Hacker News new | past | comments | ask | show | jobs | submit login
How to Write 225 Words Per Minute With a Pen (theatlantic.com)
218 points by spne on June 29, 2014 | hide | past | web | favorite | 85 comments

While Gregg shorthand is great for English, its not much use for anything technical. Many of the shorthand idioms are based on the word-sound, not the word-spelling, so it's okay if you're taking notes for yourself, reading someone else's shorthand becomes an exercise in decoding the word in context because you were not there for their experience during note-taking.

If you don't stay fresh with your own shorthand, just like your own source code, you can find yourself re-reading your own work and wondering what you were thinking!? That's kind of what this pen in the article semi-solves, by having an audio recording synched with your shorthand.

That is precisely why the author relies on his smartpen:

"After all, shorthand—at least my shorthand—isn’t foolproof. Even though, for short stretches, I can write 120 words per minute, the average American now speaks at more than 150 words a minute, so something is inevitably lost in the translation. With my trusty Sky smartpen, though, I don’t have to worry. Most of the time, when I get around to writing the story, I can rely on my handwritten notes for short quotes and background information. But when I need a longer, verbatim quote, I can use my notes as a kind of index to find the right part of the interview, and tap there with my pen to hear the playback. No more of the forward/reverse dance with a digital recorder when it’s time to transcribe. It’s an almost flawless system."

The article isn't just about a cool old technology (though he does spend a lot of time on it), but to demonstrate the novel combination of complementary old and new technologies.

My understanding is that it was not generally expected that notes taken by one person would be interpretable by another. This is what I've been told by my mother, who learned and used Pitman shorthand. She says she never had to read anyone else's notes, nor did anyone else ever have a need to read hers.

Reading other people's shorthand was common in larger companies. One secretary would take dictation, and pass it off to another secretary to type it up.

Typing is slower than taking dictation, so you'd potentially have multiple typists typing up the same person's dictated letters at once.

This is largely why note taking is a two pass process for me. The first is the pen and paper notes taken while speaking with someone else. The second is the process of transcribing and editing those notes onto some digital medium. I end up with a really nicely fleshed out document covering much more information, much more thoroughly. For example, for a more complex point in a discussion I may add an additional paragraph or two of background and context to what was actually talked about.

I've found this process to be the best I've tried, and others who I share my notes with have voiced their liking.

If you're not reviewing your notes at least once after you've taken them, you shouldn't even bother taking them. I don't think there are very many exceptions for write-once-and-never-read notes. Note amplification is stronger if your intent is to explain the notes to someone else.

> If you're not reviewing your notes at least once after you've taken them, you shouldn't even bother taking them.

Beethoven would disagree with you,

    Beethoven left behind an enormous number of sketchbooks.
    Yet he himself said he never looked at a sketchbook
    when he actually wrote his compositions. When asked, "Why
    then, do you keep a sketchbook?" he is reported to
    answered, "If I don't write it down immediately I forget it
    right away. If I put it into a sketchbook I never forget it, and
    I never have to look it up again.
Source: http://books.google.no/books?id=1YN3kc31nqAC&pg=PA148&lpg=PA...

Jill Price (the Woman who remembers everything)[0] also compulsively kept a journal. What is it about the connection between writing instrument and human brain that doesn't extend to QWERTY? Price also didn't refer back to her journals from what I can tell.

[0] http://en.wikipedia.org/wiki/Jill_Price

Yes, he may have disagreed, but most of us aren't savants with a perfect recall strategy such as this because, well, we need the inhibition to support other functions.

This is also true for me, I almost always took notes in college but it was rare for me to read them except to retrieve formulas that weren't in the course notes.

I always thought journalists were hopeless at reporting precise details, but when you compress your notes using such a lossy technique as this, it's not surprising that the facts can get skewed.

There are different types of shorthand. Some are phonetic; some are spelling based. Journalists should be learning teeline which is spelling based.

Journalists should not be losing details and if they are that's their failure to take correct notes, but not an expected feature of using shorthand.



Fwiw, I've never met a journalist who knew short land although I'm sure such exist. (I've written plenty of articles myself and don't take shorthand.) I sometimes record interviews, which is good practice for an in-depth interview. For lots of purposes though, you're looking for one of two money quotes and other material that you paraphrase or use as background.

In my experience, when I'm quoted, issues I have aren't so much that "I didn't say that" but that I said something longer and more nuanced which wasn't captured in the quote in question. It's one of the reasons you learn when giving short interviews to hit your key points precisely in a way that doesn't depend on a lot of context.

I got so excited when he said that it really was possible to write at 100 words a minute but then he said that it's nearly impossible to actually learn Gregg, I wonder if there's a website somewhere....

About the Livescribe: it's an okay pen but difficult to hold. and somewhat awkward because of the camera at the tip, but it could depend on the way you hold pens.

Just want to add that there are other shorthands which are pretty solid (and more easily available). For example, Pitmans: http://www.long-live-pitmans-shorthand.org.uk/

I'd advise against Pitmans, unless you plan to take it very seriously. Pitmans is one of the more complicated forms of shorthand, and relies on differentiating between hard and gentle lines.

I found Teeline, for instance, much easier to get going with.

I think it's only useful to learn if you actually do a lot of writing, especially in time-sensitive situations like dictation or interviews. If it's just something you're interested in as a novelty, but you don't actually do more handwriting than a few notes during meetings at work, you're probably not going to get enough practice to have any success.

I wonder how difficult it'd be to write OCR-style software for Gregg. It'd have to be more sophisticated; character recognition alone wouldn't be enough, you'd have to do a bit of Markov modelling with most probable sentences/words too. I guess it's so niche no one will ever do it, but it sounds like a fun project.

I used to have a palm pilot and I LOVED the 'graffiti' input mode with the stylus. You had an alphabet of simplified shapes and it made it very easy to enter something quickly without needing to ‘hunt and peck' on a miniature keyboard for the shift key or using shift keys to change modes.

I wondered why nobody brought gesture-based inout to iOS this way. I don't care about 'handwriting recognition' of my own writing, I want a more efficient input method, even if I have to learn it and practice to be good at it. I'm not content with tapping one button for each character I want to record :S

I think an excellent input method for mobiles is Swiftkey. It has incredible accuracy, and it learns really fast as to what and how you type. Its intelligence beat all other keyboards that I tried, and I'm able to peak at over 60 WPM using SK with two fingers, and 40 WPM with 1 finger and flow. However, it gets tricky as soon as you get sloppy or start writing technical stuff, so I tend to stick with the traditional tapping method of input.

There is a little bit of a learning curve to get optimal speed, but once you learn all the nuances and the shortcuts, it gets absolutely great. Typing long and detailed text is now reasonable on a phone, and quickly replying to something takes no time at all thanks to flow in my opinion.

A friend of mine swears by Swiftkey, too. They seem to have taken a turn for the evil on Android recently, though: still usable as ever, but bugging you every once in a while to try out new themes, or asking for overly broad permissions.

Hi there,

You can actually turn off notifications from the app by: 1) SK settings 2) Advanced 3) Uncheck the box

SwiftKey has also stated they are working on performance updates in the next release: http://swiftkey.com/en/blog/swiftkey-5-performance/

As far as the notification at the top of the screen this is a notification provided by the Android OS and cannot be turned off or hidden.


Ryan Community Manager @ SwiftKey

yeah swiftkey is cool. I just tap normally, but you can be extremely sloppy with it (increasing speed) and it'll correct your mistakes as you go.

Only dig on it is that it leaves up a notification that you can't get rid of on some android phones. That's more an android issue from what I read though.

I've had far better luck with Google keyboard than with Swiftkey, despite having been a paid user of the latter for years. Google's correction methods are simple and quick. SwiftKey feels clunky. And, even with connecting SwiftKey to my Facebook, Twitter, Gmail, etc to learn, Google keyboard is still more accurate.

I started on graffiti as a keyboard extension for iOS 8 during WWDC:


Spent too much time simultaneously playing/fighting with Swift so it's far from complete (only has shape definitions for a half-dozen letters, and some are more accurate than others) but pull requests are welcome.

> why nobody brought gesture-based inout to iOS

It's available for android:


Wouldn't character recognition be sufficient for Gregg? I don't know Gregg, but from the article ti seems like the interpretation of the characters could be left to the reader.

I don't see the utility of shorthand anymore. If you're just taking short notes, you don't need to write 225wpm...and if you've recorded everything, then you can listen to the recording later and take notes or type up a transcript verbatim.

It would seem better to learn a stenography system so you don't waste time inputting actual text, which surely you would need to do anyway? (How is the time the author saves writing in shorthand not negated by the time it takes her to type it up afterwards?) I guess if you knew shorthand and steno, you could do everything pretty quickly.

There are specific professions in which it comes in handy. Journalism is one of them. Every now and then, you'll encounter someone who is willing to go on the record, but for whatever reason, doesn't feel comfortable being taped. (And by "every now and then," I mean somewhat greater than half the time.) And a lot of people are put off by the act of typing into a keyboard or tablet while you're speaking to them, even if you're making eye contact all the while.

I'd say any profession in which you're on the go and often have to take notes by pen and paper, shorthand is a necessity. This is doubly true in my case, as a partial disability in my right hand makes longform pen-writing a PITA.

I guess I was thinking in the context of the author using this particular pen for shorthand - it records audio, so even if he's just writing shorthand, it's still recording.

Really good points though, it didn't occur to me that someone would be willing to go "on the record" but not be actually recorded.

There are some other solutions that use keyboards that wouldn't be as intrusive as a keyboard or tablet (and may be something to consider for you personally if you have a disability in only one side); things like portable one-handed chorded keyboards: http://www.handykey.com/ and http://chordite.com/

Yeah, it's kind of weird that people would go "on the record" but not want to be recorded. But there's something about pulling out a tape recorder (or in my case, an iPhone with a recording app) that freaks people out. I think part of this fear stems from the age we live in, with the 24 hour news cycle and "gotcha" journalism. People half suspect you're going to use something they said against them, as if it's a police interrogation. :)

I like to reassure people that I record them for my own accuracy, and that it actually behooves both of us to do a recording. Even still, about 50-60% of people will decline. Younger people seem more likely to decline than older people, at least in my anecdotal experience, which strikes me as interesting. There might be something worth exploring there.

Yeah. I do press stuff as a volunteer from time to time, and I've learnt the art of speaking for half an hour with no part-sentence being an abusable pull-quote, after a few times where precisely that happened.

(Simple method: when speaking of tricky matters, twist your grammar and pepper with adjectives and adverbs. Always follow with a pullquote you'd like used. Use variations on your favoured quote a few times.)

This sounds interesting, but it's not clear how it's done. Could you give a few examples?

I know a few people who maintain eye contact while typing, for instance, while I'm running a meeting. Man, it always feels creepy to me.

Am I the only one that gets that feeling?

I do this pretty often, and people pretty much always complain. Less so people my age (college-aged), but a lot of professors, TAs, and my parents all claim that it's entirely too creepy. My parents have actually told me that they'd prefer if I just looked at the screen the whole time.

But on the other hand, why would I look at what I'm typing? I already know what it says- I hardly have to read it again. A good typist knows when they've hit the wrong key and should be able to correct for it without having to read.

I wonder if college age folks are more likely to be okay with it or if they're less likely to say something about it? Probably some combination of the two, I'd guess

I usually type mostly looking at the screen and then glance up and make eye contact or nod during pauses, to re-establish that I know the person is there and I'm paying attention to them.

I don't think it's so much a practical thing as purely social. I don't think there's anything inherently wrong with looking at a person while you type. It just doesn't seem socially typical. It might be the normal thing to do at some point, in which case I'll adapt. It just doesn't seem to be the case now. It'd be interesting to do research on this, though ,and see if there's a significantly different perception between age groups.

I'd guess that most folks on this board are pretty proficient touch typists, and could look wherever they want while they're typing. Even if you're only reasonably good at touch typing, you're probably not going to mangle the words to the point where you can't run a spell check over it to fix most of your mistakes.

I'll personally probably stick to mostly looking at the screen for now, since it seems like more people are made uncomfortable by looking at them, than looking at the screen, but it's interesting to think about where it might go.

To me, it's worse. It sends the message "I'm pretending really hard to focus on what you're saying, but I'm really doing something else right in front of you!"

I say this as someone who can maintain eye contact and type while giving attention to the person I'm looking atit.

It makes people more uncomfortable than simply looking at the screen while you work and talk.

I'm the same way, and it's not a great skill to have. I've been "called out" a number of times by people who think that I'm not paying attention to them while reading/typing something on my computer, but can in fact usually recite everything they said recently verbatim.

"Maintaining eye contact" is pretty creepy regardless of whatever else is going on.

He only types up small portions of (apparently) long interviews, for quotes. In this case, the combination of shorthand and the camera and software is like a jump-to-where-I-mean feature for audio recordings, which sounds amazing, frankly, for this use case.

It does sound really cool. I just don't get why the shorthand part is important. Why not jot "grandma" and "poodle" and then tap the pen there for when the interviewee starts talking about his grandma and his poodle.

I was thinking about this. The pen sounded like a great way to create a fluid index of there conversation in a really simple outline form. No if it's setup to work like that at all

I'd actually really like something similar with a computer while I type and record.

Voice recognition software might be good enough to create the index.

Yeah. I think I like the bullet-point associations because I can provide a more abstracted index word and associate it with a point in the conversation, versus having to do a text-search, if that makes sense.

It wouldn't be hard to throw something together, either. Really I just need it to pull a timestamp from the recorded audio after every word I type and associate some meta data and the timestamp with that word location.

Might be a fun after-work project. I'm sure there are already solutions like that out there, though

If you have time to learn 200 WPM stenography then you have a lot going for you, but why not become a professional tennis player or a concert pianist instead? The promise of coding or writing at the speed of thought is actually not so good, it's like that horror story of the monkey's paw actually. Of course if you can't speak, a talking steno would be ideal for participating in a conversation.

Good luck with that as a Court Reporter.

Almost all journalists start out at regional or local publications covering Courts. It is considered the apprenticeship of mainstream news journalism.

Because a small notebook & pen is half the size of a keyboard, quieter, and doesn't need to be charged every day?

Shorthand systems also exist in other languages. I'm aware that there was something called a "fast writing" system in Korea and a version of Pitman-Graham in Japan.


For programmers, I don't think shorthand is the right technology to learn - stenographic typing has the same kind of advantages and speed, and directly inputs to the computer.

Learning shorthand does make stenographic typing easier to learn (and vice-versa), since it's the same sort of syllable-based contraction.

For a stenographic typing solution that can work on a pc check Plover: http://plover.stenoknight.com/

Here's the 2013 PyCon talk given by one of the Plover developers, Mirabai Knight: https://www.youtube.com/watch?v=Wpv-Qb-dB6g

The shorthand is based on sounds, not spellings. So the difficulty is that the brain has to translate spelling into sounds first, which is a great barrier to overcome.

There is a similar fast writing system (速写)in Chinese. I tried to learn it for a year while I was in high school, but gave it up eventually. Other than the extra difficulties of logograms vs phonograms, another barrier is that one need to translate the fast writings back to "normal" ones for easy reads later. Many times it is so hard to tell what I actually scribbled down.

> So the difficulty is that the brain has to translate spelling into sounds first, which is a great barrier to overcome.

It doesn't look like it's meant to be used to copy written text, but to take notes that have already been spoken, which is exactly the opposite to what you're describing - When I speak and you take notes, you essentially turn sounds into spelling.

I remember finding a textbook on Gregg shorthand in my grandmother's attic when I was 10 years old. It fascinated me (secret writing! esoteric knowledge!) and I spent a few afternoons practicing and inventing my own abbreviations, even using it to take notes in school once or twice, but my family bought a home computer not long after that, and all my energy went into learning BASIC.

I was really hoping this article would end with the author revealing the Livescribe pen could expand Gregg shorthand written with it into standard characters on a computer.

It would have to be a author-guided OCR though as the same symbols can represent different words.

You can probably get pretty far with machine learning.

Doesn't this have audio?

Wait so that wasn't the point of using it? The author mentions custom OCR so I assumed he was using some sort of Gregg OCR solution.

Does he mention a custom one? I just saw this part: But it doesn’t work for me, I explain, because even though I’m recording this interview with the latest model Sky wifi smartpen, I’m taking notes using a 19th Century technology called Gregg shorthand.


UK journalists still do teeline shorthand. I used to have to help upload practice dictations for a colleague who taught teeline.

Pitman's shorthand was the one used by secretaries when they took dictation. I was fascinated my my Mum's shorthand books as a nipper.

This reminded me about my Pocket PC, which used to have a writing recognition mode that was a cross between Palm's stokes and regular writing (I could've been a 3rd-party keyboard). It was more accurate than regular writing recognition that Pocket PC had and thus faster for me to use. To this day my handwriting is altered by the way I was writing on that keyboard...

I have been using the Echo Livescribe pen for 4 years now, and it has saved me on many occasions. I taught myself Quickscript before I ever had the pen, but because of the lack of OCR for it on the pen, I don't really use it much.


Loved the exposition on Gregg shorthand, I tried and failed to master it in college sadly, but it also gave me new respect for the archivists who decipher notes on the margins of older texts, as those seem to be in their own shorthand, sometimes invented by the author of the margin note.

"As a journalist, I begin most interviews by..." ...talking about my cool pen and Gregg Shorthand for a couple of minutes. Well that's a great way to spend the possibly limited time your interviewee has to talk to you.

When hand-held tape recorders were new, I'm sure many journalists opened with "Have you ever seen one of these?" shortly before asking "Is it OK if I record this conversation?". Seems like a perfectly reasonable 30 second exchange to ensure that the interviewee is informed and comfortable.

When I do user interviews, I always try to talk a little bit at the beginning about something not directly relevant. It's especially good when I can offer something about myself; if I open up to people, they'll often open up more to me. It may look like a waste of time, but the better results say otherwise.

> Gregg eschews the parade of silent letters, like the “y” in “bay” that make English so difficult to learn as a second language.

I stopped reading here. I can't take this drivel seriously.


At least for me - a native German speaker with English as a self taught second language - I find that the `y` in bay actually makes it easier to figure out the correct pronunciation.

Since `a` is pronounced as `ah` in German, just having `ba` would make the `a` sound like the one in `bra`, which would make it hard to distinguish between `ba(y)` and `ba(h)` when first learning the language.

Removing the y's would also leave one with seemingly strange words like `awa`, `ga`, `sla` and `sta`. Which quite honestly to me, look much more like mnemonics rather than actual English words.

I thought that at first as well (about the potential confusion), but then recalled that most of the "possible" alternatives aren't real English words:

"bah" is really only a sort of exclamation and the sound of a sheep, not something you'd really have appear in an interview.

As you pointed out, "awa(h)", "ga(h)", "sla(h)" and "sta(h)" aren't English words [1] that you'd confuse in context so "away", "gay", "slay" and "stay" are the only possibility for them. The "y" really is not necessary.

I really want to learn Gregg now...

1 "gah" like "bah" ends up being an exclamation, and "slaw" as a food... but I can't think of a case where it would be confused with "slay" in context.

EDIT: Clarification on gah, awa, etc

The 'a' sound in "Slaw" is a bit tricky. it rhymes somewhat with the 'o' sound in 'ostrich', 'ought' or 'taught', which are written with the symbol for an 'o'. Oddly though, the word "Father" is written with an 'a'. If it's really necessary to distinguish between these sounds, Gregg Shorthand does allow for diacritical marks over vowels, but they are rarely used, since the meaning is usually clear from the context (and they slow you down).

It may help to realize that John Robert Gregg was Irish - so imagine his somewhat British pronunciation of vowels. Since Gregg shorthand is written phonetically, the words "father" and "farther" are written pretty much the same . That becomes a bit confusing to an American speaker of English.

I taught myself shorthand in high school. I love foreign languages and found it fascinating. A "Secret Language" like someone said above.

But I agree with one of the other posters, if you do not use it often, you will find it very difficult to do well. The basics of Gregg shorthand are incredibly simple, but it is another thing entirely to master - to be able to write quickly and accurately takes practice. Still, it is not impossible and you use it quite readily only knowing the basics. Unfortunately it has become a lost art except to a very few.

It would be quite an undertaking to write OCR software for Gregg shorthand because the writer has a lot of freedom to construct abbreviations on the spot or join several small words together when convenient (rather like native Germans can come up with compound words you won't find in a dictionary). The other problem would be clarity. While Gregg Shorthand doesn't have rules like Pittman for placing certain strokes on or above a line, proportion is quite important because several letters share common shapes ("n and "m", "p and b", "t and d", "f and v" are all quite similar but vary in length or height).

So the writer would have to be consistent in their proportions and the software would need to be able to learn the writer's style.

"only possibility for them" is not "the letter 'y' is silent", which the over-downmodded OP is complaining about. That it's not necessary for shorthand is not the same as being a silent letter.

For example, the vietnamese "chicken noodle soup" is spelled phở gà. It's pronounced something like fur gah. Every native English speaker I've heard (including me) that has seen that term written tries it out first as foe gah. If what the article was saying is true and the letter 'y' was irrelevant in such words, that first attempt would at least sometimes be foe gay - and I've never heard it that way at all.

Regarding 'sla', it's probably not going to be mistaken in context, but 'sla' is also a three-letter acronym that is moderately common - Service Level Agreement.

I'm a native english speaker I'd make the same assumptions about "ba" as you would. I'm not sure what accent "ba" would be pronounced "bay" in.

Maybe it works because there is no english word written 'ba', and in fact no english word that's 'b' followed by vowel(s) and nothing else, except for 'bay'. So there's no other word it could be.

But yeah, the extreme form of gregg shorthand definitely seems, from the OP, to rely on lots of contextual knowledge for it's concision. That seems to be it's strategy, eliminate letters where contextual knowledge will suffice to reconstruct them.

> and in fact no english word that's 'b' followed by vowel(s) and nothing else

  "beau", "bi", "be", "bee", "bao", "boa"
came to mind off the top of my head and are all in the dictionary; there are probably others. Of these, "be", "bee", and "boa" should be pretty non-controversial both in terms of "word" and "English".

For the original issue, "ba" pronounced with the actual letter names does in fact sound like "b-ay". Whether that's what Gregg shorthand does, I can't tell.

One where the graph 'a' referred to the long a sound only?

I think the main reason why unpronounced letters remain in English orthography is to give hints about the language of origin, which is used (usually subconsciously) to determine pronunciation.

In this case, you have explicit phonemes in some cases (especially for digraphs), and context. It shouldn't be too hard to figure out which word was intended.

interesting link: http://zompist.com/spell.html

The vowels in Gregg shorthand refer to one sound only. So it's more like writing 'bā' instead of 'bay'.

As a non-native speaker, I never really tried to reason about what letters are silent or not. It's all seemingly nonsensical, anyway, so I might as well just take it at face value.

Are you aware that less than 15% of English words have irregular pronunciation (which is not to say the "regular" rules aren't baroque). And the pronunciation of "bay" is certainly regular: it's not that "y" is silent, but rather "ay" is a digraph that is pronounced in a regular way.

> (which is not to say the "regular" rules aren't baroque)

Well that's my point! I haven't found it useful to figure out or reason about whatever regular rules there are, because they are baroque to the point that I find it easier to just go by intuition, for the most part.

Well that's what most people do -- it's the regularity that allows you to have an intuition. For all the baroqueness, eqrtlk will never be an English word and ghoti will never be pronounced like fish (despite claims to the contrary). I didn't mean to suggest that you should try to learn the rules from a book; I just meant you are wasting effort if you learn the spelling and pronunciation as completely independent items.

Registration is open for Startup School 2019. Classes start July 22nd.

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