When Glenn Emerson, Information Architect for Xerox Global Services, read Point/Counterpoint-Is the XML Solution the Only Solution?(DCLNews, November 2007) he could barely contain his passion. Glenn was so passionate, we asked him to write his own article on debatable XML assumptions and how problematic they can be in real-world scenarios. Assumptions include, "Moving to XML is required for content reuse, Once in XML, all is golden, and Removing content from format is always a good thing." So before you go and make any assumptions of your own about your content management solution, we suggest you read what he has to say. Besides we all know what making assumptions does to you and me.

When documentation chaos finally tips the scales toward finding a solution to manage content, the question invariably becomes Is XML the only solution? I can't help but reply "No." The idea that it's the only solution implies a universal proposition. All that's needed to defeat a universal proposition is one example of the opposite.

While XML offers tremendous advantages for content reuse, adopting it as a medium for content management is not a free lunch, even with open standards such as DITA lowering the cost barrier for adoption. In the original article, it was suggested that failure to fully jump into an XML content management solution just results in old paradigms continuing. Yet, it was acknowledged that many people don't have the knowledge or understanding for a total immersion approach, begging the question, If they don't have the knowledge, how would they cost-effectively make such a complete and total change? There is, inevitably, some lost productivity during the adoption of any content management strategy. Time must be spent to analyze and convert content, train staff, model content and solutions, configure the components, develop publishing transformations and presentation stylesheets, and so on (Rockley, et al, 2003). Business does not stop during this time: Work still has to be performed, product launch schedules met, content developed and delivered.

Companies are not in the business of wasting money, so return on investment must be considered.

Companies are not in the business of wasting money, so return on investment must be considered. Breaking down the silos to achieve a unified content model costs money. That cost has to be offset by a return through some combination of increased productivity and cost avoidance.

Ann Rockley's silo model is an excellent illustration of the concept and the consequences of traditional approaches. However, Rockley concedes that it does not make sense to force all content into a unified model-for example, there is plenty of content that won't be reused, for several possible reasons (Rockley, et al, 2003). But, just because the content won't be reused does not mean it is without business value and does not need to be managed. There may still be advantages to managing that content with a CMS, and choosing a CMS that has the ability to manage legacy desktop publishing files such as FrameMaker, Quark, and Word files, instead of, or in addition to, XML content.

Several problematic assumptions are made when taking the stance that XML is the only solution:

• Moving to XML is required for content reuse-False. Content reuse can be accomplished in other, less expensive ways. Many technical communicators use solutions such as FrameMaker and WebWorks to single source content. Some proprietary Learning Content Management systems enable reuse through proprietary means. The business question is, of course, is this adequate?

• Once in XML, all is golden-False. Go ask any aircraft manufacturer who has had to support numerous DoD DTDs such as ATA and S1000D. For that matter, consider the difficulty of reusing user manual content in highly formatted, presentation-sensitive marketing brochures. Just because content is in XML or SGML doesn't help much if the documents have different content models or usage contexts. The effort to reuse S1000D DMs (Data Modules) as SCORM objects is a great case in point. SCORM is a reference model; S1000D is a structural content model. SCORM can reference anything, in the same way a library catalog index can reference any media, without imposing restrictions on the content itself. However, just because you put a SCORM wrapper around a DM does not make it an effective learning instance. Reuse between usages contexts is not impossible, but it will require thorough analysis, careful adherence to editorial style and content business rules, and mechanisms for transforming content components between the two usage contexts. Once such a system is built, change will be difficult, and any change will impact legacy content, even if it is in XML.

• Jumping into a total all-out conversion will overcome the "organizations still don't understand the technologies..." problem--False. It is best to test the assumptions of a new model on a pilot project. There will be unforeseen difficulties, technical glitches, and tweaks. Further, there is a learning curve in any such transformation. The pilot team will gain valuable knowledge and experience and can help provide support to other teams as the production-ready transformation is rolled out to subsequent projects.

• Separating content from format is always a good thing--False. Consider PC and peripheral vendors: almost all use graphical install posters. The layout and graphic design qualities of the poster itself are essential to its usability, with the goals being to reduce support calls and make the document accessible to multiple language markets with minimal translation. Though this content is part of the customer documentation set, there is no point in trying to re-use it with other content-system unpacking and setup is a one-time activity. So, this type of highly formatted, presentation sensitive content is best left in a silo, and is more readily produced in a DTP (Desktop Publishing) application than an XML editor. It could be argued that SVG is an XML format, or that XML can reference graphics. This is true, but does not buy anything for the added cost and difficulty in this instance. Real use in the real world is complex, and there are exceptions to every rule of thumb.

• Content is not viewed as a business asset worthy of effective management if it's not in XML-False. Here, effective management is defined recursively, like a catch-22. If a company does not view content as a business asset in the first place, it will be difficult at best to convince management to invest in an XML infrastructure.

XML is an important technology. Like the automobile, however, XML relies on other technologies to make its use feasible. Henry Ford invented the automobile assembly line, making them less expensive to own. A car is little use without gas stations and good roads, however. Similarly, XML is becoming more feasible and easier to adopt because of gateway technologies like DITA and S1000D. However, just as important, and not mentioned in the initial article, is the content management platform. The real breakthrough for content management in technical communications is coming from the development of component content management systems for technical communications environments like those that first appeared in the Web publishing arena ten years ago. Witness how many of the large document management vendors are recently acquiring technologies that enable them to support XML component content management, in conjunction with their Digital Asset Management, workflow automation, and other platform components.

Usually the larger the content set, the greater the potential payoff from converting to XML due to economy of scale. The greater the content volume, however, the stronger the case for a gradual approach: analyze content, audiences and contexts of use, adopt a strategy, test and refine it with a pilot project, continue the roll out as development schedules and budgets allow, training authors and managers along the way. A sound content management strategy and platform allows that migration to take place as a natural progression, and does not force the conversion of legacy content to XML just to gain the other advantages a CMS can provide.

DCLNews Editorial
March 2008