Single Source Publishing: Creating Customized Output

6 minute read
Barry Schaeffer avatar

The information world is exploding with new technology and devices of all sizes and shapes, bringing with it unprecedented growth in demand for new information products and the content to drive them.

Keep Up or Fade

Today, publish transparently across different types and sizes of output devices, and tailor your information to each user group’s specific needs and desires, or you risk becoming old hat.But as consumption grows, the burden on content creators and managers grows with it, forcing publishers to scramble for answers.

If that describes you, one answer may come partly from something called “single source publishing” or “SSP.”

Enter Single Source Publishing

Single source publishing, vendor hype surrounding it notwithstanding, is a broad approach to content aimed at enabling creation of information products targeted to specific audiences, automatically -- and without manual intervention or rework.Although decades old, the techniques of SSP have never been viewed with the urgency that they evoke today.

SSP is often synonymous with “component content management” or componentization: breaking your content into small pieces for collection in many different ways to create different desired outputs, like Scrabble where one letter per piece allows you to spell any word.

Standards often imply componentization as well:DITA is designed to manage and collect components (or “topics”), and technical standards like S-1000 also use componentization in their architectures.

So, for better or worse, any discussion of SSP is partly a discussion of componentization:

Components: How Big vs. How Many

Don’t get me wrong, componentization works: it has many adherents and a long history; it’s enshrined in standards; and it’s supported by a growing body of software.But it’s not without its limitations, a big one being “fragmentation,” the two-headed monster of component size and volume.

If you need highly granular output, your components must be highly fragmented, providing the flexibility you need but often creating a blizzard of tiny components difficult to author and maintain.With larger components, you have fewer pieces to manage but face increasingly redundant content that must be revised everywhere it occurs.

Either way, the process works but is often less efficient than it should be. Worse yet, you may not know you’re in trouble until your system is in operation and decisions are made that are difficult to undo.

Finding the Component Size Balance

The solution, in a perfect world, is having the needed output granularity without greater fragmentation or redundant content.Two techniques often overlooked in planning and implementation may hold the key.

The first, sometimes called “variant tagging” involves tagging within individual XML components making them usable in multiple contexts.If an element appears with some differences in output variants A, B and C, the original element may be tagged to allow its filtering for each output usage.

DITA provides for this capability through “conditional processing” and some commercial tools support its use.Likewise, S1000d provides a similar capability under its “applicability” tagging.

Around since the mid-90s, variant tagging has proven quite powerful, with or without a CMS, by allowing a single element to conform to multiple contexts where it is needed.

Learning Opportunities

Ford’s publishing vendor, Tweddle Litho, for example, facing 12 model years of owners manuals always live, used variant tagging with a CMS to collapse the 12 discrete model year files into a single file, extracting desired year versions as needed.The filtering process can be provided by a commercial publishing tool or, as in this instance, by custom scripts applied at publication time.

A second technique, valuable especially with technical material, uses embedded database queries in place of actual content.Reference to a certain part description, for example, may be represented by a query to a parts database for replacement at publishing time.

Using this technique, Pratt & Whitney Canada inserted queries to its engineering task systems in each change instance, allowing it to automatically include the latest required task data for each change at publication time.Combining these queries with variant tagging, PWC was able to tailor the results to multiple outputs from the same content.

Technology Aside, Careful Planning is Still Your Best Tool

Even in today’s technology-heavy world, whether or not you use a standard like DITA or s1000, without careful planning you’re unlikely to achieve the best environment possible.

Variant tagging, for example, should be planned into every phase of the content life cycle, from editorial tools, to your chosen CMS -- or no CMS, to adoption of a standard or custom content architecture.The vendor community may not make this easy for you: each vendor wants to sell you its solution based on whatever it does best, and if that isn’t right for you… well, buy anyway.

So it will be up to you to take the time and effort to look closely at your output requirements, the behavior of your target audiences, the nature, volatility and complexity of your core content, your internal resources and constraints, and describe a comprehensive plan that fits your needs. Then, armed with that plan, you can confront the vendor community and decide on software tools that effectively support your chosen architecture.

Kidde Aerospace of Wilson NC is a good example. With a tiny publication group, Kidde needed to produce unique technical manuals for nearly 150 airline users of its aircraft fire suppression and oxygen systems. Lacking budget for a CMS, Kidde used careful planning to build a fully functional publication environment based on variant tagging, supported by low-cost SGML editing/publication software.As Kidde learned, the greatest challenges aren’t technical, but are functional and organizational instead.

This kind of searching analysis isn’t trivial by any means, but if you don’t do it up front, things will get increasingly more difficult and expensive to fix when you realize you’re on the wrong track.

The good news is that if you are careful, get help where you need it and keep your wits about you, it can be done.

Editor's Note: You may also be interested in reading:

About the author

Barry Schaeffer

Barry is Principal consultant with Content Life-cycle Consulting, Inc. In 1993, he founded and led XSystems Inc, a system development and consulting firm specializing in text-based information systems, with industrial, legal/judicial and publishing clients among the Fortune 500, non-profit organizations and government agencies.