Content Modeling to Assess Standards
The Rockley Group, Inc.
With all the standards out there, can you just pick one and start authoring? Well, it depends. This article describes content modeling to assess standards, focusing on determining expectations and ensuring the standard can meet those expectations
As XML has increased in popularity, so has the number of standards created to support content. As organizations move towards implementing a content management solution to help them author, reuse, and manage content, technical publications and IT departments are struggling to learn about standards, wondering whether to use an existing one, or to create their own. And, the learning curve is steep-there's DITA (the Darwin Information Typing Architecture), DocBook, SPL (Structured Product Labeling, used in the pharmaceutical industry), and SCORM (Shareable Content Object Reference Model, used in the learning industry), just to name a few. But, how do you know if adopting an existing standard will support your needs? And, why bother creating your own when there are standards out there for you to follow. Like most things in technical communication, the answer is, frustratingly, "it depends". In this case, it depends on the nature of your content, and on what you want to do with it (e.g., how do you want to structure it and reuse it), both of which can be determined through content modeling.
Standards exist for many reasons, many of which are described in Ralph Robinson's article on International Standards and Sarah Porter's article on the SVG standard (both articles appear in this issue of The Rockley Report). Standards exist so that "things" (CDs, graphics, appliances, aircraft) can be built and used in the same way, over and over again. Following a standard makes creating a product a repeatable process regardless of who is creating it, and it makes using a product transparent to users, regardless of who made it. Following an authoring standard makes the structure (and sometimes content, if your standard includes reusable content) consistent from information product to information product, regardless of which author created it, and it makes using the information product easier for users. When users have documentation (or an interface) created according to a standard, they know what to expect, in what order. Standards set expectations. But, because standards set expectations, you have to know that the standard you follow will accommodate those expectations, from both the creation and use perspectives.
One of the most important activities in developing your content management strategy is understanding what content you need to deliver to meet your products' and users' needs. You need to understand how the content is created, how it is delivered, how it is used and by whom. We refer to this as gaining an "intimate understanding" of your content as well as the processes to create, review, deliver, and otherwise "manage" it. You gain this understanding by conducting a content audit as described in our book, Managing Enterprise Content: A Unified Content Strategy.  When you do an audit of your content, you start to determine opportunities to reuse both content and structures, which you then formalize in a content model. During the content audit, you determine expectations (what information is required to support your information products and your users) and in content modeling, you document your decisions. Content modeling takes your findings from the audit further; you identify which information products you need to create and the information (elements) they will contain. You determine how each element is constructed (down to the section, paragraph, and potentially, even the sentence level), and you determine if an element takes reusable content. You identify the metadata that describe your elements, and you create writing guidelines that tell authors how content should be written for optimum reuse and usability. You determine how elements are stored, how they will be shared, how they will travel through your workflow, and how they will be compiled into your various outputs. Thus, content models become the road map for your content management strategy.
The process of content modeling involves identifying all the information requirements for a particular project or department (sometimes, even for the entire organization) and determining how all of that information is put together. As such, the content modeling process forces you to consider all information requirements and to assess what information is available to meet those requirements. The content model becomes the "catalog" of all information products produced within your organization (or within the scope of your project) and outlines the necessary elements for each of them. Content models are the formal statement of your expectations-they state "this is what you expect the information product to contain and this is how you expect it to be structured." Creating information products based on content models ensures their consistency and ensures the process of creating them is repeatable, regardless of the author.
Why not just select a standard and start writing?
Content modeling is an important first step to selecting a model because the standard you choose may not fit your content, and it may not support your needs for reuse, or for collaboration. The content audit and modeling exercise will give you a complete understanding of your content, from both the creation and delivery perspectives, and allow you to fully assess the impact of implementing an authoring standard in your organization or department. The content model also helps you to fully understand and define your opportunities for reuse, both of content and of structure. (For a description of the distinction between reusable content and reusable structure, refer to "Issues in Information Modeling" in the June 2004 issue of
The Rockley Report.)  In short, content modeling will help you to assess whether a standard (such as DocBook) meets your needs and how you may need to extend the standard. For example, DocBook is a book-based, software documentation model that has many built-in models to cover many typical software documentation applications. If you produce software documentation, it's worth looking at because it can be implemented pretty much "out of the box". But, you need to have something to assess it against. This is where content models come in. Just as review criteria provide editors with guidelines to review against, models provide information architects guidelines to assess standards against.
Content modeling is also critical if you plan to implement DITA. DITA, as described in our recent whitepaper on "The Role of Standards in Content Management", is a powerful model that focuses on reuse with a topic-based core.  However, a common misconception is that DITA defines everything you want in your models. The DITA DTD defines only base models and its developers expect that you will create your own topic types to accommodate your own information needs. You may find that the standard DITA offering is sufficient for your needs or you may find that you need to extend it. But you won't know this until you model your content to determine your content types. The Introduction to the Darwin Information Typing Architecture states that:
The basis of the architecture is the topic structure, from which concept, task, and reference structures are specialized. Extensibility to other typed topics is possible through further specialization .... As a notable feature of this architecture, communities can define or extend additional information types that represent their own data. Examples of such content include product support information, programming message descriptions, and GUI definitions. 
So, while DITA provides general topic structures such as concept, task, and reference, you can further specialize by identifying which concept, which task, and which reference. This is what your content models will specify, semantically.
There are numerous content standards and defined structures available for you to use. (You can find many of them at http://xml.coverpages.org/xmlApplications.htm.) However, content modeling is-and will continue to be-a critical activity in content management implementations. It allows you to specify your own structure so you can determine if an existing standard will accommodate your needs. Your models are your blueprint, specific to your information products, and to your users. They are the formalization of your expectations for your information products. Standards are useful, but applied too broadly, they can become limiting, not allowing you to create information products that support your products and users. While some standards should be universal, some standards will always remain specific to your organization and to your users. Whatever standard you choose must allow you to meet those expectations.
 Rockley, Ann, Pamela Kostur and Steve Manning. 2003. Managing Enterprise Content: A Unified Content Strategy. Indianapolis, IN: New Riders
 Kostur, Pamela. June 2004. I
ssues in Information Modeling. www.rockleyreport.com.
 Manning, Steve. February 15, 2005.
Content Standards and Content Management Whitepaper http://www.rockley.com/whitepap.htm
 Day, Don R., Michael Priestly, and David A. Schell. March 1, 2001, updated June 24, 2003.
Introduction to the Darwin Information Typing Architecture. http://www-106.ibm.com/developerworks/library/x-dita1
 Examples of standards, http://xml.coverpages.org/xmlApplications.html