Some Perspectives on EPUB 3

By Bill Kasdorf, VP, Apex—There’s a lot of excitement about the new EPUB 3 format (an official recommendation of the IDPF, the International Digital Publishing Forum, as of October 11, 2011)—as there should be. It’s a major advance that was designed to bring the EPUB standard into alignment with the latest Web browser technologies and the reading systems that use them. It’s also designed to work for a much broader range of publications, including scientific and technical content, typographically rich and layout-intensive publications like textbooks and cookbooks and magazines, and to accommodate rich media (video, audio, animation) as well as interactivity and other scripted functionality.

But like most new things, it has also prompted its share of questions, confusion, misinterpretation, and general anxiety. I’d like to address some of the issues I hear about a lot.

The Old EPUB Is Still OK

The first thing to realize about the new EPUB 3.0 is that it basically builds on EPUB 2.0.1 (the previous and long-established standard) without making publications based on EPUB 2 obsolete. EPUB 3–compliant reading systems are required to be able to handle EPUB 2. That means all those EPUBs you spent time and money creating are probably just fine as is.

Yes, EPUB 3 would let you do lots of other cool things with those EPUBs. The metadata could be richer; you could control type and layout much better; you could add multimedia or scripting. But you don’t have to change them at all. Frankly, publications that really need those things were seldom created as EPUB 2 in the first place.

Sure, there were some magazines and textbooks and cookbooks done as EPUB 2, but they often fell short of what they really wanted to be. Many were created as apps instead of eBooks—at significant effort and cost. Those publishers are jumping on the EPUB 3 bandwagon big-time.

But the vast majority of books that have been successfully published as EPUB 2 (and there have been millions of them) will not need to be changed. Will their publishers want to do new books in a more sophisticated way with EPUB 3? Sure, if the new features are meaningful and beneficial. But for now, don’t panic. Your EPUB 2 eBooks have not just been made obsolete.

An EPUB Is a File But Also a Collection of Files

I often hear people ask, “Should I do EPUB 3 or HTML5?” It’s not an either/or—HTML5 is a fundamental component of EPUB 3. That means that the HTML content files you create for an EPUB 3 (okay, expressed in XML as XHTML) can be the same files you’d use online or in apps (see next section).

What EPUB 3 does is to provide a clear, complete specification for all the components of an EPUB—content files in XHTML or SVG, images, fonts, scripts, multimedia, and metadata—that organizes and documents them in a way that makes them predictable, reliable, and accessible for reading systems, packaged in a single file container: .epub.

But Are There Any EPUB 3 Readers Out Yet?

Reasonable answers to this question could be “a few,” “not quite,” “soon there will be lots,” and variations in between. What’s behind this question is the fear that this is something new that may or may not be widely adopted, thus providing an excuse to “wait and see,” which is a more palatable way of saying “can I please ignore this for a while longer?”

Here’s the real answer. EPUB 3 was deliberately based on the key existing, fundamental, and widely used open (no Flash!) technologies for publishing on the Web and to the latest generation of reading systems: HTML5, CSS3, and JavaScript, among others. What this means is that these technologies are already being used to publish content for modern browsers and to deliver content in readers like the iPad and Nook (among many others).

The files you’re currently using for cool online functionality, or for apps for the iPad or Android devices, are very likely already close to or exactly what you’d need for EPUB 3. Moreover, the EPUB 3s that you create will be an ideal (and standard!) input for standalone eBooks and their well-organized and well-documented components will be ideal for use online and in apps.

Not only is this good for publishers, it’s good for the technology companies too. Imagine how much it helps them to get files from most of their clients that are predictable, consistent, standard, well documented, and well organized. EPUB 3 gives them that. That’s why companies like Apple, Google, Adobe, and Barnes & Noble were all so actively involved in the development of the EPUB 3 standard, in collaboration with a broad range of publishers and technology companies.

What about that elephant in the room, the Amazon Kindle? No, the new generation of Kindles (led by Kindle Fire and using a spec called Kindle Format 8, or KF8) is not “EPUB 3– compliant.” Amazon wants to control its ecosystem; its software and formats are proprietary. But much or even most of what will ultimately be converted to KF8 is what you’d want in an EPUB 3 anyway. (Not everything—that’s why it’s proprietary.) Amazon even advertizes that one of the inputs for KF8 is—you guessed it—EPUB!

So back to the original question: Yes, we will soon see lots of official endorsements of EPUB 3 by the majority of reading system vendors and developers, but in most cases that will amount to confirming what they’re already doing to a great extent.

Not All EPUB 3 Reading Systems Will Be Created Equal 

There are two aspects to what is meant by “EPUB 3 compliance”: the files that publishers create and the reading systems built to render them. What people may not realize is that not all EPUB 3–compliant reading systems will handle everything in an EPUB 3–compliant file. Don’t panic—this is a good thing! If they had to do everything in the spec, we’d have few, if any, EPUB 3 reading systems. The reason is that the EPUB 3 spec accommodates a great deal of functionality that can be desirable in a given EPUB, but it will not make sense for all reading systems to be able to handle all of it.

Certain things are fundamental; for example, EPUB 3 requires that all compliant reading systems must be able to render Open Type or WOFF fonts. However, many aspects of EPUB 3 are, in a sense, optional for a reading system—with a key caveat: If an EPUB 3 reading system represents that it accommodates certain types of functionality (e.g., rendering math or handling scripting), then it must do so in compliance with the EPUB 3 spec. Thus, if a reading system says it can render math, it must be able to properly interpret MathML as specified by EPUB 3 (which, by the way, is the presentational aspect of MathML, not necessarily the semantic math). If a reading system enables scripted functionality, it must handle JavaScript as specified by EPUB 3.

So Is EPUB 3 Ready for Prime Time?

The short answer: yes! In fact, it has come not a moment too soon. The reason, as I mentioned earlier, is that publishers are already doing most of what EPUB 3 specifies (when their publications require or benefit from those things), but without EPUB 3 they are faced with inconsistencies in online environments, devices, and apps that require multiple versions of publications—and those devices and systems wind up going in inconsistent and proprietary directions. EPUB 3 provides a clear, comprehensive, practical specification that promises not just to enable cool things to be done in ebooks in a more consistent way but also to help publishers rationalize their workflows to save money and time and make their publications more adaptable to reading systems of the future.

Bill Kasdorf, a past president of SSP, is Vice President of Apex Content Solutions and General Editor of The Columbia Guide to Digital Publishing. He was Metadata Subgroup Lead for the EPUB 3 working group; he chairs the Content Structure Committee of the Book Industry Study Group (BISG); and he chairs the Mapping Subcommittee of the IDEAlliance nextPub working group (for magazines). Bill is also an active member of the NISO eBook Special Interest Group. Also see Bill’s Information Standards Quarterly article about EPUB 3.