If you were designing EPUB today, what would you change?

Kayadelenium · Yesterday, 02:04 PM

In the epublift thread, KevinH asked why WebP and why not just zip -9, and JSWolf argued that everything should be left as it is. They made me want to ask a bigger question, and not about any tool:

If you were designing the ebook format today, in 2026, what would you keep, what would you change, and what would you leave out?

Technically and philosophically both. Publishers, people who make books by hand in Sigil and calibre, people who mostly read, and people who write reading software all see different parts of this, and I would like to hear all of them. I will keep a summary of the ideas raised at the end of this post and update it as the thread goes on, so a newcomer can catch up without reading every page.

Two things to start from. Mostly, though, I would like to hear what you do today and why.

1. Compression

An EPUB is a ZIP, and the compression inside it is deflate, applied to each file on its own. Deflate is about thirty years old and works with a 32 KB window, so it cannot use what the chapters of a book share with each other. Newer codecs such as Zstandard compress better and decompress faster, but where the gain really comes from, a newer codec or compressing the book as one stream, is something I want to measure properly before I put numbers here. I opened a discussion with the EPUB working group about Zstandard in the container: epub-specs discussion #3025.

2. Images: the spec and the devices disagree

EPUB 3.3 lists WebP as a core media type, and the EPUB 3.4 draft adds AVIF and JPEG XL. On the devices I could check myself, stock Kobo (Forma, Sage) shows WebP as blank pages, while Apple Books renders it. Kindle, PocketBook and the Adobe-engine readers do not list WebP in their own documentation. So today a book that follows the newest spec can show blank pages on the readers many of us own.

Questions to start with

What do you make, convert or read today, and for which devices? Who should a book be written for in 2026: every reader still in use, EPUB 2 ones included, or the current spec?
Images: follow the spec (WebP now, AVIF and JPEG XL next), or follow the devices?
Fonts: embedded TTF/OTF, WOFF2, or no embedded fonts at all?
Accessibility: the European Accessibility Act has applied since June 2025. Does that change what a good EPUB is?
Compression and container: is ZIP with deflate good enough, or is it time for something else?
XHTML or HTML, NCX or not, scripting or not: what is the format for, and what is it not for?
What would you drop from EPUB 3 entirely?
Why are there so many ebook formats, and why did new ones keep appearing after EPUB? I asked this in the epublift thread with a list of formats: post #13. Answers are welcome in either thread.

My own starting point, to disagree with

The first thing I would fix is the name. EPUB 2 and EPUB 3 come from two very different worlds, yet they share one name and one extension, so nobody can tell which one a file is without opening it up. The one thing we can fairly ask of someone who just wants to read is to know whether their device reads EPUB 2 or EPUB 3. Everything past that, image formats, font formats, 3.0 versus 3.3, is a burden for editors, publishers, device makers and people who write software. We should carry it; we cannot hand it to the reader. Today we cannot even give the reader that one fact.

Nothing that exists should break. Every book made so far should keep working, and the old way should stay as the baseline. But new technology arrives all the time, and modern devices deserve a more efficient format that can take it in over time.

I also think an ebook should age like a book, not like a web application: boring and durable, strict enough that someone can still write a reader for it in fifty years. A book should still read in full on a basic e-ink reader. Interactive things that cannot do that belong in a different format with a different name. Notes and highlights belong outside the book, not written into it. And if EPUB grows profiles, I would rather see restrictive ones, an archival subset in the spirit of PDF/A, than supersets that fragment what a reader has to support.

That is one view. I would like to hear where it is wrong.

Summary of ideas so far

(Nothing yet. I will add each idea here as it comes up, with a link to the post that raised it.)

BetterRed · Today, 05:13 PM

I'd start by trying to implement a stitched together codex rather than a papyrus scroll - IOW I would not start with HTML.

JSWolf · Today, 07:28 PM

I would dump all of the mutli-media and I would look at what else everything Apple managed to get added in and see if it needs to go or be changed.

I would also make ePub3 more simple. I would bring back things changed from ePub2 that did not need to be changed because the changes do the same thing. I would make ePub3 accessible by default without the need for the code. I would simplify it. It's a bit (IMHO) more complicated then it need to be.

Today, 05:13 PM	#2
BetterRed null operator (he/him) Posts: 22,858 Karma: 34252730 Join Date: Mar 2012 Location: Sydney Australia Device: none	I'd start by trying to implement a stitched together codex rather than a papyrus scroll - IOW I would not start with HTML. Last edited by BetterRed; Today at 05:24 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Change date to today when moving book between library	reup	Library Management	4	11-22-2025 05:31 AM
Designing for Kobo	AlexBell	ePub	45	01-21-2013 01:09 AM
Designing a document for both print and epub	adieu	ePub	6	10-19-2011 09:10 PM
Today is the Day the Internet may Change Forever	kennyc	Lounge	6	04-08-2011 04:47 PM

Yesterday, 02:04 PM	#1
Kayadelenium Addict Posts: 222 Karma: 363834 Join Date: Jul 2026 Location: Planet Earth Device: Kobo Forma	If you were designing EPUB today, what would you change? In the epublift thread, KevinH asked why WebP and why not just zip -9, and JSWolf argued that everything should be left as it is. They made me want to ask a bigger question, and not about any tool: If you were designing the ebook format today, in 2026, what would you keep, what would you change, and what would you leave out? Technically and philosophically both. Publishers, people who make books by hand in Sigil and calibre, people who mostly read, and people who write reading software all see different parts of this, and I would like to hear all of them. I will keep a summary of the ideas raised at the end of this post and update it as the thread goes on, so a newcomer can catch up without reading every page. Two things to start from. Mostly, though, I would like to hear what you do today and why. 1. Compression An EPUB is a ZIP, and the compression inside it is deflate, applied to each file on its own. Deflate is about thirty years old and works with a 32 KB window, so it cannot use what the chapters of a book share with each other. Newer codecs such as Zstandard compress better and decompress faster, but where the gain really comes from, a newer codec or compressing the book as one stream, is something I want to measure properly before I put numbers here. I opened a discussion with the EPUB working group about Zstandard in the container: epub-specs discussion #3025. 2. Images: the spec and the devices disagree EPUB 3.3 lists WebP as a core media type, and the EPUB 3.4 draft adds AVIF and JPEG XL. On the devices I could check myself, stock Kobo (Forma, Sage) shows WebP as blank pages, while Apple Books renders it. Kindle, PocketBook and the Adobe-engine readers do not list WebP in their own documentation. So today a book that follows the newest spec can show blank pages on the readers many of us own. Questions to start with What do you make, convert or read today, and for which devices? Who should a book be written for in 2026: every reader still in use, EPUB 2 ones included, or the current spec? Images: follow the spec (WebP now, AVIF and JPEG XL next), or follow the devices? Fonts: embedded TTF/OTF, WOFF2, or no embedded fonts at all? Accessibility: the European Accessibility Act has applied since June 2025. Does that change what a good EPUB is? Compression and container: is ZIP with deflate good enough, or is it time for something else? XHTML or HTML, NCX or not, scripting or not: what is the format for, and what is it not for? What would you drop from EPUB 3 entirely? Why are there so many ebook formats, and why did new ones keep appearing after EPUB? I asked this in the epublift thread with a list of formats: post #13. Answers are welcome in either thread. My own starting point, to disagree with The first thing I would fix is the name. EPUB 2 and EPUB 3 come from two very different worlds, yet they share one name and one extension, so nobody can tell which one a file is without opening it up. The one thing we can fairly ask of someone who just wants to read is to know whether their device reads EPUB 2 or EPUB 3. Everything past that, image formats, font formats, 3.0 versus 3.3, is a burden for editors, publishers, device makers and people who write software. We should carry it; we cannot hand it to the reader. Today we cannot even give the reader that one fact. Nothing that exists should break. Every book made so far should keep working, and the old way should stay as the baseline. But new technology arrives all the time, and modern devices deserve a more efficient format that can take it in over time. I also think an ebook should age like a book, not like a web application: boring and durable, strict enough that someone can still write a reader for it in fifty years. A book should still read in full on a basic e-ink reader. Interactive things that cannot do that belong in a different format with a different name. Notes and highlights belong outside the book, not written into it. And if EPUB grows profiles, I would rather see restrictive ones, an archival subset in the spirit of PDF/A, than supersets that fragment what a reader has to support. That is one view. I would like to hear where it is wrong. Summary of ideas so far (Nothing yet. I will add each idea here as it comes up, with a link to the post that raised it.) Last edited by Kayadelenium; Today at 07:38 AM.

Today, 07:28 PM	#3
JSWolf Resident Curmudgeon Posts: 85,191 Karma: 153791427 Join Date: Nov 2006 Location: Roslindale, Massachusetts Device: Kobo Libra 2, Kobo Aura H2O, PRS-650, PRS-T1, nook STR, PW3	I would dump all of the mutli-media and I would look at what else everything Apple managed to get added in and see if it needs to go or be changed. I would also make ePub3 more simple. I would bring back things changed from ePub2 that did not need to be changed because the changes do the same thing. I would make ePub3 accessible by default without the need for the code. I would simplify it. It's a bit (IMHO) more complicated then it need to be.

Advert