Hacker News new | past | comments | ask | show | jobs | submit login
ePub 3.2 (w3.org)
89 points by rahuldottech 36 days ago | hide | past | web | favorite | 18 comments

I feel like there's a slight lack of grassroots activity with epubs, people tend to work with mature products or random e-readers. I've recently experimented a bit with making epubs by hand (bash+perl) and they tend to break readers more often than not.

So I feel we've come to the spot where there's standards, but then there's convention over (a) decade(s) that just trumps standards. Reminds me of DHCP where I've yet to come across three implementations that agree on certain aspects. Also, is 'supports epub x.y' been a selling point in any reader so far?

I'm looking through the changes and holy hell, this makes me feel like punching an engineer, and I am one! Everything is buried under three or more layers, requiring at least 3 clicks and a bunch of scrolling and usually just refers to what they've changed compared to the PREVIOUS version. While fine for a standard, this is absolutely awful for anyone who doesn't work with epub files daily (which I'm guessing is a very small minority).

If you're interested in changes written for humans, this link is what you're looking for: https://www.w3.org/publishing/epub32/epub-overview.html It doesn't contain motivation for these changes, though, so we'll go with "some engineer thought these would be a good idea"

I guess I'd like to ask the working group:

#1 How widespread would you consider high/recent version support for epubs to be?

#2 How would you sell epub 3.2 support to people developing e-readers or reader applications?

#3 Why isn't the first page filled with "Hey guv, you're an engineer who's interested, here's what epub is and does, and here's a link with all the recent changes along with update motivation information" because looking at this site, I could not find a single thing for convincing manager/C-level types to grant 200 product update hours for this.

> I've recently experimented a bit with making epubs by hand (bash+perl) and they tend to break readers more often than not.

Me too, and I wish SVG support was better on most eReader (not a good experience so far), and that there was official support for MathML or something equivalent. It hurts to see so many books using raster pictures to display a formula or a diagram, when it would look crisper and flow better if it was vector-based.

I'm in academic book publishing, and our books have a lot of charts and diagrams that are challenging to display on certain readers that have poor SVG support. Also, tables can easily be larger than the readers can comfortably handle, and certain readers (yeah, looking at you, Kindle) suck at it.

You'd think there would not be such a great difficulty in this, since epubs are constructed out of the same basic tech stack as web pages.

I think the fundamental problem is that Amazon, Apple, etc., don't have much incentive to support texts beyond the most commercial genres.

I feel like EPUB is a standard who's features are underutilized. Is there a nice high-level overview of EPUBs capabilities?

I found a German text from 2018 that covers EPUB 3.0-3.2 (which was obviously still a draft back then): https://www.dpc-consulting.org/epub-3-3-0-1-3-1-3-2-das-vers...

It links to a long blog post that covers recent EPUB development by Dave Cramer, co-chair of the W3C Community Group: http://epubsecrets.com/why-specs-change-epub-3-2-and-the-evo...

EPUBCheck is an open source tool that can check your EPUB files against the EPUB 3.2 spec. https://github.com/w3c/epubcheck

The features are underutilised because most readers simply don't support them.

There's no point using a feature if it'll just break the book for the vast majority of your audience, and it doesn't look like that situation is going to change anytime soon.

It's such a nice standard, I just wish it had better support on E-readers (looking at you Kindle).

Well, it’s only Kindle that insists on a proprietary format. Everyone else does ePub 2 fine (ePub 3 support is spotty).

Kindles can read .mobi files, which is a pretty open format. Calibre can easily convert EPUBs to .mobi (or to the similar .azw3 format) for reading on Kindle. While Amazon has introduced recently a new format with advanced typographic features that is highly proprietary (.kfx), uptake of it among publishers has been slow.

Mobipocket has been dead for the last couple of years and I can’t find a spec anyway. The Library of Congress describe it as a “proprietary, partially documented, binary format for ebooks”[1] which doesn’t really meet “open” by my definition. As for KFX, that’s 100% proprietary, yes.

Amazon in the US at least has sewn up the ebook market such that they have no interest at least in implementing open standards. The best thing to do in these situations is to vote with your wallet.

[1] https://www.loc.gov/preservation/digital/formats/fdd/fdd0004...

I love my Kobo Clara HD for its native epub support. I prefer the hardware and software over my Kindle Paperwhite too, tho I use the latter when epubs are’t available. I tried the whole Calibre conversion thing for a while, but just find it better/easier to have two ereaders.

I wish MacOS had a good ePub reader that did not insist on multiple column view and runs as a standalone app.

Kind of sad XULRunner got killed or I could run EpubReader as a stand alone app.

It looks like there was a really nice one around seven years ago, called Bookle. All of the reviews from the time seemed to love it. Unfortunately, it's no longer available for sale.

I searched high and low for a copy, and even emailed the developer. Their response made me quite sad: "I'm afraid that I don't know that I even have a build at this point."

Who knows if it would even work on modern macOS, but I bet it would, as long as it was 64 bit. macOS's backwards compatibility is better than some seem to imagine.

Apple Books? Switching it to single page view is pretty trivial.

True. I wish there was a setting to make single column permanent, but having to press Cmd-1 when opening a book is not too much of a hassle.

iBooks app? You have to fiddle with the size of the window and once you get one column, it ends up with ridiculously huge margins on it.

It was renamed to Books in Mojave. Either way, you can just select View / Single page from the menu and it resizes to a single column view with normal margins. That isn’t supported in full screen view tbf.

Yeah I prefer a single column while keeping the window at the full width of the display. It tends to shrink the window to fit it's preconceived notions of "vertical rhythm" as they call it in the text world I guess - or usually trying to keep the width to 70-80 characters. I don't like this though

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact