You need to enable experimental features in `prefs_user.config` by adding `enable_experimental_features 1` and then enable this mode using `toggle_fastread` command within sioyek.
I also experimented with more sophisticated algorithms, you can read about the experiment here:
(Developer of the Github repository in the post here)
Is it possible for me to easily integrate with sioyek to provide PDF converting ability directly within the Chrome Extension? This would be useful for the following reasons:
- For the times when you click on a link and it opens the PDF in Chrome
- Many people I know use Chrome as their main "PDF Viewer" app.
I believe this will be a great way for both our OSS projects to collaborate.
Perhaps in the future we can work on the language model as well. Very cool reasearch btw!
Correct, it's not possible in default PDF view in Chrome -- contentscript is not injected. However as long as you have full access to pages, you could intercept a PDF link, render it with, say, PdfJS, and start from there. Maybe!
I saw the stories about Bionic Reading a few days ago and ended up going down the rabbit hole for a bit. What interested me was that it wasn't the weight of the letters that helped me read quicker but the shorter lines.
The examples with and without Bionic Reading formatting were just as easy for me to read but where it fell apart was when I read something with more words per line. Put 12+ words on a line and the Bionic Reading formatting didn't help.
Having an average of 5 words per line is so much easier for me to read it kinda blows my mind.
My (totally pulled out of my ass) guess as to why this is the case (for me anyway) is that less horizontal eye movement reduces cognitive load and allows for easier processing of what the brain is seeing.
I have no way to verify this, but of course speed <> comprehension. It's better to read a book once in 100 minutes and comprehend everything, than requiring two speed-reads of 60 minutes to reach that same level of comprehension.
Not to mention newspapers. I spend so much of my life looking at screens that I enjoy reading the Washington Post print edition over breakfast. My one year old calls it the "daddy paper."
And TikTok, it’s a really stupid name for something that, at first blush, really seems to help me read. It’s not really about speed, I don’t care if it helps me read faster, I care that it provides me enough visual stimulation to make it possible to read without pain.
I saw this on my twitter and was kind of surprised that this was patented and had an expensive API to go with it. I don't see why this couldn't be on-device processing. Why utilize a cloud API for an accessibility feature?
IANAL but it's probably not a licensed use of the patent.
That said, that a patent exists does not mean the patent holds up under scrutiny. Software patents in particular are often not worth the digital paper they were printed on.
THAT said, this doesn't mean patent holders won't litigate or at least request a takedown.
Is there any actual research/study on the efficacy of bionic reading? If it improves reading speed / comprehension significantly then it's really a no-brainer.
I did a wholly unscientific test on the example listed on the bionic reading website. I finished reading the normal example in 12 seconds, and the bionic reading example in 11.5, which I'd say is within the margin of error/normal variance. I don't see the advantage of this, but I'd love to see evidence that suggests otherwise.
I think the hype is being driven mostly by a subset of ADHD folks who have discovered that this might be the “trick” that stimulates them enough to be able to read.
Like I really don’t know how to describe it but the lack of stimulation is debilitating and manifests in physical pain that I basically have to sleep off like a headache. Unless I can hyperfixate on a book I pretty much can’t read and I can’t control what the gremlins in my brain choose to hyperfixate on. It’s a miracle I make it through English.
Agreed, this sounds similar to Beeline reader (https://www.beelinereader.com/) which I didn't find helped my reading speed (per their test as well) despite lots of studies cited.
I found that beeline did help me, to the point that I briefly paid for it. I stopped when I realized it was reporting back every site I visited to the developer’s Google Analytics.
Hi Paul, BeeLine creator here. We do log URLs visited on an anonymous basis, so we know where the tool is being used and where we should optimize. These data are never linked to individual user accounts, and I think we'll remove this logging soon from the paid version, since we get enough data via the free users. But even with the free users, we never link this data to users or email addresses. This is detailed in our very short privacy policy, as we take user privacy very seriously.
Please reach out (contact in profile) if there's further info I can provide on this, or if you have other questions. We launched on HN many years ago and have been glad to have so many HN/YC'ers using our tech on their devices and in their apps.
I too have been wondering if this is mostly placebo.
For familiar words, don’t our brains automatically read entire words as a single fixation point? I.e. we don’t read in terms of letters, but see the word as an atomic unit?
I would guess that my brain, by seeing certain words hundreds of thousands to millions of times, has found a more efficient way to identify it than the same word with a few letters bolded.
Of course, this is based on no data, I’m genuinely curious if this helps.
My n=1 is that it makes me slightly slower but probably within the margin of error. I find that I recognize whole words at a time, and the shape of the word is unnatural to me when half-bolded like this, which might inhibit recognition slightly. Perhaps after an adjustment period it would be useful, but in my current state, it is not.
If someone would like to try this in Firefox (or other browsers), one could simply install a userscript manager browser extension such as Greasemonkey and use the code from convert.js.
Here's an example of how I added the functions to all websites in Firefox/Violentmonkey: https://pastebin.com/ZE75x0Uw
It's a bit strange reading that font style. It almost feels like my brain understands it unconsciously before I do. Like suddenly strapping on ice skates and going 20mph.
Years upon years ago I read about a trick to speedread by just flashing words in giant font at a set speed; your eyeballs don't need to move, and you can force-train yourself by speeding it up.
I wrote an Objective C Mac OS app (which I normally never do, so that was quite a bit of messing with completely unfamiliar tools such as Xcode) and the rarely used 'services' menu so that I could select any text and fire up this app.
It worked wonders for prose. It doesn't work for highly specific technical reading, or anything where the formatting itself is important, as that is, naturally, obliterated.
A mac or two later and I don't really want to build it again.
But, does anyone know of an app or site that does this for you? Possibly with some nice advances such as:
* Give you a little more time on complicated words.
* Possibly let you click a paragraph in e.g. a webpage and then just blast that paragraph past your eyes, letting you mix reading and looking at the layout. This sounds great for wikipedia, where there are images to look at, table data, paragraphs, etc.
If no such thing exists I might be tempted to write it as a browser extension.
When I discovered it many years ago it was first called "spritzing". There are many online apps and extensions that can do this, quick google - https://www.spreeder.com/app.php?intro=1
Nice. I read about Bionic Reading[1] few days back and I was thinking it would be lovely to have the options to try this out. I was thinking more on how do I serve Bionic Reading on my websites.
Likely because the author wants to make money from it. Which is funny because it looks a whole lot like an extension of various Dyslexic hacks used to improve readability. Previously discussed here: https://news.ycombinator.com/item?id=30787290
I've always found these "make things easier for dyslexics" things to make reading far far harder than just picking an nice clear font and contrasty colour scheme. I know a few folk who find that sticking the text on a greeny-yellow background helps them massively and they have a setting on their PC that makes their monitor look knackered, like it's lost the blue gun ;-) But that does seem to make a difference.
I don't find wobbly hard-to-read fonts help at all.
Had to read the source to figure out what this does! It bolds the first few letters of each word to make it easier for the human eye to fixate on those points.
There are also variants where they will turn each line into a colorful gradient to make it easier to know which line you're on and thus move between lines more easily.
I have been thinking of making a "syntax highlighter" browser extension for English text, so words would be colored according to their grammatical function. I'm also wondering of ways to identify keywords (based on how rare the word is?) in text and make them stand out (color, size?), to facilitate finding information in large blocks of text.
I actually found this quite jarring. I'm a decent speed reader (by no means the best, but generally able to consue vast quantities of information quickly), and having these artifical focus points completly threw my rhythm. I can only assume that I've interally developed my own way of skimming that achives the same thing, though not in the same way. So having the first letter bolded actually disrupted whatever my brain was attempting to do.
The idea behind bionic reading is similar to why I find fonts with medium-to-small x-height more readable that fonts with large x-height. Although large x-height probably makes each individual letter more readable, more pronounced ascenders and descenders allow to read words by their shape more easily, and give anchor points when reading long paragraphs. I find serif fonts found in old books to be the most readable.
A friend of mine said it would probably be even better of it wasn't automated, but someone manually made parts or the text bold according to reading flow.
For me, it seems to just give my eye a quick target to lock onto, from there I see the entire word (as well as the surrounding letters) in my peripheral vision.
https://github.com/ahrm/sioyek
You need to enable experimental features in `prefs_user.config` by adding `enable_experimental_features 1` and then enable this mode using `toggle_fastread` command within sioyek.
I also experimented with more sophisticated algorithms, you can read about the experiment here:
https://ahrm.github.io/jekyll/update/2022/04/14/using-langug...
Is it possible for me to easily integrate with sioyek to provide PDF converting ability directly within the Chrome Extension? This would be useful for the following reasons: - For the times when you click on a link and it opens the PDF in Chrome - Many people I know use Chrome as their main "PDF Viewer" app.
I believe this will be a great way for both our OSS projects to collaborate.
Perhaps in the future we can work on the language model as well. Very cool reasearch btw!
use_heuristic_if_text_summary_not_available 1
The examples with and without Bionic Reading formatting were just as easy for me to read but where it fell apart was when I read something with more words per line. Put 12+ words on a line and the Bionic Reading formatting didn't help.
Having an average of 5 words per line is so much easier for me to read it kinda blows my mind.
My (totally pulled out of my ass) guess as to why this is the case (for me anyway) is that less horizontal eye movement reduces cognitive load and allows for easier processing of what the brain is seeing.
Has anyone else found this?
I assumed it was this shorter lines effect that Pocket, Instapaper, etc put wide margins around your articles and large text.
https://www.huffingtonpost.co.uk/entry/what-is-bionic-readin...
That said, that a patent exists does not mean the patent holds up under scrutiny. Software patents in particular are often not worth the digital paper they were printed on.
THAT said, this doesn't mean patent holders won't litigate or at least request a takedown.
Like I really don’t know how to describe it but the lack of stimulation is debilitating and manifests in physical pain that I basically have to sleep off like a headache. Unless I can hyperfixate on a book I pretty much can’t read and I can’t control what the gremlins in my brain choose to hyperfixate on. It’s a miracle I make it through English.
Please reach out (contact in profile) if there's further info I can provide on this, or if you have other questions. We launched on HN many years ago and have been glad to have so many HN/YC'ers using our tech on their devices and in their apps.
I would guess that my brain, by seeing certain words hundreds of thousands to millions of times, has found a more efficient way to identify it than the same word with a few letters bolded. Of course, this is based on no data, I’m genuinely curious if this helps.
Firefox support will be ready in under 12 hours :)
Here's an example of how I added the functions to all websites in Firefox/Violentmonkey: https://pastebin.com/ZE75x0Uw
Here's a screenshot of what it looks like: https://imgur.com/a/bP3bUFP
Couldn't find a vim plugin yet.
I wrote an Objective C Mac OS app (which I normally never do, so that was quite a bit of messing with completely unfamiliar tools such as Xcode) and the rarely used 'services' menu so that I could select any text and fire up this app.
It worked wonders for prose. It doesn't work for highly specific technical reading, or anything where the formatting itself is important, as that is, naturally, obliterated.
A mac or two later and I don't really want to build it again.
But, does anyone know of an app or site that does this for you? Possibly with some nice advances such as:
* Give you a little more time on complicated words. * Possibly let you click a paragraph in e.g. a webpage and then just blast that paragraph past your eyes, letting you mix reading and looking at the layout. This sounds great for wikipedia, where there are images to look at, table data, paragraphs, etc.
If no such thing exists I might be tempted to write it as a browser extension.
1. https://bionic-reading.com
I don't find wobbly hard-to-read fonts help at all.
The trademarks and patents speak for themselves.
There are also variants where they will turn each line into a colorful gradient to make it easier to know which line you're on and thus move between lines more easily.
I have been thinking of making a "syntax highlighter" browser extension for English text, so words would be colored according to their grammatical function. I'm also wondering of ways to identify keywords (based on how rare the word is?) in text and make them stand out (color, size?), to facilitate finding information in large blocks of text.
ill put this here to exist