They do, but this simplifies the process significantly.
Think of it as a purpose built variant of typing apps. You can also do many things on Excel or Word, but sometimes it's more convenient to perform some tasks with a purpose-built app that is built for your problem domain.
I'll give this a try for sure when I'm on my computer.
I haven’t worried much about my typing speed in years (100WPM is faster than I can think when coding or writing most of the time), but I am learning German and wonder if typing massive amounts of text in a language you’re trying to learn could have some benefits?
> I haven’t worried much about my typing speed in years (100WPM is faster than I can think when coding or writing most of the time),
I hear this argument all the time and I don't buy it. Yes, my thinking is slower than 100 wpm on average, but it is extremely bursty. It goes at 0 wpm most of the time then 10000 wpm for short intervals.
I believe the fundamental concern here is not average speed, but latency. How fast can you get the current thought onto the page so that your brain can move on to think about the next thing?
I also type at 100 wpm, and I find my wetware CPU experiences stalled cycles while I'm typing, not able to continue because it needs to hold on to a buffer containing the thought I'm slowly typing out.
We type at something like 150 wpm casually and our brain is still a bit faster... although we've been optimizing it to type out chat messages for years, so that's probably why we can come up with them with such ease
What kind of keyboard do you have that polls at such a rate that it can keep up with "almost infinite" (whatever that means) WPM? The best keyboards I have found only have 1000Hz scan rate, which in the absolute best case means only 30,000 WPM. Granted, compared to infinity, every number is equidistant, but when you say "almost infinite" I expect something a bit more impressive (in the 10**80 range, for example).
Right. Time spent typing adds latency to evaluating some solution. Ideally, everyone could type with 100% accuracy at 300WPM or whatever.
But, I expect it's diminishing returns on effort.
If you type 30WPM, you'll probably benefit yourself (and your colleagues!!) from learning to type quicker. 80 WPM seems reasonable; 100 WPM is good; 180 WPM is excellent. -- But, I think above some point, the costs of training to type quicker aren't justified by the benefits.
Could it have some benefits? Sure. Maybe you enjoy the movement of your fingers. Maybe it slows your brain down to the point where you linger on each individual word enough to see new things click.
Are these benefits enough to make it the best thing in terms of ROI? I would suspect only rarely.
- If you're not yet at the point where you can understand what you're typing, you are probably served better by more traditional language learning activities.
- If you do understand it perfectly you're probably at an advanced enough level that just reading it at your normal reading speed, or writing your own stuff and then having a native speaker is a better form of practice.
- If you can understand it with effort but not at a comfortable reading speed yet - you're probably best served by throwing the individual sentences along with machine translations into a spaced repetition system like Anki. Then you can see the same sentence multiple times over the course of a few days and deepen your understanding a little more each time.
Probably in getting used to the forms of the language, but for best results you will want to stay near things you can almost understand. I had two relevant experiences: (1) graduate school advisor told the lab to copy down one paper per week by hand. After a while you learned to copy mindlessly and stopped learning from it. (2) listened to the same second language content 100+ times while commuting and it really helped to build an intuition for what was being said and how to say it.
as well as heart of darkness. he did it more than once. the quote/paraphrase attributed to Hunter on this that I love is, "I wanted to know what it felt like to write something good."
edit: I was oddly talking about this today to my partner. I've honestly never heard of anyone doing it outside of Hunter.. but that might just be because Hunter mentioned it in such a cool way.
I sometimes read to fall asleep. Usually I am already tired so I can get though 1-3 pages per session.
I started the count of monte cristo not knowing it was about 2000 pages, so that’s what I read for the whole of 2019 basically. At least when going to bed.
Typing it completely seems like a huge undertaking.
Just replied in the comment above. This is more as an intellectual curiosity. I intend to publish it eventually as a bi-lingual ebook with para by para translations. Hopefully, it is useful for those who are learning one of the languages and is not a native speaker/writer of the other.
I am not using any automated translations except to find esoteric translations (using Google Translate). Many idioms and phrases also have to be localized, so it is an intellectual challenge.
Alexander Dumas' writing is also quite beautiful and even in the limited number of chapters I have completed translating there are some excellent narratives. I am enjoying it.
My idea is to eventually publish it as a bi-lingual ebook, perhaps it would help people who are learning English/Tamil.
I use a Surface tablet as my writing device and have a Tamil Input Method Editor (IME) for Windows installed. It does the English-to-Tamil transliteration and that helps.
I disagree. Reading requires my full attention and dedication, while typing becomes almost an almost mechanical process very quickly.
With reading, I always find myself moving my eyes back a few sentences because I would think I missed some unimportant information, taking me a while before I actually got into it. Maybe it's just because technology hurt my attention span.
Neat. I coded a similar idea for fun, back when Java Applets still roamed the earth. The backend would harvest Project Gutenberg files and the frontend would present those to the user.
This is how I would practice writing shorthand. It's good too because you can practice reading what you wrote a few weeks later (this is as important as writing tbh), and if you really fucked up you can go back and see what the word was supposed to be.
I can type at a decent pace as long as it's text ; I learnt to type on gtypist . (I haven't completed all of them just the basics and some drills) .
Typing for programming is a whole lot different , and I find myself fumbling around numbers and special characters often
Sure, code written (and rewritten) (and deleted) versus net output. But maybe be more thoughtful, do less than 40wpm, and write a better solution overall.
You may be interested in typing.io, which is specifically for code. I don't think it'll help much with numbers though. Personally, I've never gotten the hang of touch typing the number row, so I use the numpad instead.
I switched from typing.io to https://github.com/jankrepl/mltype. Highly customisable.
I run it from a bash script in multiple steps with custom length. First special characters only, then numbers, letters, go and finally rust.
I do this on a daily basis just for a few minutes and I could greatly improve my typing speed.
When I practice typing and make an error, I thought that perhaps it should automatically go back to the beginning of the word. Because the road leading to fast typing is to type letters in groups, and having the muscle memories of common words might help with both speed and accuracy. The statistics could include frequently mistyped pairs and triplets, or basically, ngrams. And you can practice those in isolation.
It will be like scale practicing of musical instrument.
This is a very desirable action. Once you get some speed up
touch-typing you ( I still do ) fluff whole words. Infuriated by an
emacs not having an ergonomic offering I bound M-w to
backward-kill-word and found it so useful it's still in my .emacs
file.
This might be something you would actually overcome as you build up even more speed by focusing on learning to correct the individual letters.
I've been typing 120+ wpm for 20 years and it's actually surprising to me that this is a problem people have. It's that foreign to my experience.
I could be totally wrong. It could just be that typing happens in different ways in different people's brains and bodies and that's why I can't relate to this, but I'd challenge you to try correcting the letters rather than the words and see if that doesn't have a compounding effect on improving your speed.
Except that instruments don't come with a backspace. If you want a closer approximation simply refuse to edit your texts at all: write it out in one go and refuse to fix any mistake that you make. What you will find is that your speed initially drops to maybe 10% or even less of what it was before, but after a while it speeds up again and at some point you'll be typing faster than ever before because fixing mistakes takes a lot of time.
Your nick, by the way, is perfectly suited to this endeavor.
>Except that instruments don't come with a backspace. If you want a closer approximation simply refuse to edit your texts at all: write it out in one go and refuse to fix any mistake that you make.
The other day someone posted a webapp they wrote that doesn't let you edit and that blurs all the text except the current line while you're typing. If you want to edit your text, you need to download the file. His aim was to separate writing from editing, which I think it's an interesting approach.
> What you will find is that your speed initially drops to maybe 10% or even less of what it was before, but after a while it speeds up again and at some point you'll be typing faster than ever before because fixing mistakes takes a lot of time.
I've never been entirely convinced by this argument. I will always make mistakes, so I could argue that I need to be able to quickly fix those mistakes, so I should practice doing so to some degree. The implication that fixing your mistakes somehow trains you to make mistakes seems tenuous.
I will admit that for me, my "end goal" allows me to see what I'm typing and requires me to correct my mistakes. If, for instance, your end goal has you not even seeing the text you're typing then there's probably no reason to go back and fix your mistakes.
This is where the musical instrument comparison breaks for me. If your goal with an instrument is a performance, then with a musical instrument you can't fix your mistakes during the performance. You will still make mistakes, of course, you just won't be able to fix them.
That said, if you're making a lot of mistakes then perhaps you're getting enough practice fixing them and should stop. It's just that I don't necessarily accept a hard rule that you should carry on regardless and the outcome will be better. I've never seen any research on this subject.
No, the implication is that if you slow down to the point that you become once again more conscious of what you are doing that you will re-train your muscle memory to the point that mistakes will become less likely. If you 'practice your mistakes' you will make mistakes!
This is a problem many people have when practicing to play an instrument, they end up practicing but making so many mistakes that their muscle memory sees exposure to too many mistakes to get the right pattern down. By avoiding that through slowing down and practicing the right moves instead of a mixture of the right moves and the mistakes you will lay down patterns that are much more dependable.
When performing that pays off because then there is no way to fix your mistakes. So the best way to practice is to not make mistakes. And any practice method that helps you to do that is a good one. This is pretty much the philosophy being pianojacq.com, a sightreading / practicing application that I've been building over the last couple of years and the effect on my own ability to learn and practice without mistakes has been huge. That does not imply that this is for everybody but from my own experience to slow down, even if that means being agonizingly slow for a number of passes without mistakes has a pay-off that is much higher than practicing the same piece with mistakes at a higher speed. Once you do the right thing, speed is easy. As long as you don't do the right thing speed is irrelevant.
Ah, that does make sense but it wasn't what I was originally replying to. You didn't say why we were losing 10% speed after refusing to fix our mistakes, and I actually thought that this was supposed to be due to the software penalizing us for the mistakes when we could have got a higher speed overall if we'd corrected them ourselves.
Slowing down during your deliberate practice to the point where you don't make many mistakes could well be optimal, and something I should probably do more of. Albeit it still isn't clear to me what we should do with the occasional mistake we make even while going slow, but I accept that you think we should ignore them.
The basic idea is to lay down repetitions of the 'perfect' pattern to the point that you are no longer thinking about it and any mistake will mar that process, so it is better to slow down to the point where you are no longer making mistakes (or at least, very few, initially, and hopefully after practice so few that they no longer happen at all).
For me, depending on the complexity of a given piece that can be anywhere from 5% to 20% or so of indicated tempo for the first times that I practice a piece. One of the hardest things to take into account when practicing slow is that you should still use the exact same fingerings as you would when playing fast because otherwise all of your practice is for nothing!
Yes, but in performance it doesn't. And reason why you want to go back to the beginning of the phrase is to make sure that the pattern remains whole, if you just do the little bit where your mistake was chances are you'll end up with alternate fingerings or other artifacts that will make the fix at best useless and at worst a net negative.
On this site, I opened a book at random and was surprised to see several errors in the text. For instance, a full stop was clearly missing in the first line. So I checked the source, which seems to be gutenberg.org (the site has no reference at a book level, but mentions some sources on the front page). Yet the same book (Botchan) on Project Gutenberg does not have these errors.
So I'd suggest two enhancements to the author. Firstly, for each book, state the exact source, including the retrieval date. Then make sure that the up to date version of each book is used.
I also wonder if there could be a way to make all this typing more useful. There are so many books lacking OCR, or lacking review after a poor OCR.
"Those who have insinuated that Menard devoted his life to writing a
contemporary Quixote besmirch his illustrious memory. Pierre Menard did
not want to compose another Quixote, which surely is easy enough-he
wanted to compose the Quixote. Nor, surely, need one be obliged to note
that his goal was never a mechanical transcription of the original; he had no
intention of copying it. His admirable ambition was to produce a number of
pages which coincided-word for word and line for line-with those of
Miguel de Cervantes.
...
Initially, Menard's method was to be relatively simple: Learn Spanish,
return to Catholicism, fight against the Moor or Turk, forget the history of
Europe from 1602 to 1918-be Miguel de Cervantes. Pierre Menard weighed
that course (I know he pretty thoroughly mastered seventeenth-century
Castilian) but he discarded it as too easy. Too impossible, rather!, the reader
will say. Quite so, but the undertaking was impossible from the outset, and
of all the impossible ways of bringing it about, this was the least interesting.
To be a popular novelist of the seventeenth century in the twentieth seemed
to Menard to be a diminution. Being, somehow, Cervantes, and arriving
thereby at the Quixote-that looked to Menard less challenging (and there-
fore less interesting) than continuing to be Pierre Menard and coming to
the Quixote through the experiences of Pierre Menard." ~ Pierre Menard, Author of the Quixote (Borges)
I also immediately thought of Borges on reading the link's title.
When I was learning Spanish in adulthood, I decided for practice to re-translate this Borges story back into Spanish, having never read it in the original. Incredibly challenging (for such a beginner), but also very amusing to see the overlaps and the differences.
Would it be useful to use typing practice as a way to crowdsource scanned print publications better than OCR? (Do the OCR, and see whether a few human samples "vote" better text? Then maybe also use that to improve the OCR?)
I've been seeing imperfect OCR. And you can "vote" (like a voting configuration of redundant devices, on the assumption the majority is correct) among multiple typists.
Be aware that RSI is a real risk and once you have it, it never goes away. That is serious, even critical, if your job is at the keyboard. So be moderate.
It is, but given the repetitive nature of almost any job avoiding RSI in the longer term is pretty much impossible nowadays, unless you intend to avoid having jobs for longer than a few months.
Never is a strong word- mine went away. I had debilitating carpal tunnel RSI, and with the right ergonomics (*) it went away completely in a few months, and no sign of it for the couple of decades since, still using a keyboard for most of my waking hours.
I’d say time spent typing is much less of a factor than how that typing is done.
(*) The first big fix for me was switching from a desk to laptop on a couch where I could change positions every few minutes. The short travel and light weight of laptop keys also helped. I switched back to a desk more recently, but I kept the low-profile light keys, and the general awareness of postural stress.
There are people out there who don't know how to type properly, though I'm not sure if they'd be reading much HN. Most Americans under 40 would have learned touch typing in grade school.
Yeah, I learnt touch typing as a teenager with no instruction and as far as I'm aware most of my friends also did. My mum used to type with one finger at that time but I think by now even she would struggle to find a use for this tool
I use the Dvorak layout and have no issues on the site. I am accessing it through Firefox under OpenSUSE, though, so it may be an issue with your browser?
There is an awesome app called type-fu (type-fu dot com) which uses words, proverbs, quotes, facts, and code as typing material just as some users in the comments requested. It has many layouts, including dvorak, workman, norman, etc., and online version seems to be free.
Someone should hook this up to train an ML like which “types out a sequence of words for a human”
It’s a really small, dumb improvement but it’s something ML would be useful for and you probably won’t need much data to produce something reasonable. Current “simulated typing” looks clearly simulated, as every letter has an even delay and humans don’t type evenly (e.g. I’m sure repeat letters are typed faster)
https://github.com/jankrepl/mltype is something like this. It is tuned for programmers though. And the quality of generated text might vary a lot. But it gives reasonable char sequences.
132 comments
[ 4.6 ms ] story [ 208 ms ] threadThink of it as a purpose built variant of typing apps. You can also do many things on Excel or Word, but sometimes it's more convenient to perform some tasks with a purpose-built app that is built for your problem domain.
I'll give this a try for sure when I'm on my computer.
I hear this argument all the time and I don't buy it. Yes, my thinking is slower than 100 wpm on average, but it is extremely bursty. It goes at 0 wpm most of the time then 10000 wpm for short intervals.
I believe the fundamental concern here is not average speed, but latency. How fast can you get the current thought onto the page so that your brain can move on to think about the next thing?
I also type at 100 wpm, and I find my wetware CPU experiences stalled cycles while I'm typing, not able to continue because it needs to hold on to a buffer containing the thought I'm slowly typing out.
But, I expect it's diminishing returns on effort.
If you type 30WPM, you'll probably benefit yourself (and your colleagues!!) from learning to type quicker. 80 WPM seems reasonable; 100 WPM is good; 180 WPM is excellent. -- But, I think above some point, the costs of training to type quicker aren't justified by the benefits.
Are these benefits enough to make it the best thing in terms of ROI? I would suspect only rarely.
- If you're not yet at the point where you can understand what you're typing, you are probably served better by more traditional language learning activities. - If you do understand it perfectly you're probably at an advanced enough level that just reading it at your normal reading speed, or writing your own stuff and then having a native speaker is a better form of practice. - If you can understand it with effort but not at a comfortable reading speed yet - you're probably best served by throwing the individual sentences along with machine translations into a spaced repetition system like Anki. Then you can see the same sentence multiple times over the course of a few days and deepen your understanding a little more each time.
https://www.openculture.com/2017/06/hunter-s-thompson-typed-...
Hunter S. Thompson Typed Out The Great Gatsby & A Farewell to Arms Word for Word: A Method for Learning How to Write Like the Masters
edit: I was oddly talking about this today to my partner. I've honestly never heard of anyone doing it outside of Hunter.. but that might just be because Hunter mentioned it in such a cool way.
I started the count of monte cristo not knowing it was about 2000 pages, so that’s what I read for the whole of 2019 basically. At least when going to bed.
Typing it completely seems like a huge undertaking.
I am actually translating The Count of Monte Cristo. Completed 5 chapters so far. Still a loooooooong way to go.
What are you translating it to? From which language?
I am not using any automated translations except to find esoteric translations (using Google Translate). Many idioms and phrases also have to be localized, so it is an intellectual challenge.
Alexander Dumas' writing is also quite beautiful and even in the limited number of chapters I have completed translating there are some excellent narratives. I am enjoying it.
My idea is to eventually publish it as a bi-lingual ebook, perhaps it would help people who are learning English/Tamil.
With reading, I always find myself moving my eyes back a few sentences because I would think I missed some unimportant information, taking me a while before I actually got into it. Maybe it's just because technology hurt my attention span.
Writing code is limited by the pace of you being able to think up sound effective code.
I can't recall where I saw it but believe I heard a rule of thumb of only 40wpm pretty much mazing out most people's ability to write code.
Reference: The Internet
(For example: https://stackoverflow.com/questions/966800)
---
Sure, code written (and rewritten) (and deleted) versus net output. But maybe be more thoughtful, do less than 40wpm, and write a better solution overall.
I think the point is 40wpm allows you to code. If you wrote at 10 lines per day, 80 char per line works out to 0.0333wpm.
You'll forget what you're doing by the time get a word out.
As in once you hit 40wpm you won't be slowed by your typing as you try to code.
(WPM is characters per minute over 5, i.e. assumes average word is 5 chars long).
PS: I didn't measure my speed back then, but it definitely worked. I go at 90+ wpm these days without a sweat!
It will be like scale practicing of musical instrument.
Same. For some reason we always mix up "you" and "the", even though they are completely different words
I've been typing 120+ wpm for 20 years and it's actually surprising to me that this is a problem people have. It's that foreign to my experience.
I could be totally wrong. It could just be that typing happens in different ways in different people's brains and bodies and that's why I can't relate to this, but I'd challenge you to try correcting the letters rather than the words and see if that doesn't have a compounding effect on improving your speed.
Your nick, by the way, is perfectly suited to this endeavor.
The other day someone posted a webapp they wrote that doesn't let you edit and that blurs all the text except the current line while you're typing. If you want to edit your text, you need to download the file. His aim was to separate writing from editing, which I think it's an interesting approach.
https://write.sonnet.io/
[1]: https://news.ycombinator.com/item?id=34183244
I've never been entirely convinced by this argument. I will always make mistakes, so I could argue that I need to be able to quickly fix those mistakes, so I should practice doing so to some degree. The implication that fixing your mistakes somehow trains you to make mistakes seems tenuous.
I will admit that for me, my "end goal" allows me to see what I'm typing and requires me to correct my mistakes. If, for instance, your end goal has you not even seeing the text you're typing then there's probably no reason to go back and fix your mistakes.
This is where the musical instrument comparison breaks for me. If your goal with an instrument is a performance, then with a musical instrument you can't fix your mistakes during the performance. You will still make mistakes, of course, you just won't be able to fix them.
That said, if you're making a lot of mistakes then perhaps you're getting enough practice fixing them and should stop. It's just that I don't necessarily accept a hard rule that you should carry on regardless and the outcome will be better. I've never seen any research on this subject.
This is a problem many people have when practicing to play an instrument, they end up practicing but making so many mistakes that their muscle memory sees exposure to too many mistakes to get the right pattern down. By avoiding that through slowing down and practicing the right moves instead of a mixture of the right moves and the mistakes you will lay down patterns that are much more dependable.
When performing that pays off because then there is no way to fix your mistakes. So the best way to practice is to not make mistakes. And any practice method that helps you to do that is a good one. This is pretty much the philosophy being pianojacq.com, a sightreading / practicing application that I've been building over the last couple of years and the effect on my own ability to learn and practice without mistakes has been huge. That does not imply that this is for everybody but from my own experience to slow down, even if that means being agonizingly slow for a number of passes without mistakes has a pay-off that is much higher than practicing the same piece with mistakes at a higher speed. Once you do the right thing, speed is easy. As long as you don't do the right thing speed is irrelevant.
Slowing down during your deliberate practice to the point where you don't make many mistakes could well be optimal, and something I should probably do more of. Albeit it still isn't clear to me what we should do with the occasional mistake we make even while going slow, but I accept that you think we should ignore them.
For me, depending on the complexity of a given piece that can be anywhere from 5% to 20% or so of indicated tempo for the first times that I practice a piece. One of the hardest things to take into account when practicing slow is that you should still use the exact same fingerings as you would when playing fast because otherwise all of your practice is for nothing!
In practice it does. The common advice is to go back to the start of the phrase when you make a mistake practicing instrument.
So I'd suggest two enhancements to the author. Firstly, for each book, state the exact source, including the retrieval date. Then make sure that the up to date version of each book is used.
I also wonder if there could be a way to make all this typing more useful. There are so many books lacking OCR, or lacking review after a poor OCR.
...
Initially, Menard's method was to be relatively simple: Learn Spanish, return to Catholicism, fight against the Moor or Turk, forget the history of Europe from 1602 to 1918-be Miguel de Cervantes. Pierre Menard weighed that course (I know he pretty thoroughly mastered seventeenth-century Castilian) but he discarded it as too easy. Too impossible, rather!, the reader will say. Quite so, but the undertaking was impossible from the outset, and of all the impossible ways of bringing it about, this was the least interesting. To be a popular novelist of the seventeenth century in the twentieth seemed to Menard to be a diminution. Being, somehow, Cervantes, and arriving thereby at the Quixote-that looked to Menard less challenging (and there- fore less interesting) than continuing to be Pierre Menard and coming to the Quixote through the experiences of Pierre Menard." ~ Pierre Menard, Author of the Quixote (Borges)
https://www.jenliu.info/DIAP/Borges-Pierre-Menard_text.pdf
When I was learning Spanish in adulthood, I decided for practice to re-translate this Borges story back into Spanish, having never read it in the original. Incredibly challenging (for such a beginner), but also very amusing to see the overlaps and the differences.
https://en.wikipedia.org/wiki/ReCAPTCHA#ReCAPTCHA_v1_(human-...
I’d say time spent typing is much less of a factor than how that typing is done.
(*) The first big fix for me was switching from a desk to laptop on a couch where I could change positions every few minutes. The short travel and light weight of laptop keys also helped. I switched back to a desk more recently, but I kept the low-profile light keys, and the general awareness of postural stress.
That's interesting. Do they really use it though? What I can see from other part of the world that even developers don't use it.
It’s a really small, dumb improvement but it’s something ML would be useful for and you probably won’t need much data to produce something reasonable. Current “simulated typing” looks clearly simulated, as every letter has an even delay and humans don’t type evenly (e.g. I’m sure repeat letters are typed faster)