ebookcraft 2016

April 10, 2016

In our third year, by all reports, ebookcraft hit its stride. We had a sold-out crowd of engaged professionals from Canada, the US, and across the pond, talking about all aspects of ebooks – making, optimizing, testing, designing, indexing, enhancing, and improving. It was a dynamic couple of days. 

Day one was a day full of workshops. There were two streams of content, forcing some hard decisions about which room to position oneself in. Monotype type designer Charles Nix started with a lecture advocating for good typography practices in ebooks, taking us through a full set of acknowledged “rules.” Pointing the crowd at the Type Director’s Club resources, he cautioned us to read the content we are designing, to choose fonts designed for digital, with full glyph sets, appropriate x-height, and which are hinted, kerned and ready for deployment. “As they say at the Westminster dog academy, I am a bitch of quality.” He implored us to use true small caps, and to use bold judiciously. He railed against forced justification in ebooks – “justified type is for fools” – claiming that the screen real estate and agility is just not there. Champagne Choquer took up where he left off, going over the font embedding process, which kinds of fonts to choose (TTF, OTF, WOFF). She framed her talk as the hero’s journey: developers putting time and effort into embedding fonts are thinking about the reading experience, branding, and design. The font embedding hero, Champagne said, is also someone who can get through the average font foundry’s EULA. 

Firebrand’s Joshua Tallent walked us through the basic elements of ebook design and quality assurance testing. His talk included references to loads of handy resources, like the Google group for epubcheck and Dave Cramer’s Requirements for Latin Text and Layout Pagination, and retailer-specific issues to look for during the QA process. Joshua pointed out that the print version of the book will always be the first thing that people compare the EPUB to, so it’s important to start with that. He outlined four variables that impact how an ebook is displayed:

  • The operating system of the device the ebook will be read on
  • The display engine within the specific reading system
  • The reading system hardware
  • Hardware

He also went through the four elements of ebook Quality Assurance testing:

  1. Automated validation – Test your EPUB through a validator like the one provided by the IDPF.
  2. Retailer Requirements – Familiarize yourself with the vendors’ specs where available. Apple requires landmarks with TOC, cover and start location, whereas Kindle only needs the TOC and start location. Kobo still can’t support font obfuscation, but they do have some support in place for JavaScript.
  3. Side-loading and Manual QA – Joshua recommends acquiring an iPad for sideloading, because it supports many of the major vendors’ apps, as well as testing your ePUB on an Android tablet, as well as on eink devices like the Kindle and Kobo ereader.
  4. QA checklists – Joshua recommends putting together a QA checklist that you can use for each ebook. He’s created a detailed list on the FlightDeck website here.

Pilar Wyman lead a workshop called “Matrix Revolutions: Ebook Indexing.” Creating indexes in ebooks provides some interesting new features that are not possible in print, like implementing filters and collapsible groups, but they also pose new challenges, like figuring out how to anchor entries in the absence of static page numbers. Pilar outlined some of the improvements to ebook indexes in the ePUB 3 spec, such as index filtering, range highlighting and interactive generic cross-references. Not every book will require the same type of index – for each book, you need to figure out which kind of functionality will work best for the material and the readers. Indexes can either be linked or embedded in content files. Linked entries use XHTML anchors or unique IDs. Embedded entries are inserted directly into content as fields, XML elements, or using the program’s own unique marking system. If you’re new to indexing, Pilar advises starting with embedded over linked entries. 

Nick Ruffilo from Aer.io (Ingram) was up next talking about the basics of Javascript development. An old tool in web development, Javascript works well in iBooks and Kobo iOS and Android. “Computer science is a science. But programming is a liberal art.” Nick’s very unique and dynamic presentation included programming him to do jumping jacks and speaker-supplied snacks (as he had the pre-lunch slot). Nick looked for some ebook problems to solve – for example, a flip book – and took the crowd through the process of developing a script (using UltraEdit to write code, and Firebug as an inspector). Stating a best practice that would be common through ebookcraft, he spoke on the importance of adding comments to your code to avoid extensive code archaeology later on. Demonstrating his impressive Googling skills (take note developers!), he led us through creating a couple of sprite animations. 

After lunch, the University of Toronto Press’ Sylvia Hunter and Terri Rothman walked us through their press’ in-house XML-first workflow. Archiving content in XML enables you to future-proof it, and easily locate files and chunks of content. They walked us through their book production process, which begins with manuscripts in Word with eXtyles tags added to them. The eXtyles software enables them to export valid XML styles. They then move to XML editor oXygen for final processing, and add XSLT stylesheets so that the XML can be translated into multiple formats. The files are then loaded directly into to InDesign, and EPUBs are created and edited in Sigil. On the QA end, Sylvia and Terri warned that just because your content is valid, doesn’t mean it’s correct. Right now they do manual QA testing in a variety of apps and devices – Adobe Digital Editions, Kobo, iBooks, and Kindle, among others. Implementing the use of schematrons is on their roadmap for testing in the future.

In the opposite workshop room, I spoke on ebook accessiblity. My session’s subtitle was “It’s not just ALT text” and was focussed on levelling-up the general quality of ebooks that we produce. This session was constructed around hands-on tips and things that go beyond minimum product viability. Pointing the crowd toward both Matt Garish’s Accessible EPUB 3 and the new BISG Quick Start Guide to Accessiblity, I started with a quick lecture on why we are failing our readers by using minimum InDesign exports, and then moved on to live tips to make better EPUBs, even with InDesign as a content starting point. In practical terms this means adding HTML5 tags, a full set of epub:type semantics, and a page-list – for which I have had a hand in developing an InDesign script. I also lead the crowd though a live demo of a text-to-speech reader (during which my husband texted me a parenting Q which went up on screen – it’s not easy being a woman in tech!).

Colleen Cunningham gave us a sneak peek at the F+W Media Content Management System, designed in conjunction with Librios. Migrating to a CMS workflow can be a big undertaking for a company, and requires buy-in from a lot of different departments. Colleen talked a bit about the way that they built confidence in the system by demonstrating success, and identifying people in different departments who were willing to try new things, and bringing them on board. F+W uses their CMS as a repository of live content, the content is not exported to different formats until the very last step. The content is tagged via XML markup, and filters are used to translate between CMS tags and app styles. Corrections for both print and ebook content take place within the CMS. Colleen noted that although a CMS can significantly improve your workflow, there are some things it can’t do; for instance, make editorial decisions for you, or define what your workflow should be. Not all content will be appropriate for the CMS. Moving to a CMS workflow does have the capacity to greatly increase the efficiency of production workflows, though and to structure content in simple, logical forms that are easily kept up-to-date and searchable.

Derrick Schultz, a Lead Product Designer from the New York Times, led a workshop on managing you CSS via SASS mixins – for snippets of code that you want to use over and over. He likes to push the ebook format with a stubbornness that is unusual (impossible ebook tricks here). Before the conference he invited people to send in their ebook problems for which he would find solutions or workarounds. He demonstrated how to access iBooks inspector, media queries, how to embed an image in an <hr>, centering images with an attached caption, hacking iBooks themes, and animated section breaks. He encouraged collaborative troubleshooting via his 99 Problems GitHub repo. 


The ebookcraft main day was populated with shorter, more theoretical talks on the work of ebook developers. In my opening words, I invited people to check their print fetishes at the door and prepare themselves to be wowed by the collaboration and community on display.

Greg Albers, the Digital Publications Manager at Getty Publications, talked about the shifting fundamentals of book publishing in “Closer to Metal.” Greg says there is power and control in getting closer to metal – and by this he means that understanding and using code and markup gives all the people in the publishing process authority. He put the audience through a simple book-making craft to demonstrate the form and innovations on the form of the book. Knowing how to make books means understanding the form – that is, the code. Knowing how to read and write HTML, understand CSS, and understanding the collaborative nature of coding is key to the publishing process. Use GitHub, comment your code, put a digital colophon in your ebooks, separate content from style so that the interface doesn’t get in the way of making ebooks. His GitHub repo full of great resources is here

Sanders Kleinfeld, Director of Publishing Technology at O’Reilly Media, started out his “Ebook Developers Toolbox” talk with the inconvenient message, “Ebooks are made of code.” The main tools he advocates is XSLT as a reliable powerful alternative to RegEx processing, responsive ebook design techniques, and equation processing with MathML Cloud. He walked us through the idea of a single ebook archive for all platforms that is designed for user-configurability and responsiveness. His own iBooks reader detector can be a helpful tool.

Iris Amelia, a long-time eprdctn regular, gave a rousing talk next on the value of storyboarding in ebook production. Developers are tasked with creating and curating content that is more accessible across mobile platforms. Storyboarding is a useful tool for rethinking the content visually and establishing linearity. The focus of storyboarding is grounding the user and visualizing their experience of the content. Her steps: 

  1. Read the book. 
  2. Attack: parse it out and put it all down. 
  3. Chunk the content. Find the patterns, colour code. Bonus: get high off of markers. 
  4. Organizing the chunks. 
  5. Sketch. Work fast, don’t overthink, don’t be afraid to start over. 

Iris suggests that visualizing the content will improve accessibility and experimentation. Publishers don’t ask “What if?” nearly enough. https://twitter.com/geealbers/status/715564685030862848

John Rodzvilla, a digital publishing educator at Emerson College in Boston, spoke next on papercasting – a process designed to test user experiences without doing a software build which literally uses paper and stickie notes. He suggest that developers are not the user and need ways to think through features and aesthetics that don’t hinder usability. No one thinks about how to use a book; if instructions are necessary, there may be a serious usability problem and might mean publishers need user research – which is where low-fidelity paper prototypes come in. Experience shows that users (readers) are more likely to point out problems from a papercast of the project than from an expensively developed build. John says, “I never thought I’d be up on a stage saying ‘You need to reset the book.’” Other lessons learned are that interactivity is still unexpected, always provide simple direction, and don’t overwhelm the reader: less is more. 

Lerner Publishing Group’s Kris Vetter led a wide-ranging demonstration of CSS3 animations after lunch. She demonstrated several animations she’s worked on and initiated a discussion of how animations add to or distract from the content. Largely in the context of children’s publications, she talked about the technical needs of the artwork, and walked through the animation presets in InDesign, creating custom motion paths, using button triggers, and building in timing delays. She strongly cautions developers that just because you can animate doesn't mean you always should. Think of the end user and if animation adds context, assists narrative, and enhances reading comprehension. Reinforcing another powerful ebookcraft theme, Kris cautioned us to consider the user experience.

Teresa Elsey, a digital managing editor in the trade division at Houghton Mifflin Harcourt, took us through the mind-boggling amount of work required to maintain a substantial backlist, a task complicated by old ebook standards. Some issues that Teresa touched on:

  • Ebooks don’t show their “vintage” on the outside creating a confusing reader experience.
  • Older ebooks are ticking time bombs in terms of coding practices, accessibility, rights and permissions, out-of-date information. Pluto is no longer a planet, and racist language becomes unacceptable, for example. 
  • Metadata needs constant attention and maintenance.
  • Link and reference rot.

Teresa advocates that we develop a process, take small bites, improve our tools, make it easy to update, keep good notes, and accept that it's hard.

Our penultimate session was a talk by Tzviya Siegman and Dave Cramer on a practical implementaion of EPUB+Web. “Since the dawn of time... people have been complaining about EPUB,” said Dave. Their talk centered around how EPUB 3.1, a dynamic changing specification, is simplifying, moving closer to the web, upping it’s CSS game. Dave and Tzviya told us about some spec elements that are disapperaing – CFI, epub:trigger, epub:switch, epub:bindings, guide, for example – and one key one that is deprecating: the hated NCX. Reinforcing the accessibility message from yesterday, Tzviya is adamant that a11y not be sacrificed for design. The idea here is that EPUB should be easier to deploy on the web by making it conceptually simpler. Tzviya implored the audience to contribute by sending feedback, use cases, reporting bugs, or by joining the standards conversation. 

Understanding that the crowd would be tired at the end of day two, we planned a slightly more light-hearted session – a debate on the importance (or not) of ebook design. Moderated by Derrick Schultz, India Amos argued against, and Nick Barreto in favour of whether or not ebook design matters. We invited the crowd to participate in a Twitter poll before and after the debate to see if the speakers moved the needle on the question. While Nick ably defended design – people will pay more for a well-designed product, good design means usability and readability, design matters in everything – India was definitely the more dominant speaker. Scratching out the lyrics to an eprdctn punk song – Everybody hates your drop caps. Everybody hates your sinks. Everybody hates your CSS. Reading systems hate your design, hate your media queries, hate your CSS – she managed to move the before and after Twitter poll in favour of obliterating ebook design. 

At the end of the day, the So You Think You Can Code design contest winner was announced. Rebecca Springer of Houghton Mifflin Harcourt took home a $2500 cheque from Kobo, a year’s subscription to Adobe’s Creative Cloud, a three-month Firebrand subscription, and eternal nerd glory. We ended the day at a local pub with a trivia night sponsored by Litsy and with questions especially devised by the ebookcraft steering committee.