1. What You Actually Read, Written Down

Open a document you actually read this week, not one you feel you ought to read, and try to make the text bigger.

One of two things happens. If the lines rewrap and the page count goes up, the file is reflowable. The text isn't stuck to a page, so bigger letters just mean the words rearrange themselves and nothing breaks. Most novels work like this. If instead the page grows past the edge of the window and you end up sliding around a corner of a paragraph, the file is fixed layout. A PDF of a journal article is the usual example: the text sits where it sits. Write down which one you got. That is the first line of a reading record, and every later decision in this course gets measured against that record.

There is a comparison table at ocdevel dot com slash ereaders. It scores devices out of a hundred and it has picks. By the end of the course you'll be able to read it, argue with it and choose from it. But a table with twenty-odd columns only helps once you know which two or three of them your own reading makes decisive. You don't know that yet, and neither do I, about you. Taken now, the table is a list of numbers with a winner at the top, and the winner could easily be wrong for you. So this lesson names no brand, screen size, price or pick. It builds the record. Without one, "is this screen good enough" has no answer. With one, the question becomes "is this screen good enough for a two-page chemistry diagram at arm's length," and that can be tested.

The record has four parts: what you read, where it comes from, when and where you read it, and what is hard about how you read now.

Start with what you read, because this is where shopping most often goes wrong. Sort your reading into kinds, and the kinds aren't genres. They're shapes.

Novels are reflowable: bigger text just rewraps the lines, the first of the two outcomes above. Fixed-layout documents are the opposite: make the page bigger and it doesn't reflow, it just gets bigger than the screen.

Comics are fixed layout too, with an extra demand. The picture is the content, so shrinking it to fit isn't a compromise you can accept the way you can with a paragraph.

Textbooks are usually fixed layout plus navigation. You don't read a textbook front to back. You jump to page 214, then to the index, then back.

Handwritten notes aren't reading at all. They're writing, and they only count as a need if the notes have to end up somewhere else afterwards.

Audio is a separate thing again. An audiobook needs a speaker or a headphone connection and an app that can play it, and none of that has anything to do with how the screen looks.

Write down which of those you actually do, and roughly how much of your reading time each takes. If ninety percent of your reading is novels, that is very useful to know. It means most of the expensive columns in that table don't apply to you.

Second, where your reading comes from. Write down the store you buy from, if you buy. Write down whether you borrow from a public library, and which one. Write down roughly how many books you've already paid for, and in which account, because those books are the reason switching brands isn't free. Write down whether you have personal files: documents on a computer, papers you downloaded, things a colleague emailed you. And write down any app that is required. Not nice to have. Required, as in: if this device can't run it, I've lost access to something I need.

Third, when and where you read. Write down the light. Do you read in bed with the lamp off? On a sunny patio? On a train at dusk? Write down your hands. One-handed, lying down, thumb on the edge of the device? Or two-handed at a desk? Write down indoors or outdoors. And write down the distance from your eyes to the page. Hold a book the way you really hold it and look: is it thirty centimetres away, resting on your chest at fifteen, or propped on a desk at fifty? That distance decides how big the letters need to be, which decides how much page is left over for words, which decides whether a small screen is pleasant or maddening.

Fourth, what is hard right now. The rule for this part: write the complaint, not the wish. "My phone screen is too bright at night and I read until two in the morning" is a complaint. "I want a device with a warm front light" is a wish, and it has smuggled in an answer before you asked the question. "I have four hundred PDFs on my laptop and I never read them because reading them on the laptop is miserable" is a complaint. "I need a ten-inch tablet" is a wish. Keep the wishes out. You'll reach better solutions later if you didn't decide in advance.

Now build one from real files. Gather three or four documents you read: actual ones, pulled from your downloads folder, your email attachments, your existing reading app. If you have a store account or a library card, have it open. Those files and that account are all the starting information you need. Everything else in the record comes out of them or out of watching yourself read.

Here is how the checks come out on a made-up example: a twelve-page scanned PDF of a knitting pattern, downloaded from a pattern shop. The results belong to that file and only that file.

First check: can the text get bigger without the page turning to soup? Open the file in whatever you normally read it in and increase the font size or the zoom. The knitting pattern has no font control at all, only zoom, and at a hundred and fifty percent the right-hand column runs off the edge of the window. So: fixed layout. That tells you something specific. For this document, screen size and the device's ability to crop or split a page matter more than anything about fonts.

Second check: is there a real text layer, or is it a photograph of a page? Try to select a word with your cursor, then search for a word you can plainly see. On the pattern, the cursor draws a box instead of highlighting a word, and searching for "cast" finds nothing, even though "cast" is right there in line one. There is no text layer. It's an image of a page, and that means no search, no text reflow and no text-to-speech, however clever the device is. The limitation is in the file. This is a different problem from fixed layout, and the check tells them apart. A fixed-layout page that does have a text layer can sometimes be reflowed by software, badly, by stripping the layout out. A page with no text layer gives the software nothing to strip out, so nothing can be reflowed. If you find selectable, searchable text, write "has a text layer," and reflow and search become things to test later.

Third check: how big is the page, and what spans it? The pattern is A4, about twenty-one centimetres across, and its stitch chart runs the full width: a grid about eighteen centimetres wide with symbols in the squares. That is the measurement that matters, not the page size in the abstract. An eighteen-centimetre grid on a screen narrower than eighteen centimetres means shrinking it or scrolling across it, and shrinking a grid of small symbols until it fits is exactly how you lose your place in row forty-two.

Fourth check: how big is the whole library, and where does it live? Look at the folder and read the size, then look at your store account and count. Say the pattern folder holds about sixty PDFs, roughly two hundred megabytes, all on a laptop and none in any account. Compare that with someone whose whole library is four hundred novels inside one store account and nothing on a computer. Those are opposite situations. The first needs a way to get files from a laptop onto a device. The second needs a device that can log into that account. Neither is harder in general. They're different questions.

Four checks produce a short record. What gets read: mostly fixed-layout PDFs with full-width diagrams. Where it comes from: sixty local files, no store account. And a list of things to test on any device rather than assume: whether it can show an eighteen-centimetre-wide chart legibly, whether it can accept files from a laptop, and whether it can crop or split a page, since search and reflow are off the table for these files.

That list isn't features anyone wants. It's tests. Every line in a record turns into either a test or a filter.

Some of it stays uncertain. The checks show the file's limits, not whether any device handles the file well. And they were run on one file out of sixty. If the other fifty-nine have text layers, the record changes. So the line reads "at least some of these are scans; spot-check three more."

Three sample readers will carry through the rest of the course, described by what they read rather than by name. They're made up, none of them is meant to be you, and their job is to make the differences between records visible. Your record may look like one of theirs, or borrow from two.

First, a reader who mostly reads novels. What gets read: novels, almost only novels, about eighty a year. Where they come from: a public library, borrowed through a lending app, plus a handful bought. Almost nothing on a computer. When and where: in bed, lights off, a partner asleep alongside, the book held in one hand because the other arm is under the pillow. Also on a beach in summer. What is hard: the phone is too bright and too tempting, and holding a hardback one-handed in the dark doesn't work.

Second, a reader who mostly reads technical PDFs. What gets read: engineering papers and two textbooks. Where they come from: a university library portal and a laptop folder with about eight hundred PDFs in it. No consumer store account at all. When and where: at a desk, in daylight, two hands free, screen about forty-five centimetres away. What is hard: half the documents have figures that span the full page width, and the reading means constant jumps from page ninety back to a table on page twelve and then to the reference list at the end. A laptop can do that; it just eats the afternoon and the eyes.

Third, a reader who mostly writes. What gets read: a little. What gets written: a lot, meaning meeting notes, diagrams and marginal notes on contracts. Where it comes from: documents that arrive by email, and blank pages. When and where: in meetings, on a knee, in indoor office light. What is hard: the notes go on paper and then they're trapped there. When somebody asks for them, they get photographed, badly. The need is handwriting that arrives on a laptop in a form that can be sent.

Side by side, those aren't three points on one scale from cheap to expensive. They're three different machines. A finished record can also lead somewhere other than an e-reader. The note-writing reader's record barely mentions books; it describes a digital notebook. The technical reader's describes a document reader, and a big one. Some records describe an accessible reading setup, with very large text or spoken text, where the right answer might be a phone with the right settings and no new purchase at all. A general-purpose tablet isn't the default answer, and neither is the most capable device on the table. That device is heavier, more expensive, and worse at the one thing the novel reader does every night.

Each record turns into specific open questions.

For the note-writing reader: in what form do the handwritten notes come out, and where do they go? Can they leave the device as something a laptop can open? Does that need an account, a subscription or a companion app? Is the export a picture of the handwriting, or something searchable?

For the technical reader: how does a device show a page whose layout can't move? Can it crop the margins, split a page into halves, or reflow the text where there is a text layer to reflow? How do you jump to page twelve and back without paging through eighty screens? And how do eight hundred files get from the laptop onto the thing?

For the novel reader: can the device borrow from the library directly, and does that depend on the country and the particular library? What does its light do at the dark end, and can it be made warm as well as dim? Can you turn a page with the hand that's holding it?

One thing to do today: go to ocdevel dot com slash ereaders and look at the filters. Don't choose anything, and don't look at the scores. Read the names of the criteria and notice which ones you now have an opinion about because of your record, and which mean nothing to you yet. If your record says "one hand, in the dark," you'll find filters for exactly that. If it says nothing about writing, you've just set aside a whole family of expensive devices without doing any research.

Recent changes worth a line in your record

A few things changed in the last couple of months that affect what an owner can read or do, and all of them are shipping now rather than announced.

The first is about library loans on Kindle. Amazon's firmware versions five point nineteen point five and five point nineteen point six close a loophole borrowers had been using: keeping a device offline so an expired library loan stayed readable. Opening an expired title offline now forces a reconnection prompt, and the book stays locked until the loan status is verified. This is shipping over the air and by manual sideload, and it affects modern devices, the eleventh and twelfth generation Paperwhite, the Colorsoft and the Scribe, running firmware in the five point nineteen point two through five point nineteen point six range, mainly library borrowers on United States accounts. There is also an A and B rollout replacing the Collections feature with Tags, so two owners of the same model may see different names for the same thing. If you own one of these, check your firmware under Settings, then Device Options, then Device Info. If you don't own one yet, the action is a line in your record: if you borrow from a library, note that a loan is time-limited access and not a file you hold, and that this route is one to test rather than assume.

The second is on Kobo. Firmware four point four six fixes on-device store checkout and payment verification, which had been failing because of updated security certificate requirements, and it adds backend support for syncing your reading with the tracking service StoryGraph. It is rolling out over the air and by manual download for current models including the Libra Colour, the Clara Colour and the Clara BW, across all Rakuten Kobo regions. Owners can confirm the version under More, then Settings, then Device Information. For your record, buying directly on the device is a route that can break and be repaired, so write it down as something to verify rather than take for granted.

The third is smaller, but it touches document access. The Boox Palma 2 Pro shipped in August of twenty twenty-six with a hybrid dual-SIM tray, which means cellular data for downloading and syncing reading material without Wi-Fi. These are shipping retail units running Android with Google Play. If you read somewhere without reliable Wi-Fi, such as a commute, a cabin or a hospital waiting room, add a line saying so. Most records never mention it, and it rules devices in and out.

If you haven't yet, open one real document, make the text bigger, and write down what happened. That's line one.