Spine order in ePub

Overview

Don't fall into extensive reordering the e-book spine trap. It has no real benefits, raises the cost of production, and potentially reduces and devalues "The Book".

• Remove the minimum number of pages you can. That probably means Half-title page and any blank pages.
• All sections don't have to be in the NCX (ePub2) or TOC (ePub3), and they don't have to appear in the nav sections in the order they are in the spine.
• e-Up the various front and back matter section concepts rather than destroy them as irrelevant for e-books. For examples publishers should be looking at brand-building in ePubs of value.

Fiction and simple narrative non-fiction are not really affected by the discussion in this article. But nearly every other genre is. ePub 3 makes additional navigation structures available so you can include List-of-Figures, List-of-Equations, Page Navigation and other structures that make it easier for the user to move around the book.

This article is targeted at smaller publishers who may be producing e-books for the first time, or considering change requirements for ePub3, and need a little background on why some things are done; should they be done; what are the options; and what is "e-book law". 

There is a lot of consultant noise on the spine reordering - remove the title page, move the copyright to the back, delete the index, open the book at the first reading page... and sometimes a lot more.

I usually ask why do you want to do this shuffle-around? There are various justifications, but when you consider the ease of navigation through an ePub in most readers, it is usually change for the sake of it, rather than for a compelling reason.

Arbitrary changes usually diminish the value of the e-book rather than enhance it. For some reason emerging digital book experts feel they have to change something. If there are changes it should be about more bang for the buck on supporting pages such as About the Author, Previews and other value additions. There is no cast-off and no printing bill. IGP:Digital Publisher makes it easy to have different pages for various print editions and e-books.

The Spine and the NCX

There are four data structures in a ePub to control how things are laid out and define some part of user interaction and navigation.

1. The Manifest. A list of all the files included in the package. It's the source of ID's to everything else. Every file in the ePub must be listed in the manifest.

2. The Spine. The spine defines the linear reading order. When you use forward navigation - next page, the system moves through the document using the spine order. Each valid OPS content document is only allowed to appear in the spine once. Spine references can also be auxilary with a linear-"no" attribute. The reading device skips over these. This is very useful for documents with more interaction. To get to those pages you have to provide internal links.

3a. The NCX. (Navigation Center eXtended) ePub2 structure. Provides navigation to various sections. It allows you to jump to a location. Typically this is a start of a section. Device limitations restrict the use of the NCX for more exciting things other than a replica of the Table of Contents.

3b. TOC.xhtml. ePub3 navigation file structure. Provides navigation to various sections and also allows the inclusion of other content navigation structures such as List-of-Plates, List-of-Maps and page navigation.

4a. The Guide. ePub2 structure. This identifies the structural components of a book with a controlled grammar.  It is optional in the spec, but Amazon and Apple have made it more or less mandatory for their formats.

4b. Landmarks. ePub3 structure replacing "The guide". Landmarks allows the creation of special navigation targets to important locations in your book.

The spine defines the next/previous page navigation, the NCX/TOC.html provides section and other item navigation. The first entry in the spine defines where the ePub will open (if the reading system is not making other decisions).

Where should a book open the first time?

It is entirely your choice as a publisher. There is no best-practice. Only opinions.

Their opinion. There are experts who say the book should open at the first reading page of the book, so you can just get reading. I can give a number of reasons why that one is plain stupid. I would also say they are mostly Gutenberg book readers.  This stuff not based on user research.It's just a load of consultant opinion noise.

My Opinion. Keep the title page and open the book on the cover or title page for the first opening. Don't open at the first text page unless you decide, for a particular book, it is the best user experience. Think only of the end-user payoff you are delivering on that first interaction. Many reading systems may provide a "remember last reading position" when opening books in subsequent reading, so this can only be reliably enforced the first time. It is potentially "one-shot".

So think about your book and the experience your reading audience will have the first time they open the e-book. Children's books, novels, non-fiction, self-help, high-touch, education and academic can and probably should all have different approaches.

Why eliminate the Title page? Why not e-Up the title page concept and increase its value to both the reader and publisher. 

Some simple section advice

Front Matter

Cover

Your cover can be the first opening page defined by the spine, can appear in the NSX or TOC but does not have to. It's your choice. If your cover is a talking point, get it in the NCX/TOC. If you have the e-book rights to the print book cover artwork and it is a brand item, exploit it.  Remind the reader what a great purchasing decision they made.

Half-title

You can remove it. You can also leave it in the spine order and not have it in the NCX. Half-title on the TOC.xhtml or NXC probably looks a little naff, and the book title repeated twice is overdoing it. So at least keep it off the navigation structures.

About the Author (ATA)

Depending on the reputation of the author, this is variously retained at the front or the back. Remember there are no rules. It's your call and e-books should be crafted with the same thought as print books. So you can make up your own rules such as: Move "About the Author" to the back, except when notified to place it in the preliminary sequence.

ATA is usually kept as a simple single page for print to save paper. Seriously consider an expanded author bio, and opinion pieces, or an interview. Think "E".

Series-title | By the Same Author

If you book has a series title page this is a obvious opportunity to promote the other titles. At minimum you should have links to preview pages from your website. The fact that some reading devices don't support external links should not stop you doing this.

Your ePub may end up on a colour device. Rather than a simple list of titles, put cover thumbnails in, the summary or a short blurb and sell up. Series title page is essential in academic publishing, but can be used equally effective with other genres.

Title Page

So if you think the e-Bookshelf entry and cover is enough presentation of the title, you can probably remove this.

My opinion don't. It is an anticipation builder, and in non-fiction trade where strong design components have been used, it is a powerful tool to set the tone and design feel of the book.

Copyright Page

There is this eBook "thing" that the e-book copyright page is relegated to the back. It's kind-of a herd instinct thing - its an e-book so the copyright goes to the back right?  Why? Beats me. The only reason can be so it doesn't "clutter" the NCX or something.

You don't really have to move the copyright page to the end. It is a thing that started with the first generation of readers because they had hard-coded solutions - MS Reader for example. Given that the copyright page is probably the most important single section in a book (from an author and publisher perspective), it amazes me it gets so easily relegated.

Put your copyright wherever you think it should go. It can go in the NCX anywhere. This is even more important if you are planning a watermark sales strategy. The copyright page and title page are targets for visible watermarks and should not be relegated to the back.

Contents - TOC (Generated)

The NCX is a type of Contents, but because of weaknesses in the ePub2 spec many books also use an internally generated Table of Contents. ePub 3 has a powerful TOC.XHTML strategy. The TOC page can be used in the spine and can be reasonably styled. It is a real substitute for internally created TOC pages.

For ePub2 we advised to use an internal generated table of contents. This can have extended navigation, and allows back and forth navigation from section titles to the generated TOC page.

This is usually better than going to the NCX in limited reader devices. Finding the book NCX is sometimes hard to get to and slow to navigate. A book with more than 30 informational TOC sections definitely needs an internal Table of Contents for most mobile devices. 

With ePub3 the TOC strategy is probably strong enough for most books.

For standard fiction it is doubtful a generated TOC page has any value, but if you want it, do it.

Other List of's - LOI, LOT, LOF, LOM etc.

Ideally these are included and link to the top of their resources. EPub3 has support for a number of ListOf* structures.

These cross section links will/can perform slowly in some past limited resource devices (Sony Readers). These devices will not be around forever, and your ePub probably will be. Tag it correctly now and issue updates in the future.

This is where it gets tough for publishers and you have to make a decision, or at least have a clear set of guidelines. The correct thing to do is take that book with 80 figure references, link them, and test them on a device like a Sony PRSomething. Normal page navigation is a couple of seconds. If link jumping in a slow device is under 7 seconds, live with it. Put an apology on your blog blaming Sony.

Dedication and Epigraph

Keep them in the spine, and I certainly like to see them in the TOC/NCX, but that's your choice. They add a human dimension to the book. Some print books even put the dedication on the Copyright page. Depending on the length and layout of the copyright waffle, that may work.

Other front matter: Foreword, Preface, Introduction

From here on there are not many controversies and the e-book axeman is on holiday. These all necessarily remain in place.

Body

Second Half-title

Definitely keep them in the spine sequence, especially if a book is a collection you will need them as navigation targets. If your front matter has a lot of sections consider bringing the Second half-title into the NCX as well even though it may result in the repitition of the book title after introduction or something similar.

Part, Chapter

Obviously these are all going to stay, and in their right place and sequence.

The only comment here is whether, and how, you link the section numbers and titles to an internally generated TOC if you have one. Your options are:

1. Don't link them. Let the user use the NCX navigation.
2. Link them with Webpage Hyperlink blue and cheapen the look of your book.
3. Link them and use your CSS to maintain the default text styles and let the user discover the functionality. Give a subtle hover style.

The downside of option 3 is e-Ink devices do their own XOR text inversion on links to make them stand out, and touch screen devices like iPad do not directly support the CSS :hover pseudo element because of the non-continuous pointer.

Back matter

Appendix, Notes, References, etc.

Most back matter is safe from the e-book axeman's sword. Appendices, notes, references, etc. all obviously stay.

Index 

The Index has been a big victim of e-book torture.

When I see a SOP (Standard Operating Procedure) that says - "Delete the Index", or "Delete the Index numbers", I cringe.

We hear statements like "they can use the search tools in the reader". These do not work as effectively as well indexed books.

We never commit this crime in production. In IGP:Digital Publisher indexes are tagged and then processed in different ways for different books. Index linking is not particularly expensive and can be significantly automated.

Fortunately Small and Medium publishers are awake to the value of good index linking. When a book is created through retro-digitization of a PDF or hard-copy, it is easy to maintain the original page links. Therefore index numbers can get to within 2-300 words of their print position. About the same experience as a print book.

Index links can be dense, and cause limited resource devices to come to their knees. To address this we can optionally segment indexes. This has two benefits. It makes it easier to navigate through the alpha entries quickly, and it lowers the resource pressure on the device.There are no reverse links on Indexes, so this allows going round all references in an index term relatively friendly.

Next the page numbers can be turned to sequence numbers. This is not as good as the original page numbers which give a sense of position in the book, but is a good substitute. 

Excerpts

Excerpts are relatively high value promotions, especially if they have a little background info with them. Not all books can support excerpts from other authors, and where possible excerpts should support some value for the reader and not be just a shameless promotion (but that is an opinion). Remember there is no castoff calculation to worry about with e-books. Take advantage of that fact.

Back Ads

Usually these are removed, but not always. It depends on their currency and relevance. You know if yours should be there or not. 

Colophon

Not very common now-a-days. It is generally OK to remove this as it contains print edition information. Of course with a little effort it could be turned into an eColophon and contain information about the design and production of your e-Edition.

Summary

There is no rule or technical reason for any particular print section being included, excluded or moved. Rather than take a general approach, think it through, and put into place a strategy on a per-book basis when the content and end-user experience demands it. A continuous nail-biting novel read is not the same as browsing through a recipe-book, or a highly illustrated learning or reference book. 

And don't forget this is the next major change in publishing. Let's get some new section types defined that don't hark back to ink on paper, and the history and evolution of printed books.