Mathematics in Ebooks

Electronic books (ebooks) are used more and more in education, and mathematics is an important part of this. Design Science is working with the publishing community, ebook standards organizations, and reading system vendors to ensure that mathematical notation is handled properly. Besides this page, we have also created an Ebooks for Math and Science group on LinkedIn, the business networking website. We will use this group to exchange documents, ideas, and make announcements regarding this effort. If you are interested in this subject, we encourage you to join.

Equations in EPUB

The EPUB standard is published by IDPF (International Digital Publishing Forum). While there are other ebook formats, EPUB appears to be gaining a large following as it is an open, non-proprietary standard. EPUB is actually several standards in one, each dealing with a different aspect of the book. For equations, we are only interested in the content format. EPUB allows two different formats for the content, XHTML and DAISY. However, when the ebook world talks about the EPUB format, they are almost always talking about XHTML, not DAISY. The latest version of DAISY already has math support anyway as it supports MathML, so we also will restrict the discussion here to EPUB's XHTML variant.

Specification: Best Practices for Mathematics in EPUB

We have created EPUB Math: Best Practices for Mathematics in Ebooks in order to provide guidance to content providers and reading system developers on how best to handle mathematical notation in EPUB ebooks. The EPUB standard is complex and most reading systems do not implement all its features. While reading system vendors may be committed to implementing the missing features eventually, their implementation priorities are set according to which features are needed to support specific books and book collections. This produces a chicken-and-egg situation which we hope to break.

This best practices document recommends a method for representing equations in EPUB books involving specific use of EPUB features. This representation attempts to fit within the following goals and constraints:

• It adheres to the EPUB standard and works within it rather than extend it. Adding new features and improvements may be an activity we will be involved in someday but we want to work with the current version of the standard for the current effort.
• Equations should be scalable such that they display well on various devices and zoom levels.
• Each equation must align with the surrounding text (baseline alignment) and maintain that alignment in the face of font size and zoom level changes.
• We want to enable reading systems to allow equations to be copied to the OS clipboard to allow users to work with mathematics in other applications.
• Equations should be accessible to people using screen readers with their ebook reading systems. While DAISY books will give the best reading experience for people with disabilities, accessibility is important with all ebooks in education.
• Equations should be available for math-based search.

Our proposed solution involves incorporating MathML to represent mathematical meaning and structure, vector images in SVG (Scalable Vector Graphics), as well as fallback raster images (GIF, PNG) for those reading systems that lack SVG support.

Sample EPUB Books

The following sample EPUB books have been constructed in accordance with our EPUB Math: Best Practices for Mathematics in Ebooks specification. They can be used to test reading systems, as well as being helpful to publishers attempting to implement workflows:

• EPUB Math Sample Book (.epub file): This book contains one inline equation and one display equation and very little text.
• On Optimal Bit Allocation for Classification-Based Source-Dependent Transform Coding (.epub file): This is a version of one of the EPUB samples from Hindawi Publishing which has been modified in accordance with our specification. We thank Hindawi for allowing us to do this.

Reading System Compliance

Our EPUB Math: Best Practices for Mathematics in Ebooks contains a requirements checklist for assessing reading systems compliance. As of this writing, only Adobe Digital Editions fulfills those requirements. It is our expectation that this will change soon as most reading system vendors have pledged to add SVG and embedded font support soon, as required for full EPUB standard compliance. Whenever any reading system achieves compliance, we will report it here.

In theory browser-based eReaders, such as Bookworm and BookGlutton, should be able to render the MathML in EPUB books when opened on MathML-capable browsers such as Firefox and Internet Explorer + MathPlayer. However, in practice they currently do not. The browser-based eReaders parse ePub files to their standard and, in so doing, filter out requirements for browser rendering of MathML (for example, some browser-based eReaders serve ePub files as HTML). Browser adoption of the HTML5 standard, which includes MathML, should improve this situation. MathML support, though, is just one step to compliance of EPUB Math: Best Practices for Mathematics in Ebooks.

Ebook Compliance

We would like to extend the epubcheck tool to report the number of mathematical equations present in an EPUB book and check that they follow our best practices specification. As of this writing, this work has not been started.