Publishers may distribute E-Books in a variety of formats, where ‘format’ means that the internal structure of a given file conforms to a defined standard. Understanding the different formats can be challenging for people who are new to e-book production, and even those of a more-technical bent can find the situation confusing initially. This appendix explains the differences between the major e-book formats.
The initial factor to appreciate is the great variety of e-readers that exist. Currently, Wikipedia lists forty-two dedicated e-reading devices that are available, and an even larger number of discontinued models. Moreover, people use desktop, laptop and tablet computers, and smart-phones too, to consume e-books by means of e-reading software running on those devices.
There is also a wide variety of e-book file-formats, where those that enjoy broader support include plain text (i.e. ASCIIor UTF-8encoded files), HTML, Rich Text Format (RTF) and Adobe's Portable Document Format (PDF). However, the formats that are of principal interest to the e-book producer, and which this appendix considers, are EPUB, Mobi and KF8.
The EPUB format is a free and open e-book standard that the International Digital Publishing Forum (IDPF) developed and maintains. Here, ‘free’ means that documentation describing the standard is available to all at zero cost, whereas documentation concerning proprietary standards may not be available publicly. ‘Open’ has a looser definition, but can be taken to mean that all salient aspects of the standard are documented (which may include even the standard's development path), whereas a closed or partially closed standard will not document every aspect. EPUB superseded the older Open eBook standard, and the current version is 3.0.1. It is the format in which you should distribute your e-books when targeting the non-Amazon e-reading market.
EPUB files are simply zip files that carry a ‘.epub’ filename extension, and a painfully technical method of creating an EPUB file is to take an existing one, and edit its contents. For example, if you have a file called MyBook.epub, which you rename MyBook.zip, you can decompress that file to a directory structure on your computer. You can then delete/insert image, font and HTML etc. files from/into that directory structure manually, before re-compressing it and then renaming it with a ‘.epub’ extension.
This process also requires manual editing of certain control files in the directory structure, which complicates an already-complex and error-prone approach, and is of no value given that there are far better methods, such as usingeditors such as Calibre and Sigil, which automate matters greatly.
The Mobi format (also known as ‘AZW’) is the proprietary format that Amazon's initial e-reading devices supported. It is sorely limited typographically, and so is best suited to simple books (in terms of styling and layout rather than content) such as novels, where the content comprises essentially a linear run of text, separated only by chapter breaks, headings and perhaps sub-headings.
That is, more-technical works such as the content in this guide are impracticable in the Mobi format because any HTML tables and SVG images etc. will not be visible, nor will the ‘floating’ of elements (in theCSS sense) work either.
The truly critical point, however, if you are confused about the various formats, is that there is a difference between Mobi-format semantics and files that possess a ‘.mobi’ filename extension. The final section below explains this in more detail.
The Mobi format is now outdated, and Amazon announced in 2011 that it would be supporting it no longer. This underscores the argument pointedout in Chapter Two, Resources, that while the majority of e-books will work in the Mobi format, it is better to abandon it when producing technical publications or those with richer, more-structured content.
As a welcome replacement for the Mobi format, Amazon announced the KF8 format, which improves upon Mobi greatly. KF8 supports a subset of HTML 5 and CSS, allowing authors and publishers far greater latitude in the way they present e-book content. For example, the sidebars, SVG images, and features such as individual tables of contents that you see in the chapters in the e-book version of this guide are all possible by dint of KF8. Do not be fooled, however, as KF8 is essentially EPUB with a few changes here and there. Note also that while the format is called ‘KF8’, KF8 files have a ‘.azw3’ extension to their names, even though a ‘.kf8’ extension would have been far less confusing (and supporting EPUB directly would have been far better for all aside from Amazon, not least because it would simplify production and testing greatly for producers).
With these points in hand, another confusing issue for the e-book producer arises from the file that KindleGen or the Kindle Previewer (which incorporates KindleGen) generates. As Chapter Seven, Production, pointsout out, you should generate an EPUB version of your book first, using a tool such as Sigil. You can then upload the EPUB file you generate to non-Amazon retailers unchanged. However, to generate a file that you can upload to Amazon, you should feed the EPUB file to the Amazon KindlePreviewer application (not Kindle for PCs), which will generate a file with a ‘.mobi’ extension.
This is where the equation becomes somewhat muddled, as that file, despite the .mobi extension to its name, actually contains two versions of your book, one in the Mobi format, and one in the KF8 format (which is why this text refers to such files throughout as ‘Mobi/KF8’). When testing your book in Kindle Previewer under the Kindle DX setting, that program emulates a hardware reader that supports only the older Mobi format, and so it draws the content it displays from the Mobi part of the Mobi/KF8 file that it generated. However, if you set it to operate like a more-modern Kindle, it will draw the content it renders from the KF8 part of that file.
The same principle holds when testing your e-book in Kindle for PCs. Here you copy the Mobi/KF8 file generated by Kindle Previewer to the directory in which the software's e-book library resides, but when you open that file in that application, it will use only the KF8 version that the Mobi/KF8 file ensconces. The situation is identical when testing on a hardware Kindle, in that a first- or second-generation device will use the older-format part of that file, and a more modern device will use the KF8 part.
In line with this, uploading the Mobi/KF8 file to Amazon means that customers with older devices will receive the Mobi part of the file, while those with more-modern devices will receive the KF8 version (i.e. a file with a .azw3 extension). This implies that it is possible to split the Mobi/KF8 file into its constituent parts, and an extension is indeed available for Calibre that will do just this, generating two files, one with the .mobi extension and one with the .azw3 extension.
If you have Calibre, and you wish to explore this issue by installing that extension, you should launch the application and then click on the Preferences button. Then select the ‘Get Plugins’ option that you see, and go from there.
This is of value when you wish to ascertain the exact size of the KF8 file that customers will receive. Asnoted in Chapter One, Strategy, Amazon charges for the cost of transmitting the file to a given e-reading device (under the 70% revenue model), where a large number of raster images in a book will increase the size of that file dramatically, thus eroding your profits significantly. This means that the 35% revenue model may be more favourable ultimately, and so knowing the precise size of your book's KF8 file will enable you to make the right decision in this respect. However, a much easier way of determining the size of the deliverable is to consult the compilation details that Kindle Previewer gives at the end of the compilation phase. Chapter Seven, Production, givesdetails of this.
Amazon announced in May 2015 the launch of an updated format, known as ‘Kindle Format Ten’ or ‘KFX’. In August 2015, all Kindles manufactured no more than two years prior to that date received an automatic update, to allow them to take advantage of the new format. Amazon stated that it would provide better support for automatic hyphenation, kerning, drop capitals and the like, however, it is not possible currently for publishers to produce KFX-format files directly. Instead, Amazon process KF8 files to generate KFX files in-house.