hckrnws

I already have the info on paragraph and chapter breaks in the reader already, so this is more of a rendering issue rather than info display. The reason why I particularly like your first suggestion is that it lets you know something like a poem or different paragraph style is about to start so you can change your reading pattern accordingly rather than stumbling into it when you're already in prose mode reading at 400 wpm.

On the 2 modes thing, I feel like there isn't really a good reason to do it the other way. It was more that it happened that way rather than it needed to happen that way, and you're right that the main one should probably scroll and have bookmarks rather than making the user make a choice between 2 different interfaces.

Theme circles is definitely a solid suggestion, being a more practical implementation of visualising the underlying theme. Being able to see the actual colours of the background/text rather than an abstract blob is good, so that's a definite maybe.

Code isn't public yet, but extensions are something I've wanted to try for a while. In fact, someone sent me an email this morning asking precisely for that, to be able to paste a link and read articles. It can't be done directly in the page because of CORS policies, and doing it server side would involve routing the reading of articles through my server which is probably not ideal for everyone. An extension that pulls the article from whatever page you're on would be the most ideal version of this, from a privacy standpoint, so I'm considering it.

And finally, your last point is actually the first pushback I've had all day on the controls/interface. As far as 4 other people having difficulty finding the thing, I think you're right that something needs to be done, but I don't think hiding the options behind discovery is really the answer. There is a conflation in my brain between 'controlls are easily found and accessible' and 'controls are not hidden beneath layers of menus/bookmarks', and I think it's a false conflation, I had been treating them as one. Thanks for that.

Your method also nicely solves that thing that I keep on getting asked for which is copy and paste a url and have it fetch the article and display the text. I had somebody email me about that today! On the terminal you just get something like: curl some_url | readable and you are pretty much done. On the browser it's a cross origin request so you will want to include a server that fetchies it and well that would kinda be missing the point of having local rendering.

Something that i was wondering is if monospaced text would have made it easier for you to find the alignment for the pivot? Fixed-width is simple, since your optimal offset is just a character offset. For proportional text i'm just taking my best effort pivot character and anchoring it, and then extending word fragments from that point in both directions. Otherwise the words can tend to wander as you eye flows from character to character.

Where I think the argument starts to break down is by lumping all reading into a single category. Reading a novel for pleasure, and wading through forty pages of background information before you get to the five crucial ones, are fundamentally different activities. I'm a medical student, and the majority of my reading consists of that second type. You can't really "digest" a literature review in a way that makes it enjoyable. The only real goal is to get through it once, without getting bogged down.

Regarding the pauses, that's actually the area I put the most work into. It lingers three times longer at the end of sentences and twice as long at commas, specifically because a consistent, flat pace feels like reading a script rather than actual prose. And, of course, you can pause whenever you feel the need; there's no pressure to keep moving forward.

The part I'd really push back on is who this is actually for. The people who've given me the most in-depth feedback today are clearly avid readers. It's not really a tool for people who don't read much.

The saved toast is a reasonable hit and frankly is worse than annoying. The whole format is based on your eyes never having to leave the same area, so popping anything up next to the word display forces exactly the saccade you are designed to eliminate. It directly attacks the one function it is designed to fulfil. There’s no need for Auto-save to pop anything up at you either; it never received a request to save and there is no need to echo that action back at you.

Good call on Zen-mode, and I’m of the opinion that is what should be the default, once you start reading; when words begin moving make thechrome fade out, when words stop, make it reappear. I will consider both of these.

Was the control issue the Hold to Read option, out of interest? Just wondering what it was.

What you're getting at, about having a separate speed for each book, that’s a good idea and worth exploring. Right now, the speed is just one setting for everything. So if you slow things down for a book that needs a more leisurely pace, you have to remember to speed it back up manually when you switch to something else. Someone else mentioned the same thing earlier. It makes sense that it should remember the pace for each book, just like it remembers where you stopped reading.

What stood out to me as a major difference was pacing though - a flat interval makes for a pretty robotic reading experience since there's no sentence structure, and I found that using multipliers on the base duration was more effective, getting about 3x for sentence ending punctuation, 2x for commas, and a little extra for long words overall was more noticeable than any other adjustments I made.

I'm curious if DuoBook is actually trying to implement reading in a second language as a learning tool as the name suggests, and if so, how does that affect things? I would imagine the pacing would be overall lower (particularly for punctuation) and recognition of the words you're actually reading would have more impact on comprehension than the ones you're just learning to read.

Epub makes it manageable because the images are just files in the zip. They are positioned at a set point in the spine, so I know exactly where in the word stream a figure belongs. Show it there, pause, and wait for a press to continue instead of auto advancing. You can pace text, but you can't pace a graph. There’s no sensible way to guess how long someone needs to view it.

The detail I keep reconsidering is captions. Right now, a caption flows into the stream as regular words, so you see "Figure 3 shows the relationship between" appearing one word at a time with no figure in sight. Those should be linked to the image and displayed alongside it instead.

Your question also made me realize that tables should probably be images too. Extracting text flattens a table into a series of values in reading order, which is worse than useless. Rendering the region as an image and pausing would be much better than what I do now.

Pdf is more challenging since I’d have to determine which images are real figures and which are rules and logos. Storage increases significantly too; everything is stored in IndexedDB on your device. So, a book full of plates is very different from a novel. I’ll probably need to downscale on import.

Where did you see it before? I’d really like to check out how they handled it.

The avenue that actually looks viable is directory import. Calibre keeps everything as normal files in Author/Title folders with a metadata.opf next to each book, so if I point a directory picker at your library directory, I can traverse it, grab the opf file to get the actual title/author rather than having to guess based on the filename, and then suck the epub files directly in. Would be a properly slick bulk import.

Catch is the File System Access API only works in Chromium at present, so anyone using Firefox/Safari would still be forced to pick their files one by one. Would be a worthwhile trade-off though as it gracefully degrades.

PDF is significantly the worst. We use pdf.js, and it has to page by page to retrieve the text layer, so a large book takes a little while and the tab feels 'busy' while doing so. EPUBs are considerably faster, it's mostly a case of unzipping and parsing xhtml.

The main thing that was the difference maker was separating storage into two IndexedDB stores, one for metadata, and one for the full text. When we open the library list we only read from the metadata store, not pulling the entire content of all books you own into memory. We only pull in the one you actually want to read.

The biggest edge case by far are scanned PDFs that have no text layer (there is no OCR, so nothing is returned) that are currently not failing audibly enough to make it clear.

Tables are worse than dropped - PDF extraction de-flattens them into a stream of values in reading order, so you just get a torrent of disconnected numbers flitting past which don’t mean a great deal.

I would read figures in the normal way for this type of book and then use the tool for the prose. It’s not really fixable at the format level too, there is no good way of showing a person a graph word by word. I suppose what I really should add is some kind of flag to say there was a figure here instead of just letting the process go ahead and drop everything. Otherwise, you can’t really know that you’re missing things.

It started as a flat interval and felt terrible. Every word got the same slot, so sentences never landed and it read like a stock ticker.

Now there's a base duration from the WPM setting, with per word multipliers on top: 3x on sentence-ending punctuation, 2x on commas and semicolons, and an extra 0.7x if the word is 8 characters or longer.

The punctuation weights came from reading passages out loud and noticing where I actually stopped. Those are the breath points, so giving them proportionally more time puts the sentence structure back in. The long word bump is closer to the eye tracking literature, where fixation duration scales with word length. It felt wrong for a 12 letter word to get the same slot as "the".

The actual numbers were tuned by ear over a lot of real reading. I tried syllable estimates and word frequency weighting too, but neither was noticeably better than punctuation plus length, and they made the rhythm less predictable, which turned out to matter more than being clever.

In practice, this is probably not as good as using your own eyes and here's why, in addition to the argument above about skipping words. You no longer can skip words, which people do all the time, especially short function words. You already have to preview the next word or two in order to know when you've lost your place, so even if you had perfect peripheral vision you would be no better off than before. There is also the fact that you can no longer backtrack if you lose your place or simply get bored. In principle, of course, this could be solved by always allowing the reader to step back to a previous line, but this requires extra time and in general these features are a mixed blessing.

As to whether you would want to use such a device, and assuming you could use it with other media, I would say not particularly. Unless you are an especially fast reader, in which case there just isn't anything for you to gain. The people most likely to benefit from this sort of device are the people who normally have the most regressions or lose track of where they are. It makes the sort of prose that is meant to be read in one sitting much easier, faster, and less confusing. But it fails spectacularly on more dense material or on material where the visual presentation is important in any way. It also does not help anyone who simply can't visualize a text in the same way others do.

Some people just don't have the right kind of brain, and this isn't really their fault.

Your idea about sentences does affect the experience, but I don’t believe that makes it worse; it’s just a different approach. Once a whole sentence appears on screen, your eyes move across it again, so the no eye movement part is lost. What remains is the return sweep, the long jump back to the start of the next line, and the moment when you lose your place on a dense page. Those are real challenges in normal reading, and chunking by sentence removes both. You would end up closer to guided pacing than RSVP, which might suit a book better anyway.

Let me know how the procrastinated book turns out.