An opinionated guide to digital publishing specifications (guest blog)

(This is a reproduction, with permission, of a blog published by Liza Daly, published on the Safari's blog, on the 22nd of January.)

The World Wide Web Consortium (W3C) is a standards organization serving the “open web” — the set of freely available specifications that underpin most of the visible internet. In the years since the W3C was founded, all modern businesses have become “web” businesses, with their own industry-specific processes, jargon, and priorities. To that end, the W3C has formed interest groups for those industries which are adjacent to the web, with a goal to promote web technologies and ensure that the web is meeting common commercial needs.

I was co-chair for the Digital Publishing Interest Group for a time, and I have first-hand exposure to their work in interviewing publishers, documenting best practices, and writing recommendations for future specifications.

One of those deliverables is an intimidating table of W3C specifications and standards that were considered relevant to digital publishing. There’s a lot to digest there, and it’s unlikely that any single human is deeply familiar with all of it. I’ve provided an opinionated gloss of the most relevant or active standards, and feel free to comment if I’ve disparaged or ignored your favorite specification.

The audience

I’m assuming that the reader is one of the following:

• A developer who is working in digital publishing
• A curious non-developer who isn’t afraid of the word “normative” and acronyms that begin with ‘X’
• A standards wonk who wants to be more familiar with publishing activity

These are the “bread and butter” of digital publishing — whether it’s commercial ebooks, academic publishing, or journals:

HTML

HTML5 is a monster of a spec, but at least it’s reflective of current browser support. You should be familiar with the basics of markup, as well as the sections on browsers and common APIs.

CSS

There’s the workhorse CSS 2.1 specification which has been around for a decade. Unfortunately for the curious but lazy, all the cool new stuff is in CSS3, and that spec is broken out into many modules. Here’s a drive-by of the most interesting or publishing-relevant ones:

• Start with Dave Cramer’s highly readable Requirements for Latin Text Layout and Pagination (“Latin” here means Western languages, not veni, vidi, vici). Note that this is a requirements document, not a spec, which means much of what Dave recommends won’t actually work anywhere yet. Welcome to standards!
• CSS Text Module Level 3 is the “real world” equivalent to the above. Though it’s technical a spec in-progress, most everything in here is available in modern browsers and reading systems.
• CSS Regions Module Level 1 is a good read when you want to be angry about something. Regions can do some amazing things for advanced layout, but there’s a long and sordid history behind their implementation and deployment. There’s a lot of momentum behind getting Regions or an equivalent standard moving again, so there’s hope.

Extra credit assignments: CSS Media Queries and CSS Fonts Module Level 3. And while it’s unlikely that you’d need to actually read the SVG and MathML specs, it’s important to be familiar with those formats at a high level.

Accessibility

The simplest way to approach accessible web or ebook content is to study the semantics that are built in to HTML5. High-quality semantic markup will not only help a range of human users, it’ll aid in discovery and ranking by search engines.

Follow that up with the non-technical best practices in Web Content Accessibility Guidelines, and this overview of creating accessible interactive content.

XML

It’s not dead yet! There’s a lot of cruft in the list, but ebooks are still required to be well-formed XML documents, and academic publishing remains dominated by XML (and, sigh, PDF).

• Extensible Markup Language (XML) 1.0 (Fifth Edition) The ur-spec. If you’re new to XML, don’t try to read this.
• XSL Transformations (XSLT) Version 2.0 Even if you never write any XSLT, you should know what it is and when it’s useful. There’s a version 3, but even version 2 is only somewhat common; you may need to refer back to XSLT 1 to work in Python or many other languages.

Bleeding edge

If everything above is old hat, check out the emerging specs on the Shadow DOM, CSS Flexible Box Layout Module Level 1 (flexbox), and Packaging on the Web.