Here's a really interesting interview with the founder of Plover, Mirabai Knight, explaining how she got started and how efficient steno can be. "Steno is so ridiculously more efficient than typing every word out letter by letter that it’s possible to exceed the average qwerty speed in a matter of months, once you’ve got the phonetic system in your muscle memory."
https://geekfeminism.org/2010/10/12/plover-freeing-stenograp...
I got interested in steno once, and even designed and started making my own custom open source steno keyboard before Open Steno existed (though that's a story for another day). This led me to look in to learning steno, and I discovered that the main reason that stenographers type as fast as they do is that they memorize tons of macros (in vim and emacs terms). The triggering of a macro will result in a whole phrase, or even sentence being written out (or just a shorthand for that word/phrase is written out, and that can later be converted by hand in to that word/phrase).
On an ascii keyboard, that would be like typing something like \d and having the phrase "The defendant states that" appear.
It's obvious that simply by using that technique alone (which can be used on a regular keyboard too), one can type many orders of magnitude faster than someone who's typing out each individual word letter by letter (or even phonetically). Some steno systems will give the stenographer hundreds if not thousands of such macros to memorize.
Of course, there's also the phonetic spelling stenographers use if they don't already know a macro for a given word (or for a proper noun), and there's the funky chorded keyboard that makes it efficient to type thousands of different macros with a relatively few keys and without moving the fingers much. But that's secondary to the heavy reliance on macros.
I hate memorization, and didn't want to spend the years that it would take to master steno, so never really wound up pursuing this myself. So I don't actually know steno myself, so please take the above explanation with a grain or two of salt. But that's my understanding of how the greatest speed advantages are achieved with steno.
I'd love to hear from some professional stenographers as to whether my impression was correct.
> I hate memorization, and didn't want to spend the years that it would take to master steno, so never really wound up pursuing this myself.
I haven't learned machine stenography (for reasons which fit better in another comment), but I do know and frequently use good old pen-and-paper shorthand, which is shares a lot of properties with stenography. I also talk to people who know machine stenography to keep up with the current developments.
There's one misconception in the quote that I feel it's worth clearing up: learning stenography/shorthand does not rely as heavily on "memorization" as you seem to think.
There is a system to the madness, and most systems are fairly easy to learn (because – hey – they're systematic!) Once you know the system, you don't need to memorize anything. You will be able to type anything by just applying the rules of the system. Of course, consciously applying the rules is going to be a slow process (just like consciously hunting for keys on a Qwerty keyboard when you were still learning it), but words you type often will settle into your muscle memory, just like they have done on Qwerty. Try typing in a foreign language on Qwerty and you'll notice how much of your typing speed is muscle-memory-applied-to-full-words. The same thing is true of stenography.
You never sat down and deliberately decided to "memorize" common words on Qwerty, it just happened automatically as you were typing. The same thing is true of stenography.
So, to recap:
* "Mastering stenography" in the sense of "being able to type at a speed that is not uncomfortable" may take no more than a month. It does require memorizing the system and applying it.
* "Mastering stenography" in the sense of exceeding your current Qwerty speed may take a few months up to half a year. It does not require any memorization beyond the first step.
* "Mastering stenography" in the sense of achieving 250 WPM speeds will take several years. It will probably require memorizing custom shortcuts for words/phrases that you commonly type.
So if you compare stenography fairly to Qwerty, you see we're talking at most half a year. It probably took you longer to learn Qwerty to begin with. You're being unfair to yourself if you compare 250 WPM stenography to 70 WPM Qwerty and from that conclude that stenography is much harder.
> It's obvious that simply by using that technique alone (which can be used on a regular keyboard too), one can type many orders of magnitude faster than someone who's typing out each individual word letter by letter (or even phonetically).
It probably helps if the nature of the text you have to type is somewhat formulaic.
I'm actually in the process of learning the Plover theory. Largely for fun. After a month of drills, learning to find my way around the keyboard, I've started transcribing an old Harry Harrison story The Man from P.I.G.
It has a herd of space pigs and a substantial vocabulary so it's hardly formulaic. That doesn't seem to have been a problem so far.
Makes me wonder if we could design a programming language that is steno-friendly. I'm sure many of the requirements would translate to better human readability too.
Why do you want to type code quickly? I gave up touch typing in high school more than forty years ago but have been coding for most of the intervening time. I don't think I ever felt a need to type code faster because even with just two fingers I still type code faster than I can think it.
Because I have to pause my train of thought to type. And the longer I have to pause, the more likely I am to lose some vital part of the mental model I am juggling. Therefore, being able to type quickly and get back to thinking sooner is a good thing.
I mean, going only by lines of code written per day, one would think I would be content with being able to type only at maybe 8 words per minute, and that's on a good day! But it is obvious to anyone why being able to type only 8 words per minute can be detrimental to a programmer.
Most people who type "incorrectly" but have lots of practice can do 50-100wpm. That's my observation/experience, anyway. I don't touch type but I'm plenty quick enough. Is there a useful difference between that and 200wpm steno, for programming?
If anyone has data on that, I'd be an interesting read.
I'm talking plainly in terms of a null hypothesis, though. If we agree that there is a notable difference between 8 and 50 WPM, we should assume the effect also holds between 50 and 250 WPM, unless otherwise proven.
There is no reason to believe the "cutoff" value where the effect stops being noticable is at, say, 100 WPM, just because that happens to be a common Qwerty typing speed.
I use QWERTY and I can definitely maintain 140WPM easily when typing something from memory or when participating in realtime chat. In short bursts, ~230WPM isn't difficult. A look at sites like TypeRacer shows that there are indeed many typists who can achieve 150-160WPM, and some occasionally exceed 200.
That said, I think typing speed is highly dependent upon the keyboard, and layout is only one of many other factors. From personal experience I have a feeling most keyboards out there are extremely suboptimal, with too much key travel and heavy actuation force. I've not be able to go above 130WPM on any laptop keyboard, for example, and even trying to achieve 100 feels extremely tiring. On the other hand, my desktop keyboard happens to have extremely light keys and shallow travel with a good bouncy "cushion" at the end, and I can type for many hours at much higher speeds without feeling tired.
Steno's biggest advantage comes from having very long words produced with a single chord, but the average word length is not so high, and most faster typists I know, as do I, already somewhat "chord" keys for common sequences (otherwise transposition errors wouldn't appear at all --- "hte" instead of "the" being one example, as those with experience will tend to hit both keys almost simultaneously for that.)
It's a TurboStar KM-2601P, a pretty generic-looking keyboard with some multimedia keys on the top and a "turbo" button that adjusts the auto-repeat rate from "none" to "ludicrous". For some reason none of the big search engines (Google Bing Yahoo... and DDG) give any results(!?) for that model, but that's exactly what is printed on the sticker on the back and I do remember searching for it a few years ago and getting quite a few results... and I've even mentioned it before, along with a few others, in various forum posts etc.
On the other hand, there are mentions of the KM-2501P and KM-2601 (without the P suffix), which look the same except for the logo but I'm not sure if the slightly different model number means they have the same switch characteristics:
The total travel isn't very short --- laptop-style switches would be shorter --- but it doesn't take much force to get past the tactile "hump" where it actuates, and after that is a bouncy feeling that helps in pushing the finger back up. Perhaps what I find most desirable is that the drop in force over the hump is not very steep, unlike with many other keyboards I've used where it feels like the key just stays down with very little force and refuses to come back up until the finger is completely removed from it.
Incidentally, I have a mechanical keyboard with blue ALPS switches, but it feels more tiring to type on it than this KM-2601P.
I'm late to the party, but the website posted here is on GitHub pages. If you'd like to PR a change to remove the speed difference between QWERTY and Dvorak, I'd love the help. Could just be "top keyboard typist", and the speed is referring to the top percentile, not Sean Wrona's world record speed.
The metric that a top dvorak typer is 20 WPM faster than a top qwerty typer is not true. Dvorak is more ergonomic, but doesn't make you type substantially faster.
If I was going to go to all the trouble of learning a non-qwerty keyboard layout, I wouldn't choose dvorak, but would probably choose something else that's supposed to be even more efficient and ergonomic.
But for me it's really not worth switching from qwerty, as I don't want to have problems typing on someone else's qwerty keyboard when I need to -- which is a problem I've heard many non-qwerty typists report having, even if they were once fluent in qwerty. Some do report that they can make the cognitive switch from one to the other relatively easily, but some can't.
So for me it's not worth the risk of spending all that time to learn another layout just to find it hampers my qwerty-typing ability when I might need it.
If I was going to learn anything, it would be steno -- as at least it being on a radically different keyboard should mean that learning it shouldn't impact my qwerty-typing ability.
I was just trying colemak, from what I read the loss in qwerty is dued to people making the fast jump, I will see if learning the different layout in like the evening and keeping qwerty for the rest of the day preserves both memories
I'm a dual qwerty/colemak touch typist. I was able to learn colemak quickly with low risk of losing qwerty because I was already a very good touch typist. A large amount of typing skill is transferred which is why people with poor typing skills have such a difficult time learning a new layout.
As far as say Dvorak or colemak gaining you faster speeds, I'm pretty doubtful of that claim. It has more to do with how much effort you decide to put in.
Note that Dvorak was optimized for standard mid-20th-century typewriters. (Just as QWERTY was optimized for the original late-19th century Sholes/Remington typewriter design.)
In that context, it’s a very significant improvement over QWERTY, in particular by avoiding the bottom row, which on a typewriter is slow and awkward to reach. 15-20 WPM advantage sounds pretty plausible to me.
On a “modern” mostly flat keyboard, some other layout would be better than either QWERTY or Dvorak (personally I don’t think any of the popular layouts are properly optimized for common current keyboards, but to be fair the modern keyboard is a pretty terrible design). Dvorak is still a significant improvement over QWERTY though.
But you will have less training than on qwerty and the travel might be neglectable. Or maybe even something else, I have trouble but pressing the wrong vocals in the home row somehow (not typing Dvorak, but something similar, the german neo layout). I will still never go back to qwerty.
Although I have been looking into the possibility of a compact chorded keyboard: it would make typing on the go feasable.
...Hang on a minute: I play recorder. There are 8 keys on a recorder, and there are 8 bits in an ASCII sequence (which is what I'm typing). Suddenly, I think I see a viable design: get me the Arduino.
Very cool, but unlikely that I would be a customer of this. Most of my typing is for programming, not writing long prose, and for programming, my typing speed is typically not the bottle neck.
If anyone Dutch, German, Swedish, Danish or Norwegian is following this project, please tell me if you know of a dictionary/system/way to create a dictionary for any of those languages. I would like to use Plover for Swedish, but am hindered by the lack of a dictionary. I know the languages I enumerated has often shared systems for both machine steno and shorthand in the past, so I'm open to any such suggestions.
The website says it would be good for coders, but I'm skeptical; there are far too many particular characters that wouldnt translate as easily to steno as natural language, not to mention the need for keyboard shortcuts to compile run etc.
I'm sure there are people out there who can make it work for programming, but I would strongly advocate against trying. I'm a programmer now, but before I got into programming I was on track to become a Stenographer. I wrote at length about this two years ago when there were a few articles on HN discussing the possibility of using Stenography for coding.
Negativity aside, I'm pretty thrilled to see Open Source projects in the field. It costs 4,000 dollars for Case Catalyst, which you can think of as Microsoft Word for Stenographers. It's worth it if you're a professional, but I suspect that the price is only that high because there's no competition.
edit: In fact, I think the price might have gone up in the last two years. I could have sworn that it was 3,000 dollars back in 2014.
dont get me.wrong i think it is a wonderful project, and i would def try it out, but i dont forsee much benefit for any programmers, unless youre writing pure sql or cobol, i guess
I'm a long term programmer and a very short term stenographer (over 20 years vs. 3 months). I'm cautiously optimistic that it could be a help with entering code given a correctly set up IDE. That said - I don't think the advantage is high speed entry of text. The speed of thought is the obvious bottleneck rather than typing speeds.
Where there might be an advantage is that it might be possible to set up chords to do relatively complex structural manipulations of text and give that chord a useful phonetic name. The sort of thing that expert emacs users do almost as a matter of course. Learning emacs well would be a far more conventional approach of course but steno does offer the possibility of easier to remember commands (once you've understood and learned the phonetic system).
