EPUB best practices guide

[Bob Russell]

The new IDPF EPUB e-book format standard is taking the world by storm. Think of it as the analogy to the Adobe .pdf format, but for flowable text, so it can display well on any size screen from a PC to a phone. Plus, a standard for the container format, so multiple document files can be distributed together in a zip archive.

Okay, so maybe "taking the world by storm" is a bit of an exaggeration at this early stage, but it's certainly gaining as much momentum as could possibly be hoped for, much of which is a result of the enthusiastic and substantive support from Adobe. But, it's clear that one critical success factor for EPUB is the ability of e-book creators to create a well-formatted EPUB e-book, which is compatible with a wide range of software and devices. Coincidently, that's a great interest of many MobileRead members as well.

So I'm very happy to report that Adobe is furthering that goal, by providing the following:

(1) An EPUB Best Practices Guide white paper, created by the Digital Publishing Team at Adobe (attached in EPUB format). There is also some background information about the format that may make it of interest to non-technical readers as well.

"By following the best practices in this Guide, you will be able to produce EPUB files, compatible with the IDPF EPUB specifications, that will look great for consumers, regardless of whether they are reading your content on desktop PCs, mobile phones, or other reading devices."

If you haven't already, you will probably want to download Adobe's Digital Editions software to read the document.

(2) A web site (http://www.adobe.com/devnet/digitalpublishing) with "the latest information from Adobe on Digital Publishing".

Nick Bogaty of Adobe, who kindly tipped us on this information, says that the site will launch Wed, March 26, 2008 (looks like it's already live to me), and that it will include "access to information on Adobe Digital Editions, Adobe Digital Experience Protection Technology, Adobe InDesign CS3 and other tools and scripts that Adobe provides to help create, distribute and consume both PDF and EPUB digital publications."

It will be interesting to hear feedback from some of the practitioners here after they dissect the information with the benefit of their experience and expertise.

[NatCh]

Very exciting news! I've been hopeful that epub or DE would finally give us a de-facto "standard" format, since the people making books and devices can't seem to pull their fingers out of their respective noses far enough to do it. Each report of Adobe's support for epub greatly increases the odds of a "standard" emerging, in my estimation. Something to watch, definitely.

[jbenny]

I just took a preliminary look at the epub whitepaper and it looks good. It seems to explain things very clearly.

I am very glad when anyone supports and encourages the use of epub. However, I wish that Adobe wasn't so determined to use Flash for their reader software. I don't know why they are so married to using Flash for Digital Editions, but it causes problems, IMO. The most glaring problems are speed and poor integration into the underlying operating system.

The other issue I have with DE is not related to Flash, but with the way that DE is deployed. Requiring an internet connection to individually install and use the application is a very bad thing. How do you install DE on a machine that doesn't have an internet connection? How do you deploy DE on dozens or hundreds of PCs?

Enough ranting about DE. I am happy that Adobe is a prime supporter of IDPF and epub. How about Microsoft, or some other big gorillas?

[wallcraft]

It is interesting to see the impact of CSS by displaying this in Digital Editions and in FBReader (which has no CSS support). A minor upgrade to FBReader, adding support for the NCX navigation (table of contents) file, would help a lot. The cover page is a SVG graphic with the text of the author, title, etcetera overlaid. This is Adobe's recommended approach (end of the content section). FBReader also does not render SVG graphics, which is probably just as well since it definitely would not overlay the text.

FBReader isn't doing a very good job here, beyond displaying the text (it does better on OEB 1.2 conforming documents), but I was surprised that DE does not allow navigation within a chapter using its TOC side-panel.

One nice thing about unencrypted EPUBs is that that are just renamed .zip files. Unzipping the contents allows you to see how it is organized. This document is the very first EPUB I have seen that does not contain a human readable OPF file (no line breaks in any of the text-based files, including the .opf).

It the text in an image looks fuzzy, your browser is rescaling it - click again to get the full size.

[Hadrien]

Originally Posted by wallcraft

but I was surprised that DE does not allow navigation within a chapter using its TOC side-panel.

What do you mean exactly ? Displaying a thumbnail of the page to navigate within the chapter ?

[wallcraft]

Originally Posted by Hadrien

What do you mean exactly ?

The screenshots show the navigation pane with the chapter titles. If you click on the "+" by a chapter, the chapter subsections are listed. You can click on the chapter to go to that chapter, but clicking on the chapter subsections does nothing.

[Hadrien]

Originally Posted by wallcraft

The screenshots show the navigation pane with the chapter titles. If you click on the "+" by a chapter, the chapter subsections are listed. You can click on the chapter to go to that chapter, but clicking on the chapter subsections does nothing.

That's a problem with their file, not DE.

Try one of our books: http://www.feedbooks.com/discover/epub/83

Edit: Hmm strange problem. Works fine on my laptop with DE 1.5 beta 1, and their TOC doesn't work with DE 1.5 beta 2 on my desktop.

[Paul Norton]

Originally Posted by Hadrien

That's a problem with their file, not DE.

Try one of our books: http://www.feedbooks.com/discover/epub/83

Edit: Hmm strange problem. Works fine on my laptop with DE 1.5 beta 1, and their TOC doesn't work with DE 1.5 beta 2 on my desktop.

I think you'll find that it actually depends on which entries in the ToC you choose.

It seems I've included an old toc.ncx in the document.

An updated document will be posted soon. Sorry for the confusion.

Paul

[jbenny]

Hadrien, I haven't tried the betas of DE 1.5 yet. Did they add a "back" button?

[jbenny]

Originally Posted by wallcraft

It is interesting to see the impact of CSS by displaying this in Digital Editions and in FBReader (which has no CSS support).

Yes, CSS support is the one major feature that is lacking in FBReader. Most run-of-the-mill fiction ebooks display ok without CSS, but I've had quite a few ebooks (fiction and non-fiction) that were unusable in FBReader because of no CSS.

I love the program and I realize that it is only one guy working on it, but I wish that some other programmer would help out and work on adding CSS. It would make FBReader a killer app.

I think CSS is especially important, considering that FBReader and DE are currently the only two programs capable of reading epub. Given the drawbacks of DE that I pointed out above, FBReader would be a better choice for a lot of users if it had CSS.

[kovidgoyal]

I was looking into writing a viewer for epub, lit, mobi and other html based formats using the webkit engine (full spoort for css and even javascript). It looks like it would be pretty easy, the only major sticking point would be you'd have to give up pages and just get used to reading books like you do websites, in continuous flow mode.

[wallcraft]

If Apple wrote an e-book reader it might turn out to be a thin layer on top of safari (and work surprisingly well).

Continuous flow would be ok, if page up/down could provide a bit of an overlap. The other thing you would need is a way to remember your high water mark and last location in each e-book.

[kovidgoyal]

Bookmarking can be handled by just saving a few hundred characters from the top of the currently displayed text.

[jbenny]

Originally Posted by kovidgoyal

I was looking into writing a viewer for epub, lit, mobi and other html based formats using the webkit engine (full spoort for css and even javascript). It looks like it would be pretty easy, the only major sticking point would be you'd have to give up pages and just get used to reading books like you do websites, in continuous flow mode.

Rather than starting from scratch, why don't you add your talent to Openberg Lector? It already does epub. It does need some help with paging, TOC viewing, etc. Although having a standalone app, instead of a FF plugin has some advantages. For that matter, how about helping improve FBReader?

As for javascript in an ebook reader I question whether it is really needed. At the very least, I would make that something that the user could disable. Not having it at all would certainly make the reader app smaller.

Paging would be desirable. However, FBReader doesn't do paging. The closest it comes to it is starting a chapter heading at the top of the screen (with some formats).

Edit. More on paging: I don't know what rendering engine they use, but MS Reader figures out the paging of an ebook when it is first opened. This of course depends on screen size, font, font size, etc. You can start reading a book while the paging is being done in the background. As long as you don't try to jump to the end of a very long book before the process is finished, this method works very well.

[jbenny]

Originally Posted by kovidgoyal

Bookmarking can be handled by just saving a few hundred characters from the top of the currently displayed text.

Wouldn't it be simpler to just use a pointer into the ebook file? Isn't that how Mobipocket Reader, FBReader and MS Reader do it?