to borrow words).

In the end, steno itself is probably not a lot harder to learn than QWERTY (although I think more frustrating because the "hunt and peck" option for steno is less intuitive and often slower). But it requires sort of a "supporting ecosystem" of skills and tools that is more complex and not amenable to use cases other than natural language. That makes it much less attractive for general use.


CharaCorder is not a stenograph, individual character entry is supported, just like a normal qwerty keyboard.


Typical steno keyboards can definitely type single characters. Of course the most sensible use of them involves chords (defined according to a text-specific 'theory'), but you might still use them for mostly single-character entry if, e.g. you were concerned about RSI.


Yeah but you are not going to get the speeds they boast about without steno.


I'm not the one you responded to, but I just wanted to add that it's still not steno. CharaChorder just has chording, but no theories, or whatever you call them for steno. You basically have to memorize every single chord on the CharaChorder afaik.


Ah thanks for clarifying.

I had thought that chording to produce a words was enough to qualify as steno.

Happy to be corrected :)


Apples and oranges anyway. Steno is phonemes, computer keyboards are characters.


> In the end, steno itself is probably not a lot harder to learn than QWERTY (although I think more frustrating because the "hunt and peck" option for steno is less intuitive and often slower). But it requires sort of a "supporting ecosystem" of skills and tools that is more complex and not amenable to use cases other than natural language. That makes it much less attractive for general use.

100% agree, also it may be harder because in the US we still don't teach kids how to fucking read correctly. It's still common to have elementary school curricula that do not teach how letters are related to phonemes _at all_.


The learning curve for stenography is much higher. With a keyboard, the buttons do exactly what they say and that only. With a stenography machine, every possible combination of buttons does a different thing.

When typing the word "Unprepared" on a keyboard, you just have to look at the buttons and find each of the ones with those labels, and hit them in sequence.

To type the word "Unprepared" on a steno machine, you need to hit the buttons UPB all at once and then PRAOEPD all at once, knowing that PB is a combo that means N, that the UPB set assumes that it's a prefix to whatever follows, that AOE represents the vowel sound /i/, and that your software understands the input "preepd" to be a shorthand for 'prepared'.

You can try it out with your regular keyboard in the browser here: http://www.openstenoproject.org/demo/


Keyboards that use multiplexed row/column electrical matrix will not allow you to detect more than two keys in any box pattern simultaneously. So pressing more than two keys on that virtual steno keyboard is impossible on a regular keyboard (for many key combinations).


Any decent (gaming/enthusiast/high-end) keyboard uses dioded on every key to achieve N-key rollover (sometimes only 4KRO on the USB output due to sticking with boot protocol or whatever other reasons)


The software has a workaround to account for that, as long as you keep one key held down, the "stroke" isn't considered finished.

Obviously if you're gonna use it as you daily driver, you'd want to get a NKRO keyboard, but the "hold one key and tap the rest" trick works for the web demo.


Ah, nice, I didn't notice that while trying to use it. :)


multiplexed row/col is not problem here. By adding more components (1 normal diode per key) you can detect all pressed keys. Normal keyboard can detect even 4 keys (not all combination ofc) by clever design.


Yes, but that's usually limited to modifier keys. This app expects the user to use multi-letter key combos, which the keyboards usually don't support. I have several normal keyboards at home and none can deal with that.


This was solved decades ago by simply pulsing the rows in sequence and reading the columns for each row. You can trivially do that hundreds of times per second.


I'd argue it's a different kind of learning though.

QWERTY is more 'what you see is what you get' - you push a key and you get that letter. Sure the layout is weird and learning to type takes some effort, but there's very little additional cognitive load. It's like a WYSIWYG editor.

Steno is like Vim, you have to have all of the phrases in your head tracking a lookup table cognitively. Over time sure that becomes muscle memory and lowers the load but I think it's less gradual. You have to frontload a lot of the commands first. IME most people will never do that so it'll always remain niche.


> IME most people will never do that so it'll always remain niche.

like vim.. winces in anticipation of things thrown at him


We’re niche, but happy


Well said. We're niche, but big enough to be self sustainable. The best kind of niche


"We may not be big, but we're small." -The Vinyl Cafe


I, and I think many other people, didn't really 'work' at learning to type. We just practiced as a side effect of using the computer rather than, say, taking typing lessons. For someone like me, QWERTY has a difficulty curve. I think seno technically wouldn't have a curve. The inability to hunt and peck means that without explicitly setting aside time to train in steno, there isn't any way to make real progress. Mathematically, a curve should not contain discontinuities!


I never learned to write on a keyboard in a controlled way and I still can't 'touch type' naturally.

I tried to learn in, but I always fall back to my old habits since my right hand hurt otherwise. I don't think I will ever bother getting fast. It is probably way better to learn it properly from the beginning.


A while back I decided to fix some of my bad keyboard habits. I forget the software I used, but it took me about 2 weeks of daily 15 minute practice to fix many of my bad habits and improve my typing skills in general. I’ve been thinking about doing it again because I still have some remaining bad habits, especially with the pinkies. Well worth the effort, at least for me.


I did about the same and I got alot better, but I think I need to practice a bit more to not fall back when I don't think about it.


The best way to learn IMO is to find a way to force you into it.

What worked for me was switching to dvorak. With no keycaps to match the layout, I had no option but to memorize it (initially with the help of a cheatsheet on screen). Conveniently, back then, I had no idea how to revert back to qwerty without restarting X after using xmodmap to switch layout, and I'm the type to not restart my session (or computer) except for updates every few months or so.

It also helps a lot doing this "at home" when you've nothing particularly important to do on the computer (running a basic touch typing tutor program for an hour or two and again later in the evening while just chatting away on IRC or such during the day is great because you're typing but have no real pressure). Not during work. It sucks struggling to get things done, and you're very likely to switch back to the old method and that'll just stall progress.

After one day, I had more or less memorized the layout but of course typing was still slow and somewhat frustrating. After a few days, it stopped feeling frustrating but was still a little slow. In a week or two I had reached or surpassed my original qwerty speed. I never used qwerty during that period, it was learn or bust.


For some reason whenever I try to touch type 'correctly' I make typos. But I do get most of my fingers in play when typing normally. I just have a weird stance -- left index on f, right middle on j, right pinky on p. Index and middle fingers do a little extra work, left pinky is mostly reserved for control keys -- esc, crtl...

I dunno. I don't have anywhere near 60 words per minute that are actually worth recording.


I think the point of touch typing is not about being able to type a novel per month, but about spending less mental effort on typing. No matter whether you touch type "properly" or not, the ability to keep focusing on the screen eliminates micro context switches between thinking and typing.


I think you have a great point. It is not as much about speed but about freeing your mind while typing.


Ye since private computer chatting essentially died out I see no use for fast typing anymore. When I and all my friends communicated via computer text typing fast would have been really convenient.

Ergonomic typing however, I guess is important. And I am bad at that. Looking down to find the keys might be a bad habit? I have almost trained that away.


>since private computer chatting essentially died out

News to me


Might depend on friends. I meant at a computer with keyboard, not on a phone.


A good number of people I know use iMessage, which can be used from a Mac. Similarly lots use Facebook Messenger, which can be used from the web.

...and from my understanding WhatsApp is very common in Europe, which can also be used from the web.


Hmm. I wonder if I saved myself some embarrassment on Instant Messenger by not typing too fast to think.

I don't really understand why properly trained typing is more ergonomic, anyway. If I keep my hands perfectly on the homerow, each letter is nearly the same exact motion. The way I type normally, my hands move around a bit, so there's a couple character history built into my motions. This seems to me like it ought to reduce the repetitive motions, which are what lead to repetitive strain injuries, right?


Ye keeping the fingers at one row home position doing the same dance feels terribly unergonomic and crammed. I was only thinking about the neck looking down (dentists have problems with that).

Maybe I stress my hands more while tryhard practicing touch typing than I would in normal use though.


A keyboard tray helps since you can’t comfortably keep your eyes on a keyboard that’s under the desk.

Forcing myself to stop looking at the keys initially increased error rate to an unacceptable/demotivating level, but in a couple of weeks I was touch typing quite a bit faster than I could hunt and peck with no loss in accuracy.


We took typing in 7th grade (12-13 years old). We had a classroom set up with manual typewriters on every desk that had blank keytops.

Lesson one was learning A-S-D-F J-K-L-; and typing words and combinations on the home row, then we'd learn a few new letters each day. After that we had to learn how to use the tab function, address envelopes, and common forms for business letters. We were tested with "eyes on copy" typewriting where we didn't look at our hands.

(Other obsolete skills I learned in "Jr. High School" include how to set letterpress type, and Morse Code. I still use Morse Code today on amateur radio.)


When I was a kid I got an ATARI 600XL I'd type in code for hours into memory just to have it fail due to errors. But I got good at typing at only age 14.

Then a few years later as teenagers do in high school I choose typing class since all my friends did. So I learned more typing unusual for a guy at that time mid 1980s.

Then IRC when I got onto the Internet in the early 1990s. You had to type very fast to get in a jab or a joke before someone else did.

Now I'm in IT and typing is ubiquitous.


Mmmhm not sure I'd agree.

Initial learning curve is practically non-existent - Anybody can type a letter "A" or a word or a sentence on a QWERTY by looking and poking. It is not my understanding that I could do ANYthing on a stenography device without significant, serious training.


So-called discoverability is certainly one important metric.


> has a steep learning curve

just to make the pedantic point, learning curves show skill improvement plotted against time spent/experience, and therefore a steep learning curve means you learn quickly. A shallow learning curve is the difficult one. /pedantry

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


Your own link explains that the phrase is a misnomer. You’re not pedantic; you’re just mistaken about others being wrong.


you have it backwards. the quote that you are referring to, that "a steep learning curve" is a misnomer, means it's a misnomer if by that you mean it's difficult.

It's not wrong to say "steep learning curve" if you mean it's easy


That’s not how the phrase is used colloquially. You’re not even wrong, but your prescriptivism is at odds with common parlance.


you were incorrect to describe the wikipedia article as backing your point.

I self labelled as pedantic and pointed to an explanation, that does not make me a prescriptivist, it makes me didactic.


> you were incorrect to describe the wikipedia article as backing your point.

I’m not relying on the Wikipedia article to support my claims. I’m relying on my native English language usage. I have literally never heard anyone in my life use the phrase “steep learning curve” in the way you use it. Not even once. However, I could have heard it in an ambiguous context and assumed that the speaker or writer intended to use it in the way that you do. I’ll admit, in that case, that I may have heard it used the same way you do, and assumed that they mean to use it in the way I do.

I am not making personal attacks here, so I hope I didn’t offend, not to suggest that your disagreement implies offense. If I did, I sincerely apologize.

I wonder if this is a case of certain phrases just being imprecise or ambiguous enough to be difficult to use in conversation or writing.

For example, suppose I applied some wax to a polished wood floor, way more than I thought I did or meant to apply originally.

Suddenly, you walk in from the next room or down the hall. I stop you before you slip and fall on the excess wax, saying, “Wait, I spilled a deceptively [small | large] amount of wax on the floor, and I haven’t cleaned it up yet.”

You stop in time, and don’t slip. I clean up the wax, and all is well.

If all you have is my single sentence to go on, whether I had said that I spilled a deceptively large or deceptively small amount of wax on the floor, and no additional context to know that I spilled too much wax on the floor, you might deduce from context that there must be too much wax, thus leading to a slipping hazard.

Now we’re having coffee later and I’m telling you about my vacation, and all the new flora and fauna I saw. I tell you about seeing a rare and elusive jabberwocky, and I told you it was deceptively [small | large]. Without further elaboration or context, you may struggle to deduce whether I mean that the jabberwocky is smaller or larger than expected or than it appeared, or both, or simply that it was smaller or larger than expected for jabberwockies, given just that single sentence.

Does that make sense? I used jabberwockies because they are literary creatures that do not exist in real life. You can replace them in the example with anything which you, the hypothetical listener or reader, is unfamiliar with enough to not know how small or large they are in relation to other referent in my vacation story, or anything else I was comparing it with in that context.

It appears that we can both be right, depending on usage and context. You may not agree, and seem not to; I find my usage appropriate for the contexts I use it in, and clarify with supporting statements in context so as to not confuse others, because I know that this phrase is used in both ways colloquially and mathematically.

To the degree that you insist that the phrase “steep learning curve” only has the one meaning you describe and use, I call your insistence prescriptivist. If you deny that others use the phrase correctly when they use it in ways that make sense, given the knowledge of how other people use it in ways that differ to your own, I would call that denial pedantic, and not didactic, but instead quirky, but otherwise okay, as long as we both understand what we each mean, and otherwise understand each other’s meaning and context.

https://english.stackexchange.com/questions/6209/what-is-mea...

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

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


> I tell you about seeing a rare and elusive jabberwocky, and I told you it was deceptively [small | large]. Without further elaboration or context, you may struggle to deduce whether I mean that the jabberwocky is smaller or larger than expected or than it appeared, or both, or simply that it was smaller or larger than expected for jabberwockies, given just that single sentence.

I had no idea this was even ambiguous before your comment.

I've hitherto only considered the possibility that "deceptively large" means "large, but appeared otherwise. Having now googled this, it seems there's some debate about it, and while I can see both arguments, I think this is down to incorrect usage muddying the water.

In "A is deceptively B", "deceptively" is an adverb, so adding extra definition to the manner in which "A is B". Short of ignoring grammar rules entirely, I can't see a good case for arguing the opposite, even though it seems that examples exist where the sentence only makes logical sense by interpreting it as a negative.


The English language is fucked.


From your link:

> An activity that it is easy to learn the basics of, but difficulty to gain proficiency in, may be described as having "a steep learning curve".

That matches QWERTY exactly.


the way I explained it gets you to the right answer more quickly and frequently, plot proficiency vs experience (generally measured by time)

yes, a curve can change its slope from "learn the basics of" vs "gain more proficiency" and any place where it's not steep indicates difficulty.

that some things might have easy basics and other things might not have easy basics isn't really material, but where it matters, picking up the basics would be steep if it is easy.


A steep learning curve can also mean you make very little progress for a long time until suddenly being proficient.


you can plot things any way you wish, but the people who invented and named learning curves plotted them so that "a lot of learning" was steep.


I understand how they plotted them, I followed your link. My point was that a sigmoidal learning curve can be both steep and require much time to reach proficiency.


I heard about this years ago and hate to see the phrase anywhere now. I just avoid it entirely myself. I don't want to be knowingly wrong or create awkward situations explaining it to everyone.


To my shame I've never learned how to type. It does not really impede my coding abilities as it is fast enough to follow my thinking. But if I need to chat online my slowness drives me mad (I assume it does the same to the other party).


It’s worth practicing! I bet it’s not too hard to get reasonably fast in a few weeks. The hardest part will be un-learning any bad habits, but if you reckon you’re currently slow, you probably don’t have too many.


I've been typing with the same speed for the last 30 years. It is not improving so I really doubt special exercise will do anything but marginal improvement.


Don't practice typing faster with your current method, because you already have for 30 years. Rather invest some time to learn a better method.

I learned touch typing within a week on holiday vacation with a software similar to this website: https://www.typingstudy.com/ (Young me felt very proud of being able to type like dad.) Once you have invested the time to learn the basic movements, you will become accustomed to them naturally.


You would be surprised by hovering your fingers over home row how naturally quickly you will start typing.


Take the time to learn, it is absolutely worth it. Do not pay attention to the comments saying "you will be at some speed soon" as that's not your aim. Learn to type precisely with as little errors. It will take you several months to just learn the keys. That's fine, it's like coding, learning the slow and hard way, you are in it for the long game. Good luck!


It's worth learning and really doesn't take that much time to get to a decent wpm.

It makes it way easier to communicate via text.


I'm not sure the learning curve is all that steep. I learned it as part of a "keyboarding" class around... 1989 or something (I was in middle school). Learning to type was only a small portion of the class, and it happily included games and letter art alongside the learning.

Best class I took because of touch typing.

On the other hand, I have plenty of peers that type nearly as quickly using two fingers or simply by doing years of using a computer. People learn the layout with use. It is easy enough that switching from a standard American keyboard to a Norwegian keyboard (whose alphabet has 3 more letters in addition to some other European language letters and punctuation) was a non-issue.


True, but QWERTY is labeled and obvious how to use. Hence the great number of woodpecking that occurs out there.


That's why I got a Dvorak skin for my TypeMatrix keyboard. I know QWERTY well enough I don't have to look for it, and I can switch to Dvorak and be able to peek if need be.


I've been using dvorak for a long time. I kind of wonder if not having a keyboard with the correct dvorak keycaps actually helps learn it faster than it would if you had the correct keycaps? If you can't look at the keys to know what they are, then it encourages touch typing. Still, I can see where it'd be nice to be able to fall back on looking at the keys.


Yes, I think not having keycaps helps enormously. Force you to memorize the layout and you can't get lazy about it. https://news.ycombinator.com/item?id=29837101


I think the best argument for things other than QWERTY nowadays is ergonomics, not speed. RSI is no joke.


Agreed. I've never gotten back to my old qwerty speeds of up to 160wpm since relearning proper homerow on a split keyboard, then later learning Dvorak, then again later learning Workman. I'm in a lot less pain, though, and I'm still generally above 120wpm at least.

FYI I didn't switch off of Dvorak because I was bored of it, I'd traded my left hand pain for right hand pain, and I decided it was a bit worse. Workman is more evenly split between hands where both qwerty and Dvorak had a major bias.

Workman also seems to beat Dvorak and Colemak in efficiency, but if that were my main concern, I also considered qgmlwb and Halmak, which IIRC are even better.


> QWERTY has a steep learning curve as well, it’s just that it’s so ubiquitous that most people don’t question or remember the work that was put into.

This assertion ignores crucial differences in the basic mechanics of typing in a standard QWERTY keyboard (i.e., look at the keyboard, press one key, get the desired character) and using a stenograph/chorded keyboard. A standard keyboard's discoverability makes it incomparably easier to ramp up than guessing which key combos get you a specific character combo/word.


Every time I get a new laptop I have to adjust my muscle memory since their layouts vary so much :( still can't type the upper right letters consistently on my current laptop.


Yeah, I switched to Dvorak a few years back, it was the worst experience in the world. Granted I had to fight against my qwerty reflexes


I was lucky to never have properly learned to touch type on QWERTY in the first place, I had key positions memorized but I always used my index fingers - it actually hindered me from touch typing as the bad habit would always creep back up. Thus I forced myself to learn from square one with Dvorak, not having the keycaps on my keyboard to fall back on - it was painful, but I'm glad I did it.

Now 15 years later I really enjoy the flow of the layout, a majority of the time when typing you alternate between sides of the keyboard when typing and it just feels good.


When I learnt to touch type I switched to Dvorak at the same time. I liked the idea of improved efficiency but whenever I had to use someone else's computer or they needed to use mine it was too much hassle. Once I'd relearned to touch type with Qwerty life was so much easier. Also hjkl with Vim using Dvorak was just too hard to get my brain around


J and K are adjacent in Dvorak, and they're down and up, respectively, which makes vertical scrolling just as easy, IMO. I never really used H and L but instead rely on mostly W and B (which aren't adjacent but are close enough). So by happy coincidence, navigating in vim works just fine, I think.


Yep, I've never felt the need to remap hjkl for other layouts. Very exaggerated issue from people who didn't stick with it long enough, I think.

I was worried about it with Workman because it was slightly less logical. h still left of l, but quite far away, j and k are placed so that it's like you've inverted your camera controls in a game. Very easy to adjust anyway and I don't really think about it now. Your point about w and b is very good as well. Advanced vim users are actually less likely to use hjkl much due to having faster ways to get around in most cases.


Yep. I use vim & evil regularly with dvorak. J and K as well as H and L are just fine. H and L are placed very similarly to where they are on qwerty; H is just shifted one position to the right, and L is one position to the right and up. So they still have the "left-right" feel, and I like that H is under the index finger.


> hjkl with Vim using Dvorak was just too hard to get my brain around

Why not just remap? My intuition is that the cascading conflict wouldn't be that hard to resolve (at least compared to learning Dvorak...) but maybe I'm wrong.


You can easily and quickly add a Dvorak keyboard to someone else's computer, deleting it once done. I did that all the time when I worked as an editor, having to occasionally edit on clients' computers. The benefit of speed gained from using Dvorak outweighs the inconvenience of having to occasionally add and delete a keyboard on other computers.


Buy a mechanical keyboard with the ability to modify the firmware. I did this and it was worth the $200 price tag.

Forgetting to configure computers after finishing with them, not knowing what layout was in use in the login screen, initial configuration, etc all cost time. I also use RDP a lot which is layout roulette (sometimes it changes, sometimes it doesn’t)


Exactly, I'm a touch typist and regularly train on various websites. The biggest blocker after you start reaching high speeds isn't even the keyboard but the brain. You need to read far far ahead to write at 130+ wpm you basically are typing a word while you read the following sentence. 200+ requires writing even more ahead.

A normal person can barely "mentally process" that many words per second.


The average person speaks at 160 words so i have a hard time believing they can't process 130. Although that would explain why people say things that are so stupid lol.


That was referred to the 500 wpm.


Something I find interesting is that thoughts definitely spawn faster in my head than I can verbalize them. I wonder if you learned to think in chording if you could think faster than with sounds. And if we could, what would be the upper limit?


A bit OT, but can you recommend/do you know of any touch typing training websites which include training for the numeric keypad, or even for the numbers/symbols row above the alphabetic symbols?


Monkeytype works well for this if you tweak the defaults. Turn on capital letters, punctuation, numbers, and change the time to 60 seconds. I did this at the start of learning a new keyboard layout and got quite comfortable with the numrow and symbols.


I learned touch typing on https://www.typingclub.com/


It works without steno, too.

Like you can take a look at their quick reference guide shown in the "coder" section of their website.

I don't know how well it will work but it looks viable.

The think I'm mostly worried about is that most human fingers (not thumbs) aren't really designed (or trained) for sideways movement. Does anyone has the necessary anatomy knowledge to know if this has a increased risk to cause health issues if used long term as the main keyboard?


To me, it comes down to this line:

> have a computer program generate a predicted word

Do steno devices do this? This right there easily disqualifies it from competitions. It's like using the keyboard autocomplete. If you're using a machine to predict and correctly type the word for you, it defeats half the point of typing competitions, which is specifically about typing words without typos. The majority of time lost in these typing races is when you make a typo and have to go back to correct it.

Also, separately, I know HN skews more towards pogrammers, and I feel these keyboards wouldn't be very useful for that.


stenography involves shorthand. This is what a stenographer is actually typing https://en.wikipedia.org/wiki/File:Steno-example.gif


But does one given shorthand get hardcoded to go to a single word, or is it using some "autocorrect" system that predicts the word you want? Whatever system this uses seems more powerful than steno to me.


It says he competed at 267wpm in the article. The monkeytype site totally bans 300+ so I think that's still very impressive.


"with 76% accuracy" - I'm surprised typing competitions even allow scores below some accuracy level, e.g. 95% or so. Missing every fourth letter means it's basically gobbeldygook, and if you were actually typing anything where you even needed to be remotely accurate you'd spend at least 2x as much time revising.


The accuracy measurement is flawed, because the firmware of the device is deleting chorded letters and retyping the chorded word, which artificially reduces accuracy


If it's happening at the firmware level why wouldn't it just resolve the word internally and output only the completed word to the PC?


Yes. It’s actually artificially increasing accuracy.


Yes but the accuracy percentage is low because the software’s essentially guessing what he’s typing (not unlike swype keyboards + autocorrect on a phone though implemented differently). It would be way slower to correct errors and type at 100% accuracy because each error has to be corrected based on some cognitive process (like looking at autocorrect suggestions for example) which is far far slower than correcting a typo in a QWERTY keyboard. I can type around 110 WPM with 95% or higher accuracy and I can also feel the majority of typos I make and correct them without active thought.


Yes, but at only 76% accuracy. If you were to include the time spent on going back and correcting mistakes, I bet he would be down in the 50-100 wpm range, if that.


The video from the product site makes it clear that it's not really typing. More like matching chords to a rather small list of words:

https://www.charachorder.com/

Great for tiktok demo videos I guess.


I suppose then we are getting into the nitty gritty of what the definition of typing is. Typing to me is entering text data. For me, if I use something like Swype on my phone, where I don't even have to hit the actual letters I am trying to type, I would still call it typing. This device looks like it's matching a chord made up of the majority of the letters of a word to the expected word. I would still call that typing. Just like driving with adaptive cruise control, lane keeping assist, and automatic emergency breaking would still be called driving for me.


Yup, having to type unseen text makes a lot of difference.

Here's what I personally use to test my keyboard setups: monkeytype.com.

My best is ~95 WPM and 100% accuracy.


I like typing :)

Here's 163 wpm on a macbook keyboard: https://www.youtube.com/watch?v=-ulK2VJdQpw

Here's 175 wpm on a custom mechanical keyboard: https://www.youtube.com/watch?v=kO-A8rVJDOM


It's weird, how it looks more effortless, the faster someone types. Well done!


I think at high speeds you do actually end up optimizing force per press and time spent on a key out of necessity. It's like cycling, where they say "it never gets easier, you just get faster", or in other words, they can do a lot more with the same amount of effort.


That's amazing. Do you do any wrist stretches or recovery tasks afterward? Is there a routine you do for prep?


No idea how they can type that fast with those long finger nails!


I actually keep my nails pretty short (they're only about 1 or 2mm past the furthest-out part of my fingertip) for typing reasons!

My speed gets a lot worse if I let them grow out, so I like to file them down regularly (instead of clipping them occasionally) to keep them at a length I'm used to.


They increase stiffness. You hit the keys with the hard nail instead of the squishy flesh of your fingers, which reduces jitter and improves timing accuracy.

(I'm a fast typer, but not as fast as that one, also with long-ish --- around 0.050" --- fingernails.)


Tried that and got 109 WPM with 100% accuracy on my Macbook Pro (non-butterfly) keyboard. I usually score around the 80s with regular sentences and punctuation on other sites though, so subtract 30% from the WPM results on this side to estimate a regular WPM.


I also felt scores were inflated compared to 10fastfingers when I first tried it. Tweak the default settings and it can be a lot more comparable.


Saw the article later, they use monkey type too. I would love to see them in the 30s test.


People are getting too hung up on the "500 WPM" part. He's not saying he can type 500 WPM. He's saying that he scored 500 WPM on a specific typing task and he's doing it intentionally. He gets to say he typed 500 WPM and the website that's flagging him as a cheater has to be the one to explain why which just gives him more promotion.

It's not about how fast he's typing. It's about being able to say that his product is significantly faster than everything else.


well that and the fact that this is transparently an advertisement masquerading as journalism.


I can't stand software autocorrect, so hardware autocorrect would be a nightmare for me. This works only when you type regular text in one language.


I can type the word "a" at 1000 WPM!


That's 32 keystrokes per second, I'm impressed. Unless you're using vim, in which case a minute is a long time to type: `ia <esc>0d$1000p`


Or quicker: 1000ia<esc>


Touche!


No, it’s one keystroke, executed in 1/32 second.


In Emacs: 'C-u 1000 a'


It's probably harder to type a bunch of the same letter than real words unless your whole keyboard types that letter.


No you can't.


I used to have a pretty high WPM, above 120 if my memory serves. You can only sustain that speed if you’re working from memory or directly copying text (as is the norm for typing speed tests). In real life I can’t formulate words anywhere close to that fast, so my WPM ends up probably being less 40 if I had to guess.


Does it matter? The article is talking about typing competition. The WPM is in this context. There's lots of competitions which seem to have no real world practical value other than people believing it's enough to put together a competition for it.


I have never participated in a typing competition but I would be shocked if they consisted of typing a single canned sentence.


Based on that video, it's more of a macro keyboard. Assign each of the ~10 words to a button, click buttons in order, claim 500WPM


No it’s not, there’s another video explaining, the keys are “2d joysticks” type switches, so typing doesn’t involve removing the fingers at all.

I would think it takes a lot of time to get used, but seems like a smart approach for speed.



The performance in that video is nothing like the performance in the 500 WPM video. Accuracy is well under 100%, and he seems to be actually inputting random words - not just macroing a pre-defined phrase.


I don't mean to be rude, but did you even watch it? He explains the firmware deletes a word when it detects a chord which skews the accuracy.


That's the point the other commenters make. If that's what's happening, why does the 500wpm 100% accuracy?


Isn’t just just obviously bs? You can’t even think 8.3 meaningful words per second. Try it. It’s impossible to sustain.


You probably can in bursts. Being able to type as fast as you think is incredibly useful when brainstorming ideas.


"he did win another typing competition using the CharaChorder to hit 267 WPM with 76% accuracy."

76% accuracy doesn't sound so great to me, no matter how fast I can type on it.


Interviewer: "I heard you were extremely quick at math"

Me: "yes, as a matter of fact I am"

Interviewer: "Whats 14x27"

Me: "49"

Interviewer: "that's not even close"

me: "yeah, but it was fast"

Shamelessly stolen, but yeah, 76% accuracy is only 3/4s of characters typed correctly; if that "works" you could probably just train yourself to not even type the lesser used parts of the alphabet ...


Back in the manual typewriter days, every typo subtracted 10 WPM from your score. (Source: my junior-high typing class.) It seems to me that 76% accuracy would put you into negative WPM by that reckoning.


I like TypeRacer's approach – your time includes the time it takes you to backspace to fix your mistakes and re-type it correctly.


I don't think it's bad for 'competitive' typing but it's nothing like the real process where fixing mistakes at the end is almost a triviality instead of completely derailing your typing.


For most people, that involves relying on another service ("look to see what words have a red squiggly line under them").

It is not a good way to measure "typing speed".


That seems optimistic, odds are your typing flow will be interrupted as tou think about the mistake and go around to fix it, plus you likely only realise it a bit later and need either some movement or to remove entire words and retype them.


Since WPM is a rate, and total typos is an accumulated values, it seems that whether this would bring you negative is a function of the test length. Seem like it would make test scores not very portable.


The way it was explained to me years ago is that every error deducts an amount (probably 10, I don't remember exactly) from the running total of correct characters typed (WPM was calculated by dividing the number of characters by the time and by a fixed "average" word length (ISTR they used 6 (in Dutch)).

So if you type 100 characters and 24 of them are incorrect, the counted total is 100 - 10*24 < 0, independent of the test length.

This was in the time of typewriters; computerized text entry did exist but was certainly not ubiquitous. There was no real way to correct mistakes, so I can imagine typing correctly from the beginning was a bigger deal than it is now.



when taking notes a few typos are ok, e.g. "we got an avg of 14 complaints in our 27 stores" is "about 400 complaints" :-D


Thanks for the chuckle!


idk, wrds nt lke mth.


If software used for typing competitions is anything like a typing test in a typing tutor program (which give similar WPM + accuracy measures at the end), then it would track the number of incorrect inputs you make, but also wouldn't let you proceed (i.e. would ignore all further input) while any incorrect input remains in the buffer. With such software, you're expected to correct each word you type before moving on to the next; the moment that you submit the corrected word is the moment the word is acknowledged as "a word" counting toward your WPM.

With such programs, the WPM score measures how many total times per minute you "finished" a correct word, including any time spent correcting the word; with the accuracy score measuring how much extra work was done, on average, doing those corrections.


This doesn’t work like that. It moves on to the next word the moment you type a space. Also, https://monkeytype.com/about says:

stats

wpm - total amount of characters in the correctly typed words (including spaces), divided by 5 and normalised to 60 seconds.

raw wpm - calculated just like wpm, but also includes incorrect words.

acc - percentage of correctly pressed keys.

char - correct characters / incorrect characters. Calculated after the test has ended.

consistency - based on the variance of your raw wpm. Closer to 100% is better. Calculated using the coefficient of variation of raw wpm and mapped onto a scale from 0 to 100.

So, they assume the average word has 5 characters. Makes sense for computing wpm in random text.


You can go back and delete errors, and also the default corpus is random words which probably average out to close enough to 5 characters. There are other modes which include punctuation and capitalization, as well as longer quotes to type, as well as expert and master modes (which fails you on a wrong word or character, respectively). I've never heard about this site before but it's pretty full featured!


In addition to the other reply, if they did that the test would be so far from real world typing to be meaningless and nothing but an inconsequential curiosity.

Only in tests is there an option to check against what is expected. In the real world you could at most apply some dictionary and grammar checks, and we know those can go either way. Those tools don't know what you actually wanted to write after all, and still suggest without comprehension.


Even if you type without making corrections, you still need to go back and make corrections, so you need an estimate for the time spent on corrections. Forcing corrections midstream is a decent approximation.


It'd be neat if the test

1. used a fuzzy autocorrect of the same kind mobile OSes and word processors have (not correcting to the known text, but instead only doing what such logic usually does — only attempting to correct words if they're not recognized as words; and only moving them to the lowest-Levenstein-distance match.)

2. treated words that are correct after autocorrection as being "correct", with no ding to accuracy.

Since accuracy is about how much extra time and effort the human typist has to spend to get the text correct, and since text that can be auto-corrected requires zero human time and effort to correct, it shouldn't really be considered incorrect.

Or, to put that another way: if I'm typing really fast on my phone's keyboard, and all the right words are ending up on the screen, what does it matter if I'm hitting the right keys with my thumbs? It's the WPM and accuracy of the cybernetic system composed of me + the auto-correct that matters.

In that sense, you can sort of think of an auto-correct as actually being a very similar thing to a stenographic keyboard (at least, in the hands of a person who knows how to take advantage of it): it's really a system to enable accelerated typing, by allowing you to key in the letters of a word out-of-order or simultaneously and have the word come out in-order.


In real world, there is check for typos - your eyes. You fix them after yourself.


> but also wouldn't let you proceed (i.e. would ignore all further input) while any incorrect input remains in the buffer.

I hated these ones so much in school because I'd be like 3 letters past the typo and reflexively hit backspace 4 times to get rid of the typo, which just registered as more mistakes because it had blocked those 3 additional letters and so would only accept 1 backspace.

(I'd already learned how to type from having to quickly enter chat messages in StarCraft multiplayer)


I've been active in my typing practice since I was about 4, and once I'm warmed up I can usually hit high 120s with 100% accuracy on unseen MonkeyType using a qwerty keyboard with red Cherry switches. If I ignore accuracy and go for 'raw' speed, I am able to achieve significantly faster WPM, typically between the range of 140s-180s. The issue I run into with programs like these is that I have to read/match the expected words. Even if I can see the words ahead of time and have a moment to try and memorize, I just go so slow trying to get the correct words. To contrast this, when I write for fun, and know what I want to say ahead of time, I'm fairly certain I'm able to burst into the low 200s (don't have a solid way to test to be sure unfortunately). Depending on the complexity of the error, Grammarly usually spits out correct solutions when I'm around that 75% error rate. The vast majority of the time I'm able to auto-correct everything with a single click. The only exception is when my hands drift and I end up typing something like 'gppnst' instead of 'foobar'. What I mean to get at is that it's somewhat dependent on the types of errors that are made. If it's easily auto-correctable errors at 75%, awesome, the person is hitting an incredible WPM, if instead it needs to be corrected by human thought, then I would completely agree with you.

The real benefit to this speed/accuracy is that when writing long-form text, rather than say programming, the speed of typing can either match or exceed the thought process. There is a huge amount of utility in this approach and the keyboard becomes a true extension of the mind. I'm still waiting to find a good real-time auto-correct that doesn't screw with my flow at high WPM, I'd pay pretty good money to have something that just works out of the box.

On a practical note, I've considered trying to hit the numbers that the author claims, but I'm already so limited by my own train of thought. With the additional finger/wrist strain (And yes, there is _significant_ strain when these levels are approached for any length of time) I just don't see the costs making sense when everything can be fixed in just a few clicks after typing a great many paragraphs of text.


Monkeytype also has a 'zen' mode where it measures the speed of you typing whatever you want.


Thanks for the heads-up, I just tried it out, somehow never noticed it before!


> An activity that it is easy to learn the basics of, but difficulty to gain proficiency in, may be described as having "a steep learning curve".

It's the same for me. I doubt I can hit 200 WPM, but when I'm typing what I'm thinking (like right now) I'm significantly faster than when I type something I've never read, especially if it's just random words. On the other hand, I've never been much into measuring my WPM (I can type sufficiently fast for everything I need), so I just assumed this was a skill one can train to actually get those insane scores.


> when I write for fun, and know what I want to say ahead of time, I'm fairly certain I'm able to burst into the low 200s (don't have a solid way to test to be sure unfortunately).

What do you estimate your actual WPM to have been when you wrote your whole comment?

I find that it can take time to think about what I need to put down, and often need to edit and re-phrase for my audience.

The whole WPM thing focuses on pure technical ability and ignores that there is more to writing than being a stenographer.


I love that this thread popped up when it did, as part of a new years resolutions of sorts I decided to start writing short stories and have been thinking about this topic quite a lot. For me it depends, when I know exactly what I need to say, it's pretty quick 70-80 WPM or so, that's with the pauses, edits, etc.. If am really struggling to come up with the words, it's probably closer to 10 or 20. One of the perks of being able to burst type though, and I don't really have a better term for it, I'm able to just kinda tense my muscles and 'jiggle' my fingers in fast succession, is that I can backspace and move the cursor _very_ quickly on the edits. Typing is a means to an end though, the hardest part of writing is ensuring the communication fits the bill. If the bug is tricky, for example, my words per minute can turn into words per day :)

Being a programmer probably allows me to be faster than a lot of other speed typers for specific tasks, but I'm nowhere near as fast as they tend to be for straight word counts where the numbers tend to look like 100% accuracy at 200+ WPM. As an example, I always set my text editors to backspace both words, and sentences (ctrl+backspace, and shift+backspace). This trick, jiggling, and good searching, allows me to move through code _much_ faster than most other programmers I've meet. The only exceptions have been people who've setup and mastered vim, (sorry emacs folks, I've just never encountered any of you in the wild) and I've been slowly been working on gaining proficiency there.

Another user pointed out that there is a zen mode for monkeyType, I'm trying it out on this post. I just sat down to write for the night and I'm using this as my warmup (going this fast requires it for me). I knew a good bit of what I've wanted to say as I started this post but there have been some pauses and edits.

MonkeyType claimed 72 for this comment, bit faster than I would have guessed, I'm curious as to how realistic that number really is!


I'm guessing that hands drifting only happens when you're really pushing for speed and get sloppy?


Combination of things, if I'm cold (literally and figuratively) or I'm still warming up, if I'm really pushing hard, or when I've been going too long and the fatigue sets in. I don't really know how common it is among other typists though, I've found I've gotten sloppier over the years due to auto-correct and google search 'guessing' being pretty damn fantastic.

Another story, when I was middle school, I took what I thought was a computer programming class that turned out to be a god awful typing class. Of maybe 30 kids, over half would completely fail to complete the assignment during the hour. The students that got through it were almost all my video game/computer geek friends. I personally hated the class as I thought it was a complete waste of time as I could already type fairly well. The carrot on the stick was that you could surf the internet, play games, whatever, if you completed the the assignment. Considering that the internet was wildly faster than my dial up line at home, I would burn through the class as fast as I could, literally always finishing before everyone else, spending 15 minutes tops on it.

Well, one day midway through the semester I overheard another student asking the teacher how I was always done so fast. This guy was a completely smug hard ass and was apparently a 'computer guy' prior to his teaching career. He had the gual to say it was only because I would make so many mistakes and that I was such a sloppy typists that I wouldn't be worth much to anyone later on. During the next class I made a point to hit 100% accuracy, declaring loudly to him so that the entire class would hear. I did that the entire rest of the semester, I loved it, he never showed his smugness towards me after that. Still really bugged me though, I realized even back then that computers were getting _good_ and handling typos, and that highly accurate typing was mainly for suckers.

As an aside, this was all during the era of the 'Ask Jeves's' of the world, I switched to google simply because it was so damn good at understanding the intent behind my typos. Incredibly, they're _still_ the only auto-correct I've used that understands and corrects the drifts!


It's not 76% accuracy of the final text--it's 100% accurate, it's just that the way you get there using the CharaChorder is recognized by Monkeytype as inaccurate.

If I type the word "hello" on my CharaChorder, I type the keys `helo` which Monkeytype will recognize as an inaccuracy. The CC's internal firmware then recognizes I typed helo and wanted hello and will backspace the word and replace it, much as if you did the same thing.

Do this on enough words where the chord trigger doesn't line up with the chord and you get the ~75% number.

But the final text is 100% accurate


How does it know you weren’t intending for hole?


Every chord is explicit.

I think someone else in comment thread mentioned that it comes preconfigured with 500 chords--these cover the 300 most common English word (some duplicate chords in case you prefer to hit them one way or the other).

It's up to you (or someone else if you import their library) to setup and configure different chords.

This is one of my least favorite parts about the product--it's advertised such that it seems like it is magical but really it is just like have a bunch of macros.

So in this case, you might have h+l to indicate hole, or h+o, or h+e--whatever is most comfortable for you. Hole is not configured out of the box so there is no way to chord it...but it is just like a keyboard in that you can just type it. You can type h, o, l, then e just like you would on a keyboard.

It's this mixing and matching of character entry + chording which makes the device pretty interesting.


Thanks. That’s definitely the missing piece that wasn’t explained in their TikTok videos. He talks about creating custom cords for odd words, but not something like hole.


Yeah, way too low for programming but I wonder how low the bar is for modern autocorrect systems when writing English sentences.


one would think that programming requires especially low accuracy, since most of the things you type are either language-specific or previously defined.


I just went on 10fastfingers and got 124WPM with 100% accuracy on the 60 second test. Can't say I'm impressed either - 2x as fast in exchange for far worse accuracy and use of a predictive layer is pretty bad overall.


It depends on how it is calculated. If that is one in four key presses vs one in four words it is very different


Yes, that's terrible. The finished text was not corrected in that time, was it?


I looked into buying one a few weeks ago but ultimately decided not to move forward. One reason was that it comes with a pre-configured list of 500 chords, so it's up to users to create their own beyond the initial list. The other was that the build quality doesn't appear to be that great (with the PCB board clearly visible).

The reddit community[0] has some helpful reviews and progress updates. And their discord server[1] is quite active with lots of helpful links and advice.

0 - https://www.reddit.com/r/charachorder/ 1 - https://discord.com/invite/jMj6grUuBc


FYI links to reddit do not work if user is not logged in - at least on mobile. Sad to see its decline.


On mobile replace the “www” with “i” to get something that actually works on a phone. The new mobile site is such horrible user experience in almost every way.



old.reddit.com still works

https://old.reddit.com/r/charachorder/


Fascinating. This is the first article I’ve read where the source is exclusively tiktok videos


Keen, CEO of the keyboard company, is mentioned several times.

At the end of the article:

> Keen did not return Motherboard’s request for comment.


That's because pretty much all of their promotional content is short TikTok videos.


And Discord.


After reading only the headline, I'm not surprised.


>You may soon see that coworker with the weird monolith style trackball mouse rocking this strange peripheral and claiming he’s upped his efficiency in ways you can’t possibly imagine.

Since coding is 20% sitting in meetings, 50% reading code, 10% drawing on whiteboards or sticky notes, 15% drinking coffee, 4.5% fighting impostor syndrome and 0.5% actual typing, I think I am quite safe with my good 'ol QWERTY.


I'm having difficulties finding the source right now, but I remember reading a story about a sysadmin that was berated for his typing speed and the story ended with him replying "I don't get paid to type code fast, I get paid to press enter very, very slowly."


playing the devil's advocate, typing speed is important for these reasons:

1. It allows for minimal disruption to thinking process, close to 90% of the time even when coding is not actively typing, but when typing is subconscious and fast it removes the potential of where it interrupts the thinking process.

2. While this has no bearing on individual case (It would be Bayesian, some people consider that to be heretic), but a software engineer who does not type well have a greater chance of having had less practice.


> It allows for minimal disruption to thinking process

Willing to dive into this a bit more? My personal experience is different so I figure this is a chance to learn something new. The subjective character of typing experience for me is like riding a bike, driving a car, walking, or speaking. That is to say that unless I'm mountain biking, off-roading, on a balance beam, or trying to say a tongue twister, for the most part I'm unaware of the intention-execution-results loop, and the intention-results loop is all that consciously exists. My paltry 50-60wpm doesn't feel like an impediment to putting thoughts into text, but maybe others feel differently.


> My paltry 50-60wpm doesn't feel like an impediment to putting thoughts into text, but maybe others feel differently.

That may be high enough that you don't have the impediment, but when you work with someone who types down around 10-15 WPM... You can definitely tell they're spending a lot of brainpower on hunt-and-peck instead of the problem they're actually trying to solve.

Basically there's diminishing returns on increasing WPM and you may be at or past the tipping point. I expect it'll be different not just for different people, but also for whatever they're currently working on.


The thing is, when a programmers starts to generate input, the problem is already solved, he is merely putting the solution into whatever representation the tools demand.

I am not defending 10-15 WPM here ofc. just statingt that the majority of work happens before we start typing in code, and after we typed it.


It's a type-think-type-think cycle (or type-test-think-type-...) and the more time and effort you spend on the typing part, the more it distracts you from the rest. Typing for a few minutes isn't any better than compiling for a few minutes before you can test. It just interrupts the more interesting work.


> It's a type-think-type-think cycle

Depends entirely on the individual developer. For me its think-think-think-think-drinkcoffee-think-type.

And after the first iteration, the typing to make changes is usually miniscule compared to the initial amount.


That's due to unfamiliarity with the input device. And if they end up having to use it several hours per day, this slowness will go away in a few weeks.


For me typing is the frustrating part, because when I know what I want the code to look like, having to type it in is just an annoying obstacle I have to go through before I can move on to the next thing, whether that's testing or thinking about the next bit of code or whatever. So yes, I do find it to be a distraction and frustration. Kinda like having to type a long random password and then dig out your phone and enter a 2FA code and then navigate a hierarchy of whatever before you can get to do the thing you're really after.

I always wish I could type faster and with less effort, or -- better yet -- just wish the code into the shape I want. ~100wpm touchtypist here.


I frequently encounter typing speed as a disruption to my thinking process while developing my personal projects. My personal projects are developed interactively in the REPL - I'll make a few small changes to an existing piece of code, re-run it, and observe the difference in behavior before repeating. In this process, I usually conceptualize code significantly faster (2x-5x) faster than I'm able to type it, which translates to wasted time, where I know what code I want to write, but haven't written it.

Now, from a philosophical point of view, I believe that it's possible to have programming tools that allow a 60wpm user to program at the speed of a 120wpm user without those tools (e.g. autocomplete allows you to type 10-character identifiers with significantly fewer keystrokes) - I just haven't finished building them. That is, I believe that having a slow typing speed shouldn't impede programmers, but does with our current tooling.


The easiest way to experience this is to remove 3 key caps from your keyboard eg. from T K and C. It will destroy your flow.


It allows for minimal disruption to thinking process, close to 90% of the time even when coding is not actively typing, but when typing is subconscious and fast it removes the potential of where it interrupts the thinking process.

This arguments wholly arbitrary. Maybe someone who types slowly will put a lot of thought into each word and create fewer bugs. Who knows? Why argue random claims you pull out of the air?


Very early in my career, there was someone on my team that did hunt and peck. It was painfully slow, and definitely impacted their productivity. It was honestly painful to collaborate with with them because of this.


> painful to collaborate

National Slow Talkers of America

https://www.youtube.com/watch?v=Qvrh73BVraE


> 1. It allows for minimal disruption to thinking process, close to 90% of the time even when coding is not actively typing, but when typing is subconscious and fast it removes the potential of where it interrupts the thinking process.

I agree, and even wrote an article about that a few years ago. :-)

https://github.com/Dobiasd/articles/blob/master/how_touch_ty...


Autocomplete makes programming very different from normal typing. I don't need to type fast, I use my keyboard to trigger lists of options that I pick from. Modern programming is a contemplative interaction with the IDE/language server, normal typing should only come into play when you write comments.


Autocomplete is just a very small subset of modal editing paradigm, the entire editing process, both typing and key shortcuts (i.e. picking options) should be treated as the same.


But in this case WPM or similar are completely useless measures. With the right tooling one can pump out an extreme amount of code in a few minutes. It totally depends on the programming language and the available tooling, templates etc. You don't need quick fingers for programming. Depending on the verbosity of the programming language and if it is statically typed you are much faster in pumping out characters.


Very true for sysadmin. I can type pretty quick, but I developed a habit that when doing risky things I do a "air enter" -- I tap above the enter key, stop and think "am I sure?", before actually pressing enter. It's served me well.


When I'm about to run certain commands - any dd, most database modifications, rm, etc... - I physically lift my hands up away from the keyboard while I reread the command 2-3 times before allowing myself to lower my hands and hit enter. I think the physical cue helps the mental side.


Haha brilliant.


I think things like this are not “everyone must use this new better way of typing,” but more “hey there may be a better way of doing this task.”

I am always surprised, though at this point I shouldn’t be, that there is always pushback against any attempts at improving the status quo when it comes to typing speeds on HN - as though the creator is attacking all of us with lower typing speeds personally…

From another perspective: sure you might speed up only 0.5% of your workday - but how is that a bad thing?

Repetitive stress injuries aside, even if you only spend an hour a week typing (I suspect it’s honestly more) then if you end up increasing your typing speed by double you’re still saving yourself 25 hours a year. Assuming my a career of 35 years that’s 875 hours and you increase your time fighting imposter syndrome by 0.25%.

Scale up as appropriate for how much time you actually spend typing.


The pushback is not against the improvements themselves, but rather at the claims of what the improvements solve, and the criticisms of people who fall below some threshold at some specific task, with dubious claims about the effect on their overall performance (without actually measuring the overall performance).

It seems to be a common theme among tech people to unfairly extrapolate small, vaguely related things to judge big things, even to the point of rejecting job candidates or firing employees over one of them and then crowing about it on social media (for example the "don't hire losers" post a few days ago).

If someone doesn't like how I work, then don't watch me work. I get paid for the fruits of my work, not how I get there.


> …the criticisms of people who fall below some threshold at some specific task, with dubious claims about the effect on their overall performance

It’s unclear to me where this criticism is coming from.

> If someone doesn't like how I work, then don't watch me work. I get paid for the fruits of my work, not how I get there.

Similarly it’s unclear to me where there is anything about anyone measuring how you get your work done. That seems unrelated to the article at hand that only passingly mentions any work context (only of a co-worker making a claim about their own effectiveness) while spending most of the article on games and typing competitions.


> I am always surprised, though at this point I shouldn’t be, that there is always pushback against any attempts at improving the status quo when it comes to typing speeds on HN - as though the creator is attacking all of us with lower typing speeds personally…

I don't think it's a general attack because people are personally offended; most of the criticism right now seems to be that the headline claim is materially untrue. Rejecting snake oil is healthy and reasonable even if there's value to be found.

Edit: Perhaps more succinctly, people aren't pushing back against improvements, they're pushing back on deceptive claims.


I agree with you here entirely. I don’t really see how this is transformative compared to any chorded typing system.

The specific op I was replying to was saying

> …and 0.5% actual typing, I think I am quite safe with my good 'ol QWERTY.

And there were a host of similar comments initially.

100% on fighting snake oil. 0% on pushing back against people trying new things that might improve the status quo


Ah yeah, that's fair. I could defend that as cost/benefit analysis, but I'm with you - let's push the limits and worry about utility evaluation later! (I literally have a chording keyboard on my desk right now, so I'm quite serious when I say I want to push the limits:])


What hardware/software are you using? I’ve looked into it a few times, but the cost benefit hasn’t been there for me for doing the research without knowing anything about it yet.

Do you code with it as well, or is it primarily chat/browsing/etc?


I own a https://www.gboards.ca/product/ginni , largely because it was the smallest/cheapest option to dip my toes in:) I don't do much of anything in it right now, because I'm still learning to use it (currently at the "I have to think about each letter" stage), so unfortunately I can't say yet. I expect to end up coding with it, likely with a custom dictionary, but we'll see how it develops.


Hi, "specific op" here. I'd be interesting in knowing how what I wrote above is a "pushback against any attempts at improving the status quo".

I said I spend not a lot of time typing. Therefore, a device that is aimed at improving typing speed, has a low impact on my productivity.


I might have read too much into your use of the term “safe” as being indicative that there was some danger or concern.

I agree with this sentiment: > I said I spend not a lot of time typing. Therefore, a device that is aimed at improving typing speed, has a low impact on my productivity.

Though, setting aside that efficiency =/= effectiveness/productivity, I think the spirit of the article is that if a coworker is typing a lot, that may claim that their efficiency has increased greatly, not that every single person will see equal improvements.

FWIW: this is being typed with two thumbs on an iPhone because that’s as efficient as _I_ need to be right now ;)


Medical transcriptionists once used a specialized keyboard to quickly describe spoken terminology. "Shorthand" has existed for a while and I imagine is a dying art at this point [1].

I think most cases where you needed actual really fast typing were covered long ago and the number of careers or situations where a person needed it have been declining for a long time.

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


My point is more of it being not a bad thing to improve on.

Most of the cases where you needed actual really fast travel have been covered for a long time and the number of careers or situations where a person need to travel fast have been declining for a long time, but I think we can all agree faster travel times are better.

Side note: stenography and chorded words with something like Plover (similar to the above) is where really fast typing usually comes into play for. I have not looked into any of the above because the pain of changing my habits hasn’t been worth the benefits to me yet.

I’ve considered it when I had to do interview transcripts, but ultimately I didn’t want to make the investment. I hope future generations are able to learn on something more designed for contemporary use than QWERTY and classic keyboards - I know my wrists have thanked me for moving to a split keyboard for the ergonomics alone.


Even if you're not wrong, this is a disappointing attitude. There are so many reasons to want to improve user interfaces - reducing RSI, reducing error rates, increasing speed, etc. I love seeing new designs where people are trying to improve the state of the art even if I don't plan to use this immediately .


Most of the advantage in alternative layouts is improved comfort and less injury-prone finger movements. Faster typing speed is basically a nice side effect.


The act of coding itself is also not a constant flow of characters from fingers to screen. You stop to take a moment, go back, correct something, next line, pause to think, type out a line etc. I have never used a device like this but I'm pretty sure I type at roughly the speed I think, at least when I'm writing programs. General note taking, eg. jotting down ideas, might be different, where I do find myself thinking faster than I type.


You forgot typing documentation, emails and chatting on Slack.

Also, you might be underestimating how dumb it looks to both technical and non-technical people when a highly paid engineer can barely type.

Not that I've seen many who type slowly, probably because that's something you just pick up after decades of work.


An old very experienced co-worker once said, "code is written once but read a thousand times".

Most of the old wizards I worked with hunted and pecked, looking down constantly.


I had the opportunity to work with a guy who studied CS back in the day when punch cards were still used. Sadly he left into well deserved retirement a few years later.

He typed slowly, not hunt&peck, but really slow compared to the other engineers.

Didn't matter. The guy was a wizard. Ever seen someone implement radix sort in assembly or debugging a microservice using nothing but nc and echo? Without looking up ANYTHING either in a book or online? That kind of guy.

Working with him, I sometimes felt like Mickey Mouse in that scene of "Fantasia: The Sorcerers Apprentice", when the Master came back, surveying the flood, and fixing everyting in seconds just by waving his hands imperiously :D


your missing replying on hacker news, chatting on slack, chatting on whatever othe rchat program you use, etc.


That's the real use case here, get your comments out faster on hacker news than everyone else!


Where else does one learn to _really_ type quickly besides hurrying messages into a game chat?


I wonder how much of Carmack's time is spent typing. It'd be fascinating to get an idea of that.


> and 0.5% actual typing

I wonder how often Donald Knuth types.

And he is arguably among the 5 best coders in the world.


Doesn’t this argument go both ways? Given you have so little time left for actual coding, you should be squeezing every last bit out of it.

Depending of course if your meetings are necessary discussions, or just pointless constraints.


The argument isn't that there is so little time to type code, but that typing code requires so little time compared to everything else.


And most emacs users write programs that type repetitive patterns for them anyway


Holy wow, you’re missing a lot of quality design doc time in this breakdown


The amount of time I spend typing in Slack alone is FAR greater then 0.5%, plus writing docs, tickets, PRs, writing code, etc.


You type a lot more than 0.5% of the time.


It's a trade-off between spending some time thinking about what should be written, and then typing a little - or spending all the time typing, typing, typing, and rewriting.

I don't think I've ever been in a job where I was limited by typing speed.


Then you should definitely consider switching keyboard layouts to reduce RSI. I did, last week. Going from 80wpm to 30 wpm is definitely a job impediment (in the short term). So many things are keyboard driven:

- communicating with peers over slack

- writing code / git commands

- writing email, comments in HN, documentation, etc.

- taking notes in meetings

- all the browser shortcuts

- all the CLI readline shortcuts


Yes. I've love to be able to type fast enough to take notes so that I don't have to try to read my chicken scratch handwriting and figure out what the hell I was trying to write, especially when referring to notes where I don't have any recollection of what the contents were.


I never thought about RSI, but also I've started having pain/weakness in my right hand last month (that, thankfully, has mostly gone away). The doc said it was tendenitis.

I feel like excessive swipe-typing with my thumb may have contributed more than typing to that; but I also play keyboards and it was scary for a second to have that affected.

I got myself a vertical mouse; will consider using another layout (which one do you use?). I do type at 75-80WPM (including time to fixe typos), can that really cause RSI?

I don't know if I'll be able to pull off even 20wpm with another layout, but looks like it's the next step. Thanks for suggesting it!


In my experience, typing early and rapidly iterating and modifying, i.e. thinking on the page, is FAR more effective than spending tons of time in your own head trying to solve the problem elegantly before you ever type anything.

The exception to this might be spending time thinking about the very high level structure you want, since the initial structure you start with can lock you into pattern of trying to solve the problem that might not be the most effective.


Not an amount of time where typing more efficiently is going to matter. Once you hit intermediate competency it stops mattering.


Yes, but typing speed is still not too important.


There is some threshold where it absolutely impacts productivity. If you haven't seen this in action, you're lucky, I've ran into more than one Engineer who is otherwise very smart but somehow never learned to type.


For sure. If you're a hunt-and-peck typer, it makes meetings where we're waiting for someone to finish typing unbearable. If you're even somewhat competent at typing, then I don't think it's a big pain point.


I'd say that threshold is closer to 40 wpm than 160 wpm.


I'm glad I'm a fast typer because I can spend less time replying to pointless emails.


I also spend less than 15% of my time drinking coffee.


I'm typing even less these days with tools like copilot.


For some reason I don’t like any of this automation tools. I find VSCode unbearable even without copilot. Too many things happen when I type. I prefer linting and that’s about all I want to write my code peacefully. I use vim in iTerm and it’s all I ever need. Jump off to Pycharm to debug if needed.


About 3-5% of the time, copilot poops out an unexpectedly clever code block that makes it worth it to me, if nothing else but for the surprise factor.

For the remaining 95%… I’ve configured it to delay doing anything until 10000ms passes so it doesn’t get in the way


I use vim quite a bit too, just sayin...

https://github.com/github/copilot.vim


Oh boy..., I guess there are people who find this useful. I respect "To each their own" and "You do you".

I also like using Sublime Text with vim keymap. I prefer the low latency of editors over features of big IDEs. Sometimes, I do use PyCharm or CLion for debugging since that experience in vim sucks. GDB is terrible in CLI.


> Keen also claimed that he’d won several online gaming competitions using the CharaChorder, which he says brings up an interesting ethical dilemma. “I’m not sure if there’s any restrictions on what keyboards you’re allowed to use,” he said over a video of him playing Super Smash Bros. Ultimate on the Nintendo Switch, using his CharaChorder running through a XIM-style adapter.

As someone who's somewhat familiar with eSports, I'm having a hard time imagining what game would let this style of keyboard provide a substantial advantage, and I'm very skeptical of these "online gaming competitions" he supposedly won.


There has been conversation happening regarding tournament play legality of digital input controllers for the Smash Bros games, most notably Super Smash Brothers Melee. https://en.wikipedia.org/wiki/Smash_Box_controller


Yes, but things like regular keyboards, or hitboxes already exist. How does this new thing improve on those?


Check out https://youtu.be/Lw1tcqbFwN4?t=867 which talks about some of the advantages of stickless controllers. That link points to a particular timestamp, but the video as a whole is pretty interesting IMO.


I'm aware, I just don't see how this would provide an advantage over already existing methods like hitboxes, or regular keyboards.


Your original comment read "keyboard provide a substantial advantage" without specifying an advantage over what, so I assumed you meant an advantage of the most commonly use controllers, not advantage over other stickless controllers. Even ignoring that, stickless controllers are still very much a minority, so an advantage over traditional controllers would still be a large advantage over the pool of competitors especially at lower levels.


If it provides no more advantage than an existing non-standard controller, than this news piece is a complete non-story. Or at least, that bit of the piece is a non-story.


Tetris. There's some discussion on HN a while back on different ways of tapping the controller to make quicker turns. It will be interesting to seenwhat CharaChorder can do.


All of that has to to exclusively with techniques for improving speed when playing with the traditional controller. As soon as you can use another input device there is no need for that.


That was just for on the NES version using the standard controller there IIRC.


Kings Quest I, Zork, or Adventure maybe. :)


nethack! although it would only be cheating in the "fastest death" category


I'm fairly obsessed with alternative HIDs. I bought the Tap Strap 2 but never broke past the initial learning curve.

I'm going to buy a CharaChorder now. I'm also a musician, so I'm really interested to see what types of functionality I could map, given the additional interface dimensions. However, I'll need to test the input lag, typically anything > 20ms is rather noticeable for live music; if they have esports as a target market, hopefully they've already accounted for that.

I type ~80WPM with 90%+ accuracy, but I don't find that typing speed is my blocker, rather the speed of a coherent thought. Maybe typing at ludicrous speeds will channel some deep stream of consciousness?


> I type ~80WPM with 90%+ accuracy, but I don't find that typing speed is my blocker, rather the speed of a coherent thought. Maybe typing at ludicrous speeds will channel some deep stream of consciousness?

I agree with this, at least to some extent. If I'm rote copying something from a text (e.g., from a typing test website), I can achieve 130-140wpm easily. If I'm creating original thought or trying to actually compose a reply to someone, I think much slower than I can type. It takes a lot more effort to compose logical sentences that make sense on paper than it does to type them, in my experience.

What really makes me want to try these alternate input systems is the allure of being able to type while walking outside at speed, which is something I can't do on a smartphone. CharaChorder seems like it's nice on a desk, but the same could be said for a plain stenography keyboard/machine.


>If I'm rote copying something from a text (e.g., from a typing test website), I can achieve 130-140wpm easily.

Weird. I have the opposite problem. I'm usually in the ~80-90 range typing something I'm reading. I find it way slower to read the text, the repeat it. But if I'm typing from my brain I'm way faster because I can skip the reading step.


> type while walking outside at speed

Yes, exactly this. I didn't expect the Tap Strap would be refined enough to serve this purpose (too much ambient motion) but I like to keep looking for something novel.


Feel like when I am speaking somewhat conversationally (e.g. IM or even HN), I can think much faster than I can when doing something like Typeracer. It could be a result of me growing up with AIM and MSN Messenger, or maybe just a result of the fact that I tend to talk really fast regardless, but I almost can view typing as an extension of my brain, and as a result I do actually feel like my inability to type faster is a limiting factor.


Actually, I think that is completely normal. When writing with pen and paper, it's so slow that you can't help but think ahead. You're thinking in parallel because you have time before you write the next clause or sentence. You might scratch out a word or sentence or two, but overall your thoughts will probably be more coherent and well-considered.

Even typing at 150 WPM (actually fast) is about the same speed as dictation. Anyone who's done extensive dictation knows how slow that actually is.


This is my experience exactly. I speak fast, and typing for communication just feels really slow, since I'm thinking at my normal speed.

Typing commands or programming isn't so annoying though, because my speed of thought is slower then... Except for the occasional long commands such as 'sudo systemctl restart asterisk' or something.


> I don't find that typing speed is my blocker, rather the speed of a coherent thought.

In my case (which probably shares at least a few similarities to yours), I (a) have some tasks where I really do need to type fast (usually either when I'm transcribing my voice notes to text, or when I've already formed an idea into a sentence in my head and just need to get it out) and (b) have highly irregular rates of thought, where sometimes I'll have things I want to write at 300 WPM, and other times I don't have any ideas for minutes on end.

In the former case, faster is always better. In the latter case, while you might not be directly blocked on typing speed, it does allow you to get the typing out of the way faster, so that you can then move on to more thinking, or another non-typing action.


I type ~80WPM with 90%+ accuracy, but I don't find that typing speed is my blocker, rather the speed of a coherent thought

I’m on typeracer.com with an average typing speed of 154 WPM with 99% accuracy. I don’t find my thoughts to be limiting for typing purposes.


I don't really know how people get past 120-130wpm with sustained accuracy. I find that I trail off around 100-110wpm. It really depends on the text too. Some text with a lot of difficult double repeated letters, fancy punctuation, etc. in a row or hard to read words - not so great.

I just did the practice twice. Once at 100wpm (keyboard twister kind of text). Once at 130wpm (easy sentences). So much variation just from the practice text alone. This is also on a macbook pro keyboard literally on my lap - which I fucking hate and find horrible to type on and mess up on all the time. (I think it is also dysfunctional/semi-broken)

I don't think I could ever hit 150wpm. I just don't see how it's possible with a normal keyboard - at least for me. Never seen anyone do it sustained either. Must be some <1% skill - as I'm the fastest typer of just about anyone I've met and I don't feel fast.

Speed of thought is mostly the issue for me too though even at 130wpm. If I am saying things faster than that - whatever I am saying is probably not worth reading outside of a chat conversation.


I don't really know how people get past 120-130wpm with sustained accuracy.

Speaking as someone who can type in the 130-140 range and has gone over 200 in short bursts --- the keyboard makes a huge difference. Look for one with a low actuation force, short distance-to-actuation, and a "bouncy" feeling that helps your fingers return. I'm using a cheap no-name rubber-dome keyboard, but it's definitely on the softer side compared to most others I've used.

Unfortunately, the only searches on Google for its model number (KM-2601P) are the posts here where I've mentioned it on other keyboarding articles: https://news.ycombinator.com/item?id=24761580 (Don't you find it disturbing that a keyboard which was probably manufactured in the millions of units has literally only one person on the whole Internet mentioning it!?)

I also have a mechanical keyboard with clicky Blue Alps switches, which is impressively loud and satisfying to type on, but reaching 120 on that one is far more difficult; and I've also typed on a Thinkpad (pre-island style) keyboard, which has been praised as one of the best laptop keyboards, but maintaining even 100 on that one is very tiring --- it has far too much actuation pressure.


My ex-wife used to say that she could not randomly mash the keys on a keyboard as fast as I can accurately type. I've had people ask me if I'm human based on the speed of my typing. But honestly if you watch YouTube videos of typing competitions - there are lots of people that are 160+.

I first learned to type in 1982 and that really accelerated when I got online in 1985. I did not touch type at that point - I just used 3 fingers on each hand. The first year I learned to touch type my typing instructor typed 60 WPM and I typed 75. Then it just went up from there after 4 years of formal typing instruction.

For what it's worth, I use a Das Keyboard mechanical keyboard.


You're making me feel a little inadequate about my current record of 97WPM with 99% accuracy.


On the other hand, it's making me feel pretty good about my thinking speed. ;)


I don’t think my problem is being able to get the words from my brain into the computer fast enough. My typical HID use case is composing meaningful written communication, a process which is unfortunately much slower than 500wpm.


I consider typing speed the I/O of the operation. Sure there's some processing time while you think of what to type, but regardless of how long that takes, typing will slow it down further. I'm not perfectly parallelized in thinking/typing pipelines.


I agree with the analogy, but for me (and I think most people) processing time far exceeds IO. Once you get to a good enough IO speed, it doesnt make sense to optimize further, as the returns are diminishing.


For programming, I can see that, definitely. But when I'm writing an essay, there are many times where I can't type as fast as I can think, and it drives me nuts. I end up resorting to using a voice recorder and transcribing it later. Because sometimes the ideas can come out quickly, but it's still easy to forget them if they're not written down.


For programming I only see this if you have a language with a lot of unnecessary overhead (not in the syntax but in what you need to type out) no (good) IDE or only solve mostly memorized leet code problems, or only write pretty brain dead code the 100ed time (in which case you could optimize it away with code-gen).

For other thinks I don't see this, not because I think slower then I type. But to some degree thoughts and typing are out-of-sync and while each though is faster then a typing, for much code you have one thought about how to type it, but also many more about contexts of your solutions and interaction with other code and what you do next etc. you type. And I don't think increasing typing speed would change this much. Except if I increase it to a point where I now need to fully focus on typing, which would be counter productive.

TL;DR: I type and thing, not type then think then type. (Though biologically seen I maybe don't do it actually in parallel but micro-task like how multi-threading on a single core non SMT system works, but it doesn't matter much for the end result.)


I think the more significant I/O trend would be ai guessing your intentions. The program will know the user so well it will be effectively mind reading.


Yeah. The sheer amount of button presses to write a 1k line script? That'd be roughly 10,000 words let's say, so 500 wpm would bang that out in a mere 20 minutes. Meanwhile the thoughts to compose 1k lines of code? That could take weeks of reasoning to deduce the proper logic.


These days I type at about 80-90wpm which is not as fast as I once could (there was a point where I could achieve 130wpm on Dvorak but I was both younger and had less of a life). As is, though, it's very noticeable that I can type faster than I can actually compose text. I've also found that typing very fast tends to lead to a much higher rate of what I call "wordos," in which you mistype by completely swapping a word with another or omitting a word... these feel like symptoms of a sort of "buffer underrun" when typing that doesn't get noticed because I'm already having to move on to composing the next sentence.

So yes, I think there's a significant effect of diminishing returns on very high typing speeds, and I'd take a guess that it starts really kicking in at something not that high like 60wpm.


It’s surprising to me that stenography-esque chording hasn’t had a breakthrough moment among programmers. Most programmers love trading hotkey/shortcut recommendations, and software from vim to blender to photoshop show users have a deep interest in using shortcuts to become more efficient.

Hardware restrictions notwithstanding, if my keyboard driver/OS had convenient support for importing standardized maps that let me type `h-down, e-down, release` instead of “hello” I’d gladly (slowly) start using it


I also love to improve my input methods.

I'm still not convinced steno works well for code. One because I haven't seen a convincing example (all the examples I've seen were youtube videos where it's obvious someone's working hard trying to remember the inputs they just fed into their dictionary to produce a specific piece of code). Two, because I think code is by nature very different from natural language and you have to be extremely precise with every dash, underscore, capital letter and symbol, and there are lots of identifiers that are abbreviations or not natural words at all.


I feel this is already what we get with IDE auto-completion and snippets.

Then some people push it further with personalized dictionaries and TextExpander like extensions at the OS level. I am not sure how much a dedicated hardware device would be an improvement, at least it wouldn't warrant the learning curve and portability issues to me personally.


This is interesting to me not because of the typing speed but because of something almost everyone will find weird: As a developer I'm not a fan of typing due to health reasons.

The long story: I have an undiagnosed disease/syndrome which, among other things, makes my fingertips become red and cold during winter season, and it gets worse when I'm typing. So in the summer I can pretty much type all day. It's not that my hands don't get cold while playing video games on my Playstation controller, they just get worst. I assume that the impact on every finger makes it worse, because I can wear all the layers of cloth in the house + have the air condition turned on and all I can get is my face and ears all red, but the fingers still cold as ice cubes.

From my standpoint, I see the potential use as an accessibility tool. If I can write words and avoid the step where my fingers suffer the impact of every keypress that's a big win for me. It's not like I can't type (obviously) but I'm wondering if I could train myself to code using something like this to avoid discomfort. I once entertained the idea of using my PS4 controller since Steam uses something similar.


You probably already got checked for this, but just in case: this sounds like a textbook Raynaud's syndrome presentation.


Try going gluten free for 2 months. Seriously. Family member had issues like this and tried many docs. No results after lots of testing. Tried going off gluten for other reasons and everything cleared.

To go gluten free, only eat foods that are certified gluten free. Anything with any kind of grain that is not certified likely has cross contamination.


Envavo Heatbuff was invented to help this. Unfortunately they seem to have gone out of business


Has anyone tried this with coding? It looks like this is very focused towards writing English, which is fine, but as someone who spends a good chunk of my day around code editors, a keyboard that is crap for code editing is a non-starter.

However, if it's good for that, I will probably be buying one.


> chorded typing allows the users to input several letters at the same time and have a computer program generate a predicted word. Instead of typing h e l l o, a chorded typer mashes h, e, l, and o to produce the word.

It sounds like it relies heavily on autocomplete, which means it's unlikely to be useful for coding instead of just English text.


No, that means it's likely very useful for coding instead of English. You just need a custom dictionary. Code has a smaller vocabulary than prose, even with the domain-specific words that get used as identifiers . You could probably do pretty well with a dictionary that just has all the keywords in your favorite language, plus some common variable names. Heck I bet it'd be a win just to make the punctuation used in code have convenient chords.


Seems like it would be better to have the device be dictionary agnostic and leave it to your editor do know whether you're writing a comment, or code, or a string containing sql...


I wonder if it would make sense to latch into the syntax highlighter for something like Vim. Most syntax highlighters have a reasonable understanding of what context you are currently in, so conceivably the device could, for example, see that I'm inside a comment and revert to vanilla English, and then see that I'm back in code and change to F#.


I think it would. It seems like you'd want some amount of symmetry between contexts, otherwise learning separate ones would be too difficult.

Like the gesture for python's "create a for loop" would align with the english word "repeat" and in functional languages it might map to some manipulation that sets up recursion with a shortened list.

Actually coming up with a list of such a list of abstractions for app developers to map actions onto sounds like a dissertation for some psychologist somewhere.


He posted an earlier video on his TikTok account showcasing the device's coding capabilities. I don't know if the videos he posts are entirely reliable (eg. as another commenter mentioned, he claimed a 500 wpm typing speed by typing one memorized sentence). I do know that it needs autocorrect (built-in) to work at high WPM which doesn't seem great for coding which might use nontraditional words and spellings. Personally it seems like it would be too great a learning curve for not enough benefit, given that no software is designed for devices like these. Maybe if this somehow takes off in the future.


I'm gonna bite the bullet and buy one, I think. I'm ok with a "good enough" coding experience, and I never bought myself a Birthday/Xmas present last year. This will work.


i don't think there's anything to gain here that you don't already get with good autocompletion.

i wonder how a trained stenographer on a chorded keyboard would compare to a trained typist with a good english language model backed autocomplete and a ui built for speed on a classic qwerty keyboard.


I'd be ok with it being "as good" as my current coding setup, at least if it improves my other typing speed. A vast majority of my correspondence these days involves me typing (as I've made , so being able to speed that up would still be cool, though I doubt it's going to be a categorical difference in my day-to-day life.


curious how it will work out. you'll have to learn new chords for the standard alphanumeric characters and then a chunking/chording scheme where the space of chunking schemes and associated chords for computing and coding is vast.

my understanding is that classic stenography is actually phonetic. so the chords match up to phonemes or phoneme like chunks which are then postprocessed to reconstruct english text.

a new approach using english language could be more data driven. a simple mapping could be one chord to one word with words sorted by frequency and easier chords assigned first. more complicated approaches involve chunking up the words into frequently used chunks and then doing the same.

code is harder, there are frequently reused strings, but they change from technology to technology. overall there is far more entropy in computing than english on a character by character level, so designing a chording scheme that is more efficient and isn't tied too much to a specific domain is actually a really hard problem, especially once you consider that the effectiveness of a given scheme is not only a function of how well it fits the problem (how often the user is actually entering things the scheme was designed for) but also how well users are able to learn the vocabulary of chords and the dexterity to execute them quickly.


> Instead of typing h e l l o, a chorded typer mashes h, e, l, and o to produce the word. With practice, it’s much faster

If this works like swipe typing on phones, you'll type blazingly fast, but you have to retype each word several times until the AI finally guesses the right one.


Even better, you have to create a keyboard shortcut for each word.


> Instead of typing h e l l o, a chorded typer mashes h, e, l, and o to produce the word.

So, no, it won't be typing "cadddr" for you in your Lisp code when you mash c, a, d, r. Probably, "card".

Of course it's banned from typing competitions; you're not actually typing the word; the algorithm is.

From the contest's point of view, it's morally equivalent to a regular QWERTY keyboard equipped with firmware that corrects typos.


Chording input devices hardly count as keyboards. It's not even that it's fast, it's that it's an entirely different paradigm of input. Stenography is an entirely different skillset than typing.

According to its website, the device mentioned in the article basically serves as a device that does autocorrect on top of traditional chording, which is quite literally cheating in typing competitions.


> article basically serves as a device that does autocorrect on top of traditional chording

This is normal for chording input devices. Both old-school stenography machines and modern steno software like Plover are context-sensitive to consecutive inputs.


Years ago I had a friend who was in IT, but had a side job doing data entry.

When he got a new set of data to do entry on, he would develop custom data entry environment specifically for that data. He said it effectively gave him some large multiplier in typing speed, making a minimum wage job fairly profitable for him to do.


That thing isn't really a keyboard is it? It doesn't have keys and it isn't a board. It's more a joystick controller.


Those joysticks do have switches at the ends (that click) though…

I don’t know, it’s a very philosophical question


Is it me, or does this article look suspiciously like an ad?


On the one hand, yeah the "controversy" here seems rather dumb. It's banned from a website where people compare their typing speeds, big whoop. And the discussion about its possible advantage in games like Smash is similarly stupid, seeing as there already are similar controllers.

On the other hand, it admittedly seems like it could be a nice development in consumer stenography. I'm only aware of open-source stuff like Plover, which I never got around to trying.

I admit that if the sensationalism and fake drama were stripped, the article would probably read like a product endorsement. So I guess me saying, "I'm OK with ads if they're cool" is itself a dumb defense.


I'm more curious what the ergonomics are like on this device and how it compares to keyboards like the Kinesis Advantage for preventing injuries.


Maintaining such typing speeds look like a recipe for burnout to me. I think I could even see the CEO's hands trembling in the TikTok when he lifted them from the device after demonstrating 500 WPM. It certainly incurs a massive cognitive load, perhaps similar to what professional classical musicians can pull off—and they don't have invent the score at the same time as they are performing.

Someone capable of decent touch typing on QWERTY will hardly benefit from the additional typing speed IMHO; at least not when factoring in the time spent learning, and the inevitable decrease of typing speed on QWERTY keyboards which will be very hard to avoid when working in a team or multiple locations.


> they don't have invent the score at the same time as they are performing

Do you mean “improvising”?


No, I was referring specifically to classical musicians who typically don't improvise.

Improvising musicians will almost certainly not sustain high note counts except in repetitive patterns. If you have an example that proves this wrong I would very much like to hear it ;)


> Keen is experimenting with a modified version of the device that he thinks might even allow babies to type and communicate similarly to the way some babies communicate with sign language.

I feel like this warrants its own article. Jumping from understanding but not being able to reliably speak words to chording without reading or letters in between sounds like a really interesting direction. Are the chords chosen to relate to orthography? To sound? To semantics? How would you teach it? A caregiver can easily demonstrate a sign. A chord is a lot less visually salient and harder to demonstrate.


...he did win another typing competition using the CharaChorder to hit 267 WPM with 76% accuracy.

Isn't this accuracy abysmal?


Yes, IIRC stenographers in the US are required to type 225 WPM with 98% accuracy.


If my accuracy was 76%, I'd be upset. I'd definitely work on accuracy over raw speed at that point.


Everytime I see thread on typing speed it makes me want to re-surrect this: https://github.com/janmechtel/rockettype

I wanted to start small with measuring real-life typing speed. And then work on two epics:

1. more aggressive Auto-Correct: why can you swipe "messy" on your phone and everything just works, all typos corrected fine, but the same doesn't work on PC?

2. typing technique trainer, webcam record your fingers and learn to use the right fingers for the right keys

Is anybody interested?


> "Chorded typing allows the users to input several letters at the same time and have a computer program generate a predicted word. Instead of typing h e l l o, a chorded typer mashes h, e, l, and o to produce the word. With practice, it’s much faster"

There's nothing wrong with this, and many of us probably use similar techniques daily (macros, shortcuts, etc.), but he's not typing 'words per minute' without computer help. So, makes sense it's considered 'cheating' for comparison purposes.


A lot of people are expressing doubt about the utility of faster typing speed--rightly so, since that is the focus of the article. However, as a prose writer, I'm actually quite excited about this kind of device. When I'm typing 50,000+ characters of prose per week, every week, that's a lot of finger strain with a QWERTY keyboard--and I feel it. If a device let me accomplish that work with less net impact on my hands, even if it wasn't any faster, I would embrace that in a heartbeat.


Split keyboards have done wonders for me. But a different keyboard alone won't solve the strain problems. A combination of using an ergonomic typing device, exercise, regular breaks and time off the keyboard are essential.

When you already feel the strain it's high time to do something against it, damage might already be done. Carpal tunnel or other inflammations are really painful and can take months to recover, during which your productivity will be quite low. Better to type only 40,000 characters a week and give your hands some rest, than squeeze 50,000 out of them until the damage is done.


Did you try voice transcription?


Back in the 90's I worked with a guy who used a Dvorak. Any fans still out there?


Posted above: Been using Dvorak for 16 years now, but I didn't necessarily "switch" as I had been using a memorized hunt-and-peck style with QWERTY and needed to force myself out of that habit by removing my ability to fallback to that practice (so I had never properly learned to touch type before starting with Dvorak).

I'll sing its praises until the day that I die - but it does have disadvantages in that most applications design their keyboard shortcuts and other inputs around QWERTY users. On macOS I can use the Dvorak-QWERTY Command layout to deal with this in particularly annoying cases, but there's nothing comparable I've found on Windows or Linux and that doesn't help with things like Vi.

Not to mention I frequently find myself having to change layouts back in forth in games. A lot of games published even today have a nasty habit of using the character code instead of the keycode for keybindings, and I've gotten tired of redoing bindings in everything just to avoid pressing Windows+Space to change my layout. Additionally, since my keyboards still have a QWERTY layout of keycaps it makes it challenging when I get prompted to hit "Y" (which doesn't get used for important things usually as it's a stretch for the index finger from the ASDF position) and hit F by mistake since I look for the "Y" keycap on my keyboard - ditto when something prompts me to hit "F" and I hit "U" instead (basically the mappings between Q/' E/. F/U C/J V/K are easy enough as they're used often and are within natural reach, but once it goes outside these I start looking at keycaps and screw up).

Anyway, I'll always highly encourage people to give alternate layouts a try and I'm never going to stop - but I'd say there's nothing wrong with using QWERTY either.


Yup! I learned Dvorak probably in 1999 or early 2000s (and I have a coworker who also still uses Dvorak).

The kind of people (like me) who got into Dvorak at that time would today probably learn something like Colemak DH, although there are many alternatives available nowadays and a big (discord) scene of folks optimizing and designing both their keyboards and the layouts they use on them.

I expect Dvorak will die with people like me, because anyone willing to switch away from Qwerty is probably better served by other alternatives.


I disagree with your assessment that Dvorak has been obsoleted by other layouts. I used Dvorak 1998-2001, and again since 2018. Before I re-learned it in 2018 I did a bunch of research on Colemak, Workman, and the rich set of other optimized layouts people have created this century. There are compromises in every layout, and I came to the conclusion that Dvorak was within spitting distance of minimally pessimal, other named layouts less so.

The caveat to my perspective is that I don't care where hotkeys are. I minimize mouse use, and with both hands on the keyboard it just doesn't matter very much. I've used emacs and vim with both layouts. My faded recollection with respect to emacs is that it was equivocal, and Dvorak is actually a bit better for vim.


Yeah, I used it for probably 6 years! Despite becoming very fluent I eventually came to the conclusion that the hassle of was not worth it.

This was in the early 2000s and I had to use a lot of Windows and Remote Desktop, so there were at least several incidents a week in which the keyboard would start making the wrong letters and I had to figure out why. It doesn't help at all that Windows' default layout-switcher switches using Ctrl+Shift so any key combo that includes that pair (including select-by-word!) will swap your keyboard layout too.

Sometimes I'd remote into a machine and after dozens of failed password attempts realize I was being treated to Double Dvorak, a much less well known layout in which the Qwerty -> Dvorak mapping is applied twice, due to the map being loaded both locally and remotely. This is all w/o third party software, literally nothing more than Windows just not coordinating with ... itself.

And finally, despite claims that Qwerty and Dvorak could be maintained at the same time, that wasn't true for me. The faster I got at Dvorak, the more speed and accuracy I lost on every other keyboard in the world I had to type on.

So now I'm typing this on Qwerty. It ain't as comfortable, but the number of hours I spend each week trying to get the computer to show which letter I'm pressing is now zero.


I tried learning Dvorak about 9 years ago, and I got good enough for IMing and emails and whatnot, but I found it extremely difficult to context-switch between coding-editor keystrokes and conversational keystrokes. I know it has its fans, but I could never really get the hang of Vim with Dvorak, and I also didn't type any my emails any faster than I was with QWERTY, so I abandoned it and haven't tried Dvorak since.


I've been using it for 15 years or so. One of the tricks I use for getting around the problem other commenters here have is that I have bound my keycodes to be qwerty ones and (virtually) moved my key positions. This means if you set the OS to qwerty and type Dvorak the correct letters will be input.

When working in offices I'd have my company supplied keyboard still plugged in as a guest keyboard so others could work with me seamlessly.

This also came about as windows used to have the most ridiculous behaviour of setting the layout per window, so if you changed it when someone else came over you'd end up in typing hell. They thankfully fixed this to be a global setting a few years back.


I use Colemak, but I don't really buy into any of the claims of improved comfort or speed. It took me somewhere between 1-2 years to get to about the same typing speed as QWERTY (somewhere between 120-150 WPM depending on the test), and I have completely lost my ability to type QWERTY without looking at the keyboard. As others have said it is a big hassle when working on other machines or through things like remote desktop.

That being said, I don't regret anything and Colemak is way more comfortable, for me. I never typed "properly" in QWERTY, I would use every finger on my left hand, but only two on my right hand. I started to notice some pain in my right hand and so I tried to retrain myself to type properly, but it never lasted more than a day because I was typing half the speed with proper technique. The only way I could force myself to use a proper typing technique was to just completely switch the layout.


I've used Colemak primarily for the past 5 or so years. However, the benefit is primarily in typing comfort and hand strain, not quite as much in speed, although it is probably marginally faster at its limits. I'm still fluent in QWERTY (I'm typing this message with qwerty to make sure it still works :)), so I can switch if needed, or if I'm using an unfamiliar computer (or a phone, etc). Moving to Colemak completely solved the frequent wrist and hand pain I got while typing using QWERTY. Others have had the same experience.


I've used Dvorak for more than 15 years and love it. I'm "fluent" in QWERTY too but am faster in Dvorak.

I also use it on my phone. Dvorak is terrible on swipe keyboards but I'm used to it.

It was annoying to learn but I forced myself to power though my initially slow typing speeds, and it was 100% worth it. I highly recommend it.


I use Dvorak Kinesis Advantage 2.

I switched to using copy and paste with the mouse with it since Cmd+c and Cmd+v are so hard to reach.


> I switched to using copy and paste with the mouse with it since Cmd+c and Cmd+v are so hard to reach.

I don't use it myself - but on macOS there is a separate layout called "Dvorak - QWERTY Command" that shifts the layout to QWERTY when the command key is held to alleviate this issue.


I switched to Dvorak when I bought a kinesis vertical keyboard. It's great. Much less strain.


This is probably the best advertisement they could get ;=)

Not saying that it's bought or anything, just that as a company focused on a value-delivering-product such articles are just awesome.

Especially when the article only focused on your good sites.

(No ideas about bad sites, but price ($250, not absurd for that market), compatibility with less usual hand-forms, and it not fitting well with a travel laptop setup (more size needed) are probably some).


Looking at the CharaChorder Lite version — can’t this be done via software implementation with a n-key rollover keyboard?

I’d like to try it but not $200 like to try it.


I agree. It really feels like it should be a PLOVER dictionary.


This reminds me of an old electronic organiser / PDA that was available in the 80s/90s that had just five (I think) keys for typing. Letters were formed by pressing the keys in combination. Skilled users could type very quickly on it.

I can't remember the name of it now. "Jaguar" is in my head, but I think I'm conflating it with the Atari console.


Anecdotal: Growing up my school had various computer classes, starting in elementary all the way through high school. In high school we had a computer proficiency class which was basically "how to use a computer for dummies 101" type of class and one of the competencies the teacher had us work on was typing.

Well me, having grown up with computers, I was beyond proficient and in particular was quite a competent typist. Being an engineer at heart (lazy) I realized I could rig the system and be just fast enough each week to win but also faster than my classmates to remain on the leaderboard. The teach was thoroughly pleased with my weekly progress and applauded my efforts. All the while I was busy gaming (Nerf Arena, Quake) on a LAN with some others in the school who were "computer proficient".


I wonder how effective this CharaChorder is at preventing RSI (repetitive stress injury). It still seems to have the same flaw as QWERTY in that it requires you to use your small finger muscles, which are more prone to RSI than say your biceps.

ps: Ugg, vice.com. I don't care for them.


twitching your fingertips left and right seems like a terribly unnatural repetitive motion to me, I would be concerned.


As a flute player who has experienced my right pinky locking up after passages involving moving between the footjoint keys, you're right to be concerned.


If this works--and I have my doubts--it will revolutionize the stenographic industry and bring stenographic recording to the masses. Furthermore, the article barely mentions the role of AI, and its role here is significant. It is AI and word/phrase completion that allow Chinese writers to approach 200 wpm and regularly hit 125 wpm. Only expert English language typists regularly reach 100 wpm. It wasn't that long ago that the ability to type only 40 wpm qualified you as a clerk/typist in the military or the government.

If it works half as well as the claim it would be an extraordinary advance. The issue is the length of time it takes to become proficient.


Steno devices are well-known to the speed typing community. Some competitions allow them, some don't. You can't bring this to one that doesn't for the same reason you can't bring a F1 to a stock car race.


Sidenote: Ghost in the Shell had people with cybernetic hands that would convert from ten fingers to dozens for super rapid typing.

Which was a bit weird in a show about cybernetics with tons of direct mind links: why not just thinktype?


If there's one scenario where I'd dread autocomplete-based/chorded typing is when coding or writing on a terminal. My coding usually comes _after_ some consideration about I want to write, and having to tack an additional "let's proofread what the software thought was close enough" step on top of that would make it seriously annoying. I already have an expansion plugin on my editor that allows me to insert potentially massive amounts of boilerplate for me if I happen to need to, and I already trust it.


Very interesting design and good logical extension of the whole chord keyboard approach which I recall the early https://en.wikipedia.org/wiki/Microwriter

One of those interfaces I've always wanted to give a go, but never really invested the effort in trying.

Though the prospect of breaking a finger and being locked out of your computer as your unable to enter your password; Is one aspect of chord keyboards that would compound a stressful day.


Donald Knuth used to mentioned that he went to secretary school to learn typing more than 80 WPM. Then he found that he couldn't think as fast, so he resorted to writing with a pen and then typing out what he wrote. In work, I find that people who type smoothly and flawlessly often need to think fast. I don't break my stroke of typing a long command line because I can't type fast but because I have to pause to think what I need to type next.


UH, you guys dont recall that this keyboard came out in the 80s?

And it was designed so that you can have them on your thighs and type whilst standing with arms at relaxed hanging?

I loved the idea of chorded-keyboards and I posted in the past about an engineer famous at intel for coding on them on his recumbant bike he would ride to the santa clara campus down san thomas expressway - connecting via a satellite phone and coding in binary in his head as he rode his boke and typed on these guys...


typing speed is rarely the limiting factor on productivity. HN can be though. get back to work.


Agreed, I don't think editing text has ever been a bottleneck for developers at any point of time. The vast majority of time is spent on reading, thinking, discussing, explaining, and debugging code. This is also why I rue the time I wasted learning vi.


This looks awesome. Whether 500 WPM or not, I would give this a try.

Long ago I learned the BAT one hand chorded keyboard. For one hand typing, I got to around 50 or 60 wpm - but only prose. It was much too slow for programming, as symbols and modifiers required their own chord plus the chord of the final character. But with two hands, or even some modifier key foot switches or a second chord hand (like the CharaChorder), it should be fine.


The idea of typing 500+ words a minute sounds silly to me. If you're writing in a natural language and you can touch type reasonably well, your typing skill is not going to be the bottleneck. Your speed at putting things into words is.

Stenography or other data entry are an exception, of course, but I'm guessing the intended audience for these pitches is the average joe who dreams of writing an email in seconds.


What about the mouse?

I use mouse a lot and don't feel like I really giving this habit up because the general ecosystem of apps you have to use if programming is not your primary job and terminal is not your primary environment is not particularly keyboard-friendly.

Does this device offer an efficient (perhaps something trackpoint-like) way to move your mouse pointer without moving your hand between it and the mouse?


Oddly enough no. I would have though that had been thought I to the design since it’s such a natural extension. But no, the CEO has TikTok’s playing Valorant, and he’s using a mouse and only the left side. Which also makes it slightly odd when he makes claims like “characorder helped me win” because when has the speed of keyboard entry ever been the limiting factor in an FPS?


The device currently offers a switch, available to either hand, that moves the mouse up, down, left, right. You can use them both at the same time for diagonal movements and same direction movements to increase the speed--but it's nothing you'd want to use in an FPS and that's why you see Riley playing with the CC in one hand and mouse in the other.


It's not the keyboard itself but the software implemented that allows users to type this fast by predicting the order that you intended to hit keys allowing you to press all the buttons in "hello" at the same time, the fact it translates the ambiguous keypress into a dictionary word is the assistance.

It's clearly "cheating" in the context of a typing competition.


This gets me curious about a Japanese version.

Due to the larger number of basic characters (46) Japanese smart phone input already uses flick-joystick like motions, with pretty amazing speed (for a smart phone). The 46 basic characters are input using 10 keys with 5 directions each: in (press), left, up, right, down.

Does CharaChorder support a press as well, or just left/up/right/down?


Japanese input’s core part is less the direct input and more the conversion to the target character (which relies on the IME, so we’re not far from the “chorded” mechanism of this keyboard)

For instance there is a “kana” mode that let’s you direct input each kana with one keystroke only, which is way faster than romaji input, but is ignored by 99% of the population because the ROI on learning it is extremely low as you’re still converting to kanji at the end. To your point, a ten-key style input on desktop would probably be slower than kana input.


Is Kana input really that much faster than Romaji? I've almost never seen it used (presumably because you have to learn romaji input to type in English anyway, so why learn a second input method), so can't speak to its speed.

However, is it possible that ten-key style input could actually be faster than Kana input? After all, direct Kana input in Japanese is kind of equivalent to Roman input in English in the sense of one key = one character, but this keyboard claims to be faster than direct input due to not needing to move your fingers around. Couldn't the same thing be true for Japanese? In fact, for Kana input the problem of finger travel is even worse due to more keys = longer travel.


I've never seen it used IRL as well.

People tout a 30% speed increase over romaji: https://persol-tech-s.co.jp/i-engineer/technology/key-kana

> this keyboard claims to be faster than direct input due to not needing to move your fingers around

That would perhaps work if you remapped the movement to not be linked to romaji, but direct kana mappings. Then it's not just 26 keys combination for the alphabet alone, but 46+ more arrangement, and more for ° and ". With 10 fingers it feels like hell just for standard text input (to note, you can only use 9 for combination as you need a transformation key, and I guess you also need backspace and arrow keys ?), and you add to that punctuation etc.

I'm not sure there can be an actual human being effectively typing faster with this...


thanks for the link!

> That would perhaps work if you remapped the movement to not be linked to romaji, but direct kana mappings.

Yes, I picture the same 12 key basic input mapping as on smart phones. The CharaCorder has enough joysticks, but looking again at their website it doesn't look like it would work—seems like each joystick only supports 4 directions?... too bad.


Does anyone actually generate 500WPM of content? I sure don’t.

That’s almost 18 tweets or SMS per minute. And I’ll bet that due to limitations of the human brain that if you tried to come up with 500WPM of content that it would have about the same quality and information density as 18 average tweets.

I guess if you were a stenographer or audio transcriber then maybe this is interesting.


For programming, I wonder if there isn't a lot of low hanging fruit just by key remapping. For instance, a large proportion of keystrokes must be for opening and closing brackets, which requires two fingers for shift + the key.

I bet I press brackets more than "d", "f", "j", "k" - the best positioned keys in QUERTY


I have a keyboard[0] that supports a concept of layers which means if I press a key (or hold it down) the layout changes.

With this, I have it set such that the keys i o, and k l, are for curly braces and parentheses. My left hand holds down the trigger to swap layers and so that my right hand doesn't have to move much if at all to hit these keys.

[0]https://ergodox-ez.com


Yes! I dove into this (and later, Moondlander) because I have mobility issues and had to reduce travel, but the resulting 90WPM and (unprovably) faster coding have been awesome side-effects.


500 WPM! I wonder bandwidth measurements we'll soon have to consider with regards to transfer rates in human computer interfaces. I've seen blind people use computers with synthesizers much faster than human are able to talk, maybe even read.

We are still far from programming languages which can make reasonable use of such bandwidth though.


All of this seems like an advertisement. It does seem pretty cool though if you have a need to be able type a lot of text really fast; I don't think most programmers need to type fast though. My daily work as a programmer is like 5% typing code and 95% talking to people and thinking about things.


Hmm, finally a chording device that doesn't look like total BS. I'm unwillingly intrigued.

As for the underlying story, both sides are right. The chording device seems significantly better for text entry in terms of pure speed. But chording isn't typing, and it's typing, not text entry.


In the far distant future, supertypers will use prescience to achieve 10000WPM, like guild navigators


https://news.ycombinator.com/item?id=29816098#29847333 Please refer back here and notice the bias that the thread has taken against me as a sexual assault victim.

The thread has been hijacked with gaslighting and victim blaming, name calling, and unsubstantiated claims including unknowable thoughts on me. Please contact hn@ycombinator.com to support removing this thread if they admins can't apply policy in an unbiased way.


And what is about another (non-English) languages? Or not natural languages at all? Looks like chorded mode is useless for anything but natural English text. And I'm doubt, that it can be easily adapted to non-English layouts even in "char per char" mode.


I saw the Microwriter [1] demoed at the Nottingham Micro Computer Club forty years ago. Nothing new under the sun!

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


When i switched to dvorak, i quickly realized that the main benefit i got was improved comfort, not a ton more speed. My point is only that one should focus on comfort with any keyboard, especially when you're using it multiple hours per day.


This reminds me of the speedrunning community ‘any percent’ ladders and how anything goes to get to the finish. Such a grind, but very specific applicable knowledge, and no meaningful generalizable growth at the end to bring to another game.


10 words in 2 seconds isn't really a very impressive way to do 500 words per minute.


I ordered one. This could be a huge benefit for someone like me who writes for a living.


How fast can he write Perl code?


Is it possible to edit the dictionary used for the auto completion? Is it possible to edit the characters it inputs from the defaults, even to characters not in the English alphabet? Which operating systems is it compatible with?


Yes, you can manage and edit your chord library. Default device has 500 chords for the 300 most common English words.

It supports other character sets--have to join their discord and look to see which ones.

It's compatible with any system a keyboard would be. The OS doesn't know it's not just a USB keyboard.


I could see this for prose, but this seems all but useless for programming. Even if you combined it with copilot or something I think the time you'd spend fixing what it presented would make you slower in the long run.


Make it mobile so that I can use it as a mouse without taking my hands off the keyboard, and I'll go through the trouble of learning to use it.

Could even make both halves mobile so that I can use each one as an independent mouse.


Is this wireless? I've always thought it would be cool to be able to walk and type at the same time. I always seem to get great ideas while walking and it's tedious having to stop and write them down.


Is there any reason you wouldn't use a standalone pocket sound recorder, or alternatively the sound recorder app on your phone?


I always just dictate a note in Keep on my phone. I don't think there's a more effective mobile note-taking system.


Is it a chording keyboard? because those hav existed for decades, and are demonstrably faster for typing. They just have a steeper learning curve.

Follow on question: does anyone know how court stenographer keyboards work?


> Monkeytpe won’t let Keen run up the leaderboard, but he did win another typing competition using the CharaChorder to hit 267 WPM with 76% accuracy.

Oof. That’s bad. No other mention of accuracy anywhere else.


Anyone here remembers datahand? Both of them seem to have similar input space of 4-5 directions per finger and could use similar input methods for mapping that to actual inputs.


Here is the direct link to the CharaCorder website https://www.charachorder.com/


This is really cool.

I'm not sure I want to _use_ it, my experiments with chorded phone keyboards have uniformly been blah. But, I really like it and want to see people using it.


This has got to be a paid advertisement, not a real article


This is just a PR piece by a guy breaking the rules but saying nah it's OK because my thing I made is cool and you should allow it because I made it.


So it lets you type faster than you can think but that’s only useful if you are copying existing text and for that you can just use OCR.


I like this typo in the article: "*Monkeytpe* won’t let Keen run up the leaderboard". The author is a fsat tpyist.


It looks more comfortable to type on than a traditional keyboard. That to me is more important than the speed improvements.


I think the idea is good but if its auto-completing and readjusting as you go I dont think your even typing properly.

Just my 2 cents.


Is the bottleneck the keyboard or my mind? Not sure if having faster keyboard is all that helpful day to day.


Will take me 2 years to learn this that’s more than the time saved in my life time to type a bit slower


I still have a "Twiddler" chord keyboard in my drawer - the Granddaddy of this thing.


Why do a lot of people use "suspect" instead of "suspicious" :|


This is how sus became the one true way.


"Man on bicycle moves so fast he's banned from running competitions"


Looks like it’s banned because it is aided by predictive text software?


This feels like native advertising.

But effective! I may buy one just in case it works.


Is this one of those submarine advertising things?


The picture up top is upside down..


of course it's not cheating to use the best available tool for the task


Cool but who needs this? I'd rather see a keyboard that limits you to 10 WPM so you are forced to think before you type


I've never completed the leap to a chordal interface, but my interest in them is about ergonomics, not speed. For instance, a habit I had to break to avoid hurting myself on the keyboard was rotating my hands to reach a key. This is simply not a problem on chordal keyboards.


The ergonomics angle is a good one, I failed to consider that


Sometimes it's useful to be able to run a series of quick experiments, each based in some complex way on the outcome of the previous ones, to develop an empirical understanding of how something is working.


Not everyone thinks while typing. I do all my thinking ahead of time, then I just need to record it on the screen after I'm done.

I'm usually thinking of other things at that time, since I just need to transcribe the words in my head, not think about them.


This seems more my style too.

I notice that I can listen to rap and write code this way, but if I shift to trying to write a comment or a docstring then I just get stuck. Its like all English circuits are busy, please hold. But the effect goes away when I know the lyrics very well.


How much text do you memorize before you type a round?


It's not a specific amount, and it's not memorization of specific words, it's more like assigning a sub-processor the task of converting the ideas into typing, while the main brain moves on to other things.

You've never had that with programming or other tasks? You think and figure out what to do, and then it's just a matter of getting the idea down on paper (or computer).


Transcribers would probably appreciate it. But then again, there are other regular stenographic keyboards out there.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: