The Most Useful DTD You’ve Probably Never Heard Of

I’m often asked to recommend a DTD, and after the usual "it depends" discussion--journals or books? what kinds? what will the XML be used for? who’s going to be working with the XML, and with what tools?--it becomes obvious that there is no one right answer. There are standards that have strengths in certain areas (NLM for journals, TEI for scholarship, DocBook for technical documentation, etc.), and there are standards that are useful for specific purposes (e.g., XHTML for online publishing and its close cousin, OEB PS, the Open E-Book Publication Structure, for e-books). Most publishers need to relate to one or more of these models, and modifications are often necessary to address all those "it depends" issues.

The process usually involves some combination of paring down, augmenting, and adapting a standard to fit particular products or purposes. One key issue is how "granular" to get--just tag the basics, or tag every little thing, or somewhere in between? The goal is to balance between, on the one hand, the richness and complexity that the content deserves and the users demand and, on the other hand, the practical issues like what the available tools, staff, and vendors can actually do, and how much it all costs.

It’s a shame more publishers don’t factor the DTBook model into their thinking. Ironically, although most people agree that accessibility is a good thing, to most publishers it’s a complication that they’d rather think about tomorrow. It may surprise you that the standards created for the purposes of accessibility can be, well, accessible. Reasonable. Not rocket science. Worth a look!

The DTBook DTD is part of a broader standard for Digital Talking Books (DTBs), defined as ANSI/NISO Z39.86, which encompasses a broad range of audio, text, navigation, and other file formats. The standard for markup of text is the DTBook DTD, and its main purposes are (1) to provide sufficient structure so that the content can be properly navigated by reading systems, and (2) to make it easy to incorporate alternative versions of things (like verbal descriptions of visual elements) and tell what they’re for. Rather than being a whole new vocabulary, the DTBook DTD is based on HTML, which is widely known and needed by most publishers anyway.

This DTD was originally created by the DAISY Consortium (Digital Accessible Information SYstem), an international body that did not mandate any particular degree of markup. Basically, the more markup, the better. The consortium recommends, at a minimum, the markup needed for print publication. In 2005, NIMAS (National Instructional Materials Accessibility Standard) made things even easier by creating an implementation of the DTBook standard that specifies both a Baseline Element Set (the minimum requirement) and the Optional Elements (the rest of the DTBook standard).

The NIMAS Baseline Element Set is really the "no-brainer" stuff that you’re probably marking up one way or another already: basic document-level tags for structure and hierarchy (e.g., <level>, <level1> through <level6>, <h1> through <h6>), block elements (lists, notes, citations, sidebars, definitions, etc.), inline elements (emphasis, page numbers, etc.), and some tagging for tables and images. The Optional Elements Set includes a lot of interesting and useful things (and it will become even more useful as publishers do more things with their content), but the key is that it’s optional. (In focusing on markup, I’ve glossed over metadata; suffice it to say that the metadata specifications are similarly reasonable and useful.)

I’m not suggesting that you just adopt the DTBook DTD wholesale--far from it. That’s why I started by pointing out that there are all those "it depends" factors; there never can and never will be a "one size fits all" DTD. But I really want to encourage you to have a look at the DTBook DTD and its implementation by NIMAS. I think you’ll find that it will provide a useful perspective that’s surprisingly practical. And I think you’ll be pleased to discover how easy it should be to produce at least minimally accessible files.

Next SSP Tech Topic: The new Open E-Book Standards (which, by the way, incorporate DTBook XML. . . .).

Acknowledgments and some useful links:

Most of what I know about accessibility was either taught or pointed out to me by Rick Bowes. Definitive information can be found at the DAISY, NISO, and NIMAS Web sites mentioned earlier; all these sites have both detailed technical information and user-friendly explanations.

Are you interested in other types of scholarly publishing standards? Visit the Standards page on SSPNET.ORG.