Plover is a free, open-source stenography engine | Hacker News Reader

Plover is a free, open-source stenography engine

openstenoproject.org

182 pointssmitty1e124 comments

One thing on the rhyming - you say on one page it rimes with 'lover' and on another that it rhymes with 'hover'. In many people's accents, these are spoken quite differently...

And both options are different to how I first read it (rhymes with clover).

That's how I always pronounced the bird (rhyming with clover), but looking it up now, it does actually rhyme with "lover." It's just "lover" with a "p" in front.

(And it's the same for US vs British English, that is, they can both use the rule "p + lover" and pronounce it correctly respectively.)

"Hover" is more ambiguous, though. It can rhyme with "lover" or have a rounder "o" sound.

It's funny because I wouldn't have thought twice about how to pronounce it if it just never mentioned the rhyming thing, but now I can't stop saying plover, hover, lover. Plover, hover, lover.

I've never seen a title like "Redis (rhymes with "Candice") is an in-memory..."

But you have probably noticed that close to nobody pronounces debian correctly (deb-ian, like the names), or gif (whichever is right), or PNG (ping), or the people pronouncing dogecoin as dog-e-coin, etc.

Confusion about how to pronounce something is extremely common, due to most open source communication happening in written form, and English being a weird language where pronunciation and spelling seem mostly unrelated to each other.

Maybe we need to have a latin/binomial nomenclature like in species naming. So pluvialis apricaria instead of golden plover. mensa elephantus for PostgreSQL. mensa parvus for SQLite.

The purpose is to have an agreed name in all languages, and not worry much about pronounciation. Now what about Oracle...

mensa pythia?

mensa diabolica?

Freaking English. Can't we revamp the dictionary to make the alphabet truly phonetic? So many of the words have dubious pronunciation. To the point that if you have just read a word in a book, there is no guarantee that you can pronounce it.

That's impossible to some degree because English pronunciation varies so much across dialects.

Or rather, you could make writing phonetic but then spelling would have to vary by dialect.

I just realized it's done precisely like that in my language! Never even thought about it. But there is a standard "written" version of the language and dialects are mostly spoken.

Isn't that true for every language?

Nope. Most languages have no written standard, and some have more than one (E.g., American vs. British vs. Australian English, or Nynorsk vs. Bokmal Norwegian).

please specify the language so people can learn things, it's the internet

That's really hard with old languages. Accents and new words surpass the alphabets with time. Turkish for example switched to using Latin alphabet on 1928 and they were able to map every sound to a specific letter. If you know the Turkish alphabet song you can read every text with perfect pronunciation. There aren't even any letter combinations to consider, letters have one sound and are individual.

We have an alphabet that can be used to give something close to certainty regarding punctuation. ɪt wʊd biː ˈkʌmbəsəm - to begin with, but I’m sure we could get used to it.

I think I’d prefer it. I enjoy reading text which has been written ‘in an accent’, where the usual way is to bend the standard alphabet so ets cluen yer en aboot tha propa wuy ta sey et.

The problem here is that a vowel is split in some dialects but not others. Phonemic spelling would not solve this problem.

Okay, so you're saying we should just discard the centuries or even millennia of etymology encoded into spellings because it's better to make sure symbols on the page directly correspond to your preferred pronunciation.

In at least England, this would cause spelling to have regional variations. E.g. 'bath' in southern English is often pronounced as if was spelt 'barth' while in the North it would be 'baf'.

Not while using an alphabet with just 5 vowel letters. Switching to IPA would do the job, though.

5 vowels, and then 25 more vowel combinations (e.g. ea)

I don't think those 25 combinations are enough to cover all the cases of unique vowel sounds, diphthongs, and pairs of distinct adjacent vowels that you find in English. 'ea' for instance is ambiguous between sounds like those in 'great', 'deal', and 'creatine'.

Yes, they are different in standard British English:

hover /ˈhɒvə/

lover /ˈlʌvə/

rover /ˈrəʊvə/

The bird "plover" is like "lover": /ˈplʌvə/

Not only that, but (at a guess) I'd say think the southern /ʌ/ sound in Plover might come out more /ɔ/ in northern English?

https://en.wikipedia.org/wiki/Open-mid_back_rounded_vowel

https://en.wikipedia.org/wiki/Open-mid_back_unrounded_vowel

I know you’re correct, but I cannot for the life of me remember or imagine how people pronounce these differently, even after having it explained many times. Is “hover” done like “over” (long ‘o’)?

> Is “hover” done like “over” (long ‘o’)?

No, it's not that - at least in British English - referring to https://en.wikipedia.org/wiki/Help:IPA/English, "hover" uses ɒ ("LOT, blockade, song") while "lover" uses ʌ ("STRUT, untidy, trustee, sung").

In British English, plover (the bird) rhymes with lover but not with hover.

I always have difficulties in the opposite direction, namely imagining how such vowels can be merged - a classic for me was a programming manual once that told me "char" (the keyword) should be pronounced care, "like the start of the word character". Confused the heck out of me, since where I come from the vowel at the start of the word character is nothing like the one in care.

This British pronunciation of hover https://dictionary.cambridge.org/pronunciation/english/hover does not sound at all like the vowel in 'lot' or 'song'.

I'm British, and sitting at my desk saying "hover" "lot" "song" over and over again. They're definitely very close. I say hover like hoh-ver, which is very similar to the vowel I say in lot (loh-t) and song (soh-ng).

Yes. It does. Source: have been British for 54 years

Hover is “hovver”

Lover is “luvver”

Did you listen to the audio file at the link? I take it the British pronunciation at dictionary.cambridge.org is an unrepresentative example?

No, that's correct and so is what they're saying

Look up lot and song on the same website, it’s the same vowel. Then compare with lover, an entirely different vowel is pronounced.

I did look up lot and song at the same website. The audio examples of British pronunciations provided for those two words use an entirely different vowel sound.

Lover (and plover, the bird) are pronounced with the "o" sounding the same as the "u" in "cut" or "gut" or "puck".

Hover is pronounced with the "o" as in "hop" or "hot".

Fun fact: the "o" [in "lover"] sounding the same as the "u" in "cut" is actually an "a" in spanish.

As fun as that, also in italian. Though of course the British pronounciation of vowels is far from being as "clear cut" as in spanish or italian.

Not in my accent! I think you'll have to be more specific.

(both accents fairly neutral, English from England, Spanish from Spain)

That's curious. I have never heard the word "cut" being pronounced without the "a" sound (in the vecinity of how "cat" sounds in "category") by someone who had a neutral English accent. Where is it from?

...and this is why IPA was invented.

Indian pale ale?

Seems like this has upset some people.

https://www.google.com/search?q=ipa&oq=ipa&aqs=chrome.0.69i5...

First hit on Google is for India Pale Ales.

I always heard the phonetic alphabet referred to as the NATO phonetic alphabet. It doesn't seem very international given that it is entirely in English.

Also Russia has its own version https://en.m.wikipedia.org/wiki/Russian_spelling_alphabet#:~....

So IPA is a bit of a misnomer I feel.

The International Phonetic Alphabet isn't just "Alpha, Bravo, Charlie"—in fact, it has absolutely nothing to do with the NATO "Phonetic Alphabet", and anywhere you heard that referred to as the International Phonetic Alphabet was just simply wrong. It's an actual alphabet, designed to represent all the phonemes humans actually use in speech—not just in English.

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

Either way the acronym is ambiguous at best.

How is IPA ambiguous when talking about pronunciations of words?

If someone has no idea what IPA is, how are they supposed to figure it out when there's a bunch of other things that have the same acronym?

Also, ALL acronyms are ambiguous. That's why it's best practice to say the actual thing you're talking about then put the acronym next to it the first time it's used.

Well I googled IPA and the first Wikipedia hit (2nd result) was India Pale Ale, which is the acronym I am familiar with.

The first link was to https://ipa.co.uk Web results The Institute of Practitioners in Advertising: IPA

So I think calling it ambiguous is ok, you are if course free to disagree, but what I said wasn't fundamentally incorrect. But it seems like HN has a bee in its bonnet.

Ho as in hot, lo sounds like lu from lush.

I think hov-er and huv-er.

This is an oversight. We settled on "lover" a while ago for this reason, it should be updated.

Huh, I learn something new each day. I definitely would consider these words to be "very" rhyming and can't even envision which one mutates in other accents. Thanks!

Just don't pronounce the P as a separate letter.

I don't think that's the problem the clarification is designed to address. A plover is a type of bird, and in some accents it doesn't rhyme with either lover or hover.

Nevertheless, the parent is right, in my accent hover and lover don't rhyme either. All three have different vowels.

In my accent, lover, hover and clover have three different vowels in the first syllable. Plover rhymes with lover for me. Does it rhyme with clover for you?

Yeah, it does.

I suspect it's not the standard pronunciation in my region, since Wiktionary lists it as an alternative American pronunciation and I'm Australian, but for whatever reason that's how I pronounce it.

I expect that many people pronounce "plover" as "clover" if they saw it written down first before they heard someone who knows the pronunciation say it. There are plenty of words like that in English.

I pronounced it as rhyming with "clover" too before learning that it's actually incorrect.

That's possible too, though I don't subscribe to the idea that a widely used pronunciation can be incorrect.

And mover has a fourth, now that I think about it.

I think GP was alluding to the drug P (meth).

A simpler explanation would be "pee-lover".

The big drawback with Plover-type stenography is that it requires a massive base dictionary for the language it's going to be used with. That's a big up-front investment unless you're lucky and someone has already done the work.

In Europe, stenography usually works on a slightly different orthographic principle, meaning you still chord words, but generally you're explicit about each character. That makes it much more flexible in terms of new languages.

I'm still waiting for a solid, free QWERTY implementation of that. I started working on one[1] a while ago but I just can't find the time to get it done and maintain it.

If anyone would pay me competitively to work on anything I wanted, this would be it. Hugely important for marginalised groups in society, seems like it could be very convenient for regular people too.

[1]: https://github.com/kqr/qweyboard/

Plover comes with an expansive English dictionary by default (developed by one of the creators, used in their work over many years). Its ready to use out of the box. It also has support for some other languages, with some other (smaller) base dictionaries.

Which are some other? I can only find French and Spanish (and "beginner Italian") with a quick search. These are the three biggest Latin/Germanic languages -- not a huge selection.

Interesting! Do you know how it deals with ambiguity? I saw the part about using key order to disambiguate which key was the first letter. But what about words with remaining ambiguity? Or words when there's an "s" in the word, but also a plural form ending "s" that you might want instead of the singular?

The honest (and unfortunately short) answer is: No. I don't know.

Maybe it makes sense at least in some instances to repurpose the Z as a plural S, but I suspect it will be somewhat dependent on what comes before it.

Thanks for the answer! Perhaps someone could write a script calculating what percentage of common words can be expressed unambiguously using this keyboard? It seems like the percentage would be quite low to me (considering e.g. 80% to be "low" given how often you'd have to issue slow corrections), but I might be misunderstanding the rules and I expect it would improve in English with your Z-as-S idea.

The general idea has been used for decades in Europe by people doing closed captioning for live television and other types of speed-demanding work. So the idea is tried and true, and I have no worries about ambiguity from that perspective.

The bits I'm uncertain about are the specific details around substitutions and such.

I use Z for plurals exclusively; that's one common system. You might be interested in my post: https://blog.waleedkhan.name/my-steno-system/#orthography

For homophones, I use a different vowel (OEU = oi, the least used vowel sound; AE is an extra vowel sound not representing anything). For proper vs common names, i often duplicate the stroke (e.g. MAT = "mat", MAT/MAT = "Matt").

Why do people make products with names that they know they have to tell people how to pronounce? Why would you make marketing, stickiness, word of mouth etc an uphill problem for yourself?

In my local accent "plover" has the same vowel as "over".

They have these birds called Plovers in California and that's the pronunciation (puh-lover) I immediately went to. https://en.wikipedia.org/wiki/Pacific_golden_plover

There is a bird called a plover with the same pronunciation. I’m guessing this product is named after it.

Think Danny Glover and that’s how most Australians would pronounce it.

I need to launch a product with an ambiguous pronunciation, and then “helpfully” suggest a heteronym as a guide.

I heard three pronunciations of "Wacom" yesterday and I'm still not sure if there's an official one and if so, which it is (if it's not a fourth).

Not really a good comparison (if we're talking about choosing a pronounceable name when incorporating) because the pronunciation is not at all ambiguous* (ワコム) in the original language, and the country it was founded in mostly uses that language.

* - while the ワコム part is not ambiguous, you can technically choose to pronounce the rest of the word (株式会社) differently - since these characters have multiple pronunciations - though choosing to do that would be considered a very strange thing to do in this context, since its otherwise got a standard pronunciation.

Probably because the people making the products are developers and not concerned about or aware of such things.

Developers, rhymes with interlopers

"Check out this open-source project!"

HN: "OK, but how do you feel about dipthongs?"

My keyboard is supported! Never tried stenography, so I hopped to http://qwertysteno.com and tried the tutorial - it makes sense, but has a steep learning curve - you certainly need a lot practice more than the 5 minutes I gave it.

I wonder if something like this for programming rather than writing is feasible, and if one then could code at the speed of thought.

I mean, the speed of code is something like 10 lines per hour for good code... Sometimes less or even negative so I'm pretty sure the input method is not the bottleneck. That said I still use Vi mode in vs code so maybe there is something to it.

But to do that, you often have to navigate through thousands of lines, and you use the keyboard for that, too.

Tried plover (with a Georgi) for a few months.

It is fun, addictive almost to learn, but it would take a long time before i feel comfortable to write in it on the daily.

I think learning it for ergonomic reasons is far more beneficial than any perceived speed increases, unless you're really willing to sit down and grind.

How long does it usually take to learn stenography? If my regular typing speed is about 75 wpm, what kind of wpm can I expect to reach?

In the school I went to, 2 years was the expectation with graduation requirements being 240 WPM with 98% accuracy.

How quickly you actually got there was very much dependent on the person. Some do it quicker than 2 years with little practice, some do it slower than 2 years with a lot of practice, and some appeared to be unable to reach the requirements at all.

I don't really remember too well at this point but it maybe took me a few months to get to 30 - 40 WPM, maybe a year to get to 120, and year and a half in total for 240.

I don't think it's practical or worth it for the vast majority of people to learn stenography to be honest. That being said, Plover is a nice and much needed initiative, I'm glad it exists.

What was this course? Was it a dedicated stenography course or taught stenography as part of something else?

It was at a small accredited school which offered associate degrees in court reporting (also degrees in a handful of other subjects such as accounting). I had to complete a few other courses not strictly related to stenography, but the degree itself was an associates degree in court reporting.

That's a very healthy qwerty pace. It would many months, perhaps years, of daily practice before you exceed that. I've been lurking in that community and unless you have A LOT of prose to produce which can justify the up-front learning investment, steno is IMHO best approached as neat hobby/game that also happens to get you fast text input ... eventually. Else you are in for a frustrating wait.

Oh, and to your question, very good stenographers head off into the stratosphere at 200wpm+ (elite in 300wpm+) for extended periods, rates not really doable on qwerty except for short burst of predictable prose IMO.

From what I have read, you can go to over 150 in 6 months. Of course, it would depend on how much time you spend on it.

My biggest concern is that there seems to be no standard for programming. I have seen some people develop their own cords, but that seems like quite a time investment to me.

Another think that might be an issue is if you have to be able to type in different languages. I guess you could adapt easily enough as long as the language already has a chord-table, but it also means more time learning.

> My biggest concern is that there seems to be no standard for programming.

The standard would definitely have to vary by language, and you would probably define custom strokes for each defined symbol in the source, similar to the strokes for English words in ordinary steno. It could be workable if the stroke definitions could be stored with the original source code as custom comments, and then imported by your stenotype tool.

Intellisense for Typescript gives you a sort of stenography with zero time investment. I barely type anything in full except definitions or strings.

Mapping dot and Enter to thumb keys on an ergonomic keyboard does wonders in combination with this. I actually have both mapped to the same key (in different layers).

If your IDE is good at predicting which class member you will type, it can be lightning fast.

I think much slower than I can type, so for something where I'm the author, typing speed has never been the limiting factor.

Yet the idea of being able to author something at 250 wpm is utterly fascinating. A novel in around six hours...

People often say things like "I think slower than I type", but I don't think that this is correctly modeling the problem, because it's only true over long timescales.

Consider this: the average American uses only 7 GB of mobile data per month. Would they be happy with 50 kilobits per second? That's much faster than their average usage rate, so in some sense, the data rate wouldn't be the limiting factor.

It's very much the same with typing speed. Yes, your internal word generation rate might be slower than you type if we're looking at an hour by hour, or even minute by minute scale. But once you have a thought crafted, it's beneficial to get it out as quickly as possible.

When I hear Plover I think Mark Jason Dominus.

I offered the Open Steno folks a subdomain of plover.com, so they could set up a redirect or something, but Mirabai said they didn't need it.

Mirabai Knight's talk about this is one of my all time faves: https://www.youtube.com/watch?v=Wpv-Qb-dB6g

Inspired me to try to create a live coding musical instrument based on it: https://github.com/jarmitage/stenophone / https://iclc.toplap.org/2017/cameraReady/stenophone_camready...

I got recommended this on my YouTube over the weekend and learnt about steno, something I don't know much about. https://youtu.be/nRp_1S7cj6A

Fascinating stuff!

Ha, same. YouTube must have been recommending that video to more people.

How relevant is stenography today, now that we have ubiquitous voice recording and voice recognition? What are the use cases? I'm genuinely curious.

I'd imagine that accuracy is a key driver — e.g. for something like a courtroom setting, I think I'd more trust a professional stenographer to record proceedings accurately that a programme that's doing automated speech-to-text and then someone going through to correct errors afterwards while listening to a recording.

As an aside, I'd consider a different name if you need to say "rhymes with" (especially as without that I'd naturally read it as rhyming with clover).

I might be immature, but I originally pronounced it in my head as P-Lover...

Good explanation on the value of human captioning, https://www.youtube.com/watch?v=XKe1O7ppyqQ

voice recognition might be ubiquitious, but good voice recognition is not, and until that changes it's relevant. E.g. professional event live-capturing is going to be steno most of the time, for people doing subtitling on videos it's a useful skill for speed reasons, ...

Wherever the GDPR applies (and possibly similar regulations elsewhere), you cannot record someone's voice without prior consent. So, taking notes in meetings or courts is usually not possible/practical via voice recordings.

This project got me to buy a full NKRO keyboard.

I dislike how every system/theory has their own dictionary for words and in addition to that, it's heavily encouraged to customize your dictionary for the work you're doing. The customization makes it much harder to learn. Think extensive emacs/vim shortcut customization but worse (in my opinion) and much, much more common.

It is fun essentially typing syllables instead of individual letters. When I was learning, I found a lot of mismatches between how I thought a word sounded, and what the dictionary thought. It felt like relearning English.

Other downsides (at the time), was that there's no real slowing down. You're either typing syllables quickly, or you're not writing anything. There's a way to type individual letters but it's pretty slow, and you have to stop plover (toggle it off), to type normally (which is slightly faster).

For those who just want to see what doing stenography is like, there's a demo here that works with compatible keyboards: http://www.openstenoproject.org/demo/

I am currently working on a keyboard with a plan to put plover on it, so that I can use it at work without installing any software on the work PC.

Check out stenogotchi [1] if you haven't already. Plover running on a battery-powered Rasperi Pi Zero.

[1] https://github.com/Anodynous/stenogotchi

That's a cool idea! Are you posting your progress anywhere? I'd love to follow along.

I don't post my private projects because after work and kids I have barely any time for them.

The keyboard is going to be mechanical with Bluetooth, trackpoint and maybe tiny OLED display. The whole point of this is that it runs my software and so I can put on it whatever layout, logic and other shenanigans I want.

I already have it working as both mouse and keyboard and got two sets of Cherry MX silent red switches.

While the trackpoint is working great, I don't yet know how I am exactly going to fit it between keys. An idea I have is to sacrifice one key completely and make it a trackpoint with a custom keycap (rubber red?)

The way this works is that controller detects that trackpoint is in use and maps some other keys on home row of left hand side to function as mouse buttons.

There is going to be microsd for software and configuration updates.

I have not yet started with plover. I just looked at the code and I know this is probably going to cost major effort to get it to run on STM32.

There is also this post from Plover developer on additional challenges to get it done: https://www.reddit.com/r/Plover/comments/6ucy15/stenography_...

I don't agree that this is necessarily a huge problem that can't be resolved.

Sounds neat! Thanks for sharing. My only comment from research (no practical experience) is that I'd worry about the spring force of the MX making chording tiresome. I know the Georgi goes to some lengths to get switches with a lower force.

> I don't post my private projects because after work and kids I have barely any time for them.

I hear ya. I was excited because I've been mentally toying with the same project. But you know, first I need to finish my remodel, fix the truck, build a deck....

Cherry MX silent reds are very light, there should be no problem with chording.

Has anyone tried Plover with Kinesis Advantage? I love my keyboard, I am deeply habituated to use it and I tried many other (similar) boards, but I just don't feel as much comfort with anything else. I wanted to learn Plover for a long time, but I don't think my keyboard is a good option here. What do you guys think?

As someone learning stenography (but with a different keyboard) - try it and see, however your ability might be hindered without the right keyboard, key switches and keycaps.

I bought an ortholinear keyboard with DSA keycaps to get started, which is absolutely fine when you're chording fairly simple words, however once I reached a point where words involved chording multiple keys from the top and bottom rows - it started to become a little bit painful. After switching to F10 keycaps [1], it made a huge difference in the enjoyment of chording more advanced words.

[1] https://pimpmykeyboard.com/f10-flat-key-keyset/

I haven't tried it, it looks to me like it doesn't support N-key rollover out of the box which would definitely be a problem. You can try it out but if you wanted to put more serious time into it, I'd highly recommend picking up something more dedicated like the "Georgi" listed here: https://github.com/openstenoproject/plover/wiki/Supported-Ha...

would stenography help in taking class notes, where the goal is not to reproduce the material verbatim?

would there be additional cognitive load on the brain trying to find the right set of keys, even as one tries to ingest the materials being taught?

You can get proficient to the level where it's the same additional mental overhead as regular typing, unless there's significant domain-specific jargon. It could be argued that the struggle in recording notes is what makes you learn, and typing them up is a hindrance in that sense.

Pardon my grumpiness, but you know a name is sub-optimal when one references the pronunciation right in the title. Or maybe it's another cutesy pun, or an 'in-group' reference. Those always work out, right? :)

This is a great rule. It’s right up there with “if you have to spell out your new kid’s name every time, don’t call them that (or at least don’t spell it that way” and “when you write a comment, explain it to your duck, then delete the original and replace it with what you told the duck.”

I'm a bit annoyed because I have a daughter whose name was pronounced correctly by everyone when she was young (it was an alternate spelling of a popular name, there was a fairly well known actress with it spelled and pronounced that way).

In the past decade though, there's a new name spelled the same way but with a different pronunciation and everyone now says her name wrong... she's taken to going by her middle name now because of how much it annoys her.

Ugh, that’s annoying. Not much you could have done about that one. :S

Reminds me of the movie, That Thing You Do, with the original band name, The Oneders, or as the record producer asks them, "The Oh-need-ers?".

It's an actual word in the English language, so if one were so inclined one could look up the pronunciation in a dictionary.

Brings to mind a certain database frequently discussed on HN… ducks

nice colossal cave adventure reference!

Reply on news.ycombinator.com