Last month, in Part 1 of this exclusive DCL interview, Ann Rockley shared current trends in companies moving to XML and the importance of content analysis. This month we continue the interview with Senior Consultant, Steve Manning, discussing content reuse, tools to convert content to DITA, component content management, and dynamic personalized content delivery. Then Ann joins us again to discuss an upcoming CMS Watch report she is co-authoring.

DCL: Steve, in organizations that are converting legacy documentation to a structured format what kind of reuse are you seeing, both before and after conversion?

Steve Manning: We have found that companies have, at minimum, about 35% of their content duplicated in different places. This means that if they had the ability to reuse it, they could reduce the volume of their content by about 35%. We've estimated the amount of duplicate (or redundant) content to be as high as 70% in some cases.

The actual amount of reuse ranges from company-to-company based on the kinds of information being created. For example those companies that have documentation that supports multiple related product lines tend to have a lot of reuse because Product A1 may be the basic unit, Product A2 has a few different features, and Product A3 might have a larger screen or some other feature. In those instances you find an enormous amount of reusable content because a lot of the core information is the same. In other instances, where you might be looking at a single product but different kinds of documentation such as technical manuals versus marketing materials, the percentage can be less.

As companies move through the content analysis Ann mentioned earlier, they get a good picture of their content and begin to understand how they can reuse it. When an organization starts producing a lot of documents and has the same piece of content in different places, they get a lot of benefit from maintaining that piece of content in one place. Create it once, and reuse it in different places. Translate it once, and reuse the translated piece in different places.

DCL: What kind of tools are your clients using to identify and restructure their reusable information and convert legacy documentation to DITA?

SM: Companies seem to be using a combination of things. We work with companies who use conversion services to help them convert large batches of content. Some companies use individual conversion software utilities that import content into an authoring tool. Then they modify the content to fit the new structure. And some companies basically start from scratch. So they might have an old version in a Word file and cut and paste into an XML editor to move the content forward. That's especially true when they have a lot of restructuring to do. It depends on how much of a transition they are making. If they are moving from a very loose type of structure in their legacy documents to a tightly controlled structure, it's typically going to involve the help of human editors.

If you have 10,000 pages of content, then you have to look at what an automated conversion will buy you. So the issue of conversion becomes a "buy versus build" decision. The question is, 'How far will it take you?' In companies that have a large amount of documentation to convert, sometimes it's not just about reuse. It can be about getting it into the new structure so they can work faster.

Hiring a firm like Data Conversion Laboratory to build some intelligence into the conversion process may get you 70% of the way. That's going to end up being cheaper than having people do 100% of the conversion work by hand.

DCL: Are your clients able to convert straight to DITA or is a certain amount of restructuring of the content necessary first?

SM: Many companies, because they are moving from paper-based content, must rewrite or restructure some of their content to make it ready to be reused. Sometimes they use conversion services for this, sometimes they use other tools. I haven't seen a way to do 100% conversion from legacy documentation into structured documentation and have it fit right away. Some type of manual restructuring has to be done.

DCL: Why do you think DITA has become such an important standard?

SM: DITA is very popular for different reasons. One of those reasons is that people know they are going to get benefits from XML, structured content, separation of content from format, and the ability to mix and match components into different outputs. And let's not forget the benefits provided by being able to automate processing of content into different outputs. These are the business drivers that are pointing people to XML. When you start thinking about moving to XML, you start looking at specific applications for XML. Starting from scratch is expensive, so DITA provides an immediate starting point which can be useable out-of-the-box in some circumstances, but has been designed to also support customization (or specialization).

DITA is topic-based, which gives you component-level content reuse and content management. It saves information as topics and associates those topics using DITA maps to generate a specific output whether it's a PDF file or an HTML page. It allows you to reuse lower-level content so that an individual section could be reused in multiple topics. It supports all those things that you need for reuse. It's XML and it provides style sheets, which is the mechanism for moving from XML to actually generate the output that you need. It's a fairly complete package.

DCL: DITA is getting a lot of attention these days. But are there any other competing standards that people should also be considering?

SM: I think DITA is here to stay. One of the reasons is because it comes prepackaged with a better approach, which is component-based, or topic-based, authoring. For several years a few software vendors have offered tools that provide users with the ability to create small chunks (topics) of content. But because DITA comes in a package, it gives you a leap ahead. It pushes you further down the implementation path than starting from scratch creating your own XML tags.

However the question shouldn't always be, 'Can I use DITA?' but 'Should I use DITA and what would I need to do with it to make it effective for me and my company?' If you are a multinational company with 200 users all over the world, you might need more control over the structure of your content than DITA out-of-the-box provides. But if you are a four-writer shop, DITA out of the box may be all you'll ever need. The answer for most companies is probably somewhere in between.

DCL: We talk a lot about moving legacy documentation to structured formats like DITA for reuse, but is there another goal that goes beyond just reuse?

SM: One of the things that makes DITA effective is that you are not embedding content in any way. You are just referring to it. That makes it easy to create variations. So you can take a DITA map and easily create a new DITA map pointed to the same topics in a different order to give you a different kind of output. A lot of companies are moving away from creating static content, which leaves it up to the end user to get out of it what they need. Instead, they are moving toward a different kind of content delivery model - an on-demand model--whereby content is delivered dynamically to the user when they request it. The on-demand model also allows us to provide the right information to the user, personalized to his or her needs based on information we know about them. For example, the government client we mentioned in Part I has 8 or 9 different defined user types. Those users are going to approach content in different ways. This client is moving to a component management system that gives them the opportunity to adjust the content for these different user types-personalize it. That's what we mean by dynamic (or on-demand) personalized content delivery.

DCL: Why is the concept of dynamic personalized content so important?

SM: Companies are changing their focus toward the user experience for different reasons. Primarily it's based on usability. Documentation is an important part of that. If people can't use the tool, no matter how effective it is, they will declare it ineffective. Companies realize they have to provide, not only information, but the information users want and need. To entice sales, they want to show people what their product does. They also want to make information more effective so they can cut down on support calls.

When you approach users from a dynamic perspective, you must be able to interpret where they are in your process. If they are novice users you can give them more details. As they become more experienced you can provide just the information they need. Companies are also trying to build persuasive smarts into their web sites. This occurs when users go to a site, and based on what they click and certain logic assumptions, what they see changes.

For example, if you go to site for a vacation resort and you click on links related to activities for kids, it makes the assumption that you are a family. Displayed next would be activities directed toward families. Companies are putting an enormous amount of effort into this. It's not intrusive. It doesn't ask you for personal information. It just makes certain assumptions based on the link choices you make. Then they may send you an email later that says 'Hey, you signed up for this activity. You might like this other one too.' That's dynamic personalized content.

DCL: You mentioned earlier using DITA to better manage content components. How is that different from content management?

SM: We are moving away from being document oriented-or page oriented-which is the management support that has been provided in the past by content management systems. We now look at content as building blocks that we can use to assemble and reassemble different outputs to meet customer needs. Those building blocks, or components, bring different challenges to content management.

When we were doing document-oriented management we just needed to check documents in and out of the system, make sure there was an audit trail, and save different versions. But with components you need to be able to build relationships between components and identify where an individual component is used and reused. Or you might need to manage conditions and associated components into an output in a certain hierarchy or sequence. There are things that component content management systems need to do that traditional content management systems don't do.

DCL: Ann, I see you are working with CMS Watch on a new report evaluating Content Component Management technology. What kind of technology are you talking about?

Ann Rockley: In the past CMS Watch has looked at Web Content Management Systems, which are geared toward web sites. They've also done a report on Enterprise CMS, which is a broader management of documents, records management, and things such as email support. But those have been more about managing from the perspective of output than managing the authoring process.

This will be the first report that focuses on component content management. It focuses on the creation side. Instead of authoring a book, users are authoring individual pieces of content. These are the components we create . . . this is how we build links between these things… this is how it flows through the review process and the translation process… this is how it flows through the content management system itself.

We will be looking at a specific set of tools and will be updating the report on a regular basis. There are an enormous number of content management systems, but this report will help people focus on the type of system they might need to do DITA or component-based content management.

DCL: When is that report due to be released?

AR: Summer of 2007.

DCL: Finally, can you give us a clear picture of what it really takes to start, and get through, a project of moving to DITA.

AR: First, it all builds from the analysis because we are talking about answering the question, 'Is DITA the solution?' or 'Is XML the solution?' So we ask our clients, 'A solution to what?' It really begins with a clear understanding of 'What hurts? What is the problem, and what do you need to do?' Because that's going to take you forward in terms of how you proceed.

For example, we have worked with publishing companies, and they don't have a reuse need, but they need other things such as a consistent structure so their journals have a level of quality and completeness. So moving to structured authoring is great for them, but they probably wouldn't need to break content into components like DITA. But for them XML is a great solution for automating multi-channeled publications such as print and on-line PDF or HTML versions.

It's also important to understand the life cycle of your content and see if there are any problems in workflow, writing, design work, or translation. Then look for opportunities to eliminate redundant content and to break the content up into components that can be quickly reconfigured into new content for a different user.

Then to get through the process, companies are usually motivated by need. Having a clear understanding of the whole picture really helps you get your business case in line and helps you work out the expenses. That financial argument is a very persuasive one. We recently worked with a company who saved 60-70% on translation costs, while tripling the number of languages they used. For other companies, who don't have the translation issues, conversion has simply made their life-cycle processes so much easier.

DCL: Those are all compelling reasons for companies to start looking at managing their content and content components in a better way. Thank you both for taking time to discuss this. We'll be looking forward to the release of the new CMS Watch report.

The Rockley Group Inc. has an international reputation for developing customer-centric enterprise content management strategies and underlying information architecture. They have helped organizations develop structured content reuse and unified content strategies across departments, divisions, and the enterprise. Working with customers in the Life Sciences (Eli Lilly, Guidant, Johnson & Johnson, Medtronic, Wyeth), Financial (Citibank, Bank of Canada, Deloitte & Touche, PWC) and High Technology (Business Objects, Cisco, HP, Lexmark, Symantec) industries, The Rockley Group has developed structured content reuse solutions to reduce the cost and effort of complex information creation and management.

DCLNews Editorial
July 2007