Having interviewed Keith Schengili-Roberts and Joann Hackos, we thought we would be running out of ideas by the third interview on DITA and localization. Talking to Julian Murfitt and Joe Pairman from Mekon brought a whole new perspective to the subject – probably because they are British and their culture as well as the accounts they have worked with influence their perception.

Julian Murfitt Mekon CEO

When Julian Murfitt founded Mekon in 1990, the documentation world was unstructured and used Framemaker and Word. It was labour-intensive, with much typesetting, document conversion and customer training. Today Mekon’s business has evolved to helping his customers generate intelligent technical content, content that contributes to the final purchasing decision because it is tailored, short and answers exactly their concerns. Their solutions integrate best-of-breed technologies such as XML, DITA and S1000D. Each intelligent content project starts with strategic planning and rolls out with software customization and customer coaching. Mekon’s market is a global niche, with staff and customers around the world.

Julian, what new trends have you seen in your intelligent content projects?

There are two concepts that our customers understand and have been bringing up in the last eighteen months: the first is taxonomy, which is the science and practice of classification. It is the key to semantic richness and searchability. The second one is directly correlated to it: metadata, i.e. data that provides information about other data. Five years ago the DITA adopters tended to overlook both of these and use only a small portion of DITA’s capabilities. They grafted new tools onto old processes, outputting print and PDF files. They may have been translating, but not localizing to suit local audiences. They were not thinking about delivering content in an agile way or for a particular device. Our recent projects have  involved understanding the content in a more meaningful way. What is this piece of information going to be used for? Is it a description or a general concept? Does it correspond to a single piece of equipment or does it belong to a system?  Does it refer to a person, a fault code? What is the information’s lifetime? This brings us to building a common terminology base, shared by all the company’s departments. Believe you me, it can get messy!

An additional concept is starting to emerge with some of our customers: truly localizing the content, which consists in translating and tailoring the content to fit the audience’s region. DITA enables this — and we are helping customers to plan and design the content so this works in practice. We have for example developed for one of our customers an automatic regional filter to his case studies, so that the potential customer can genuinely relate to the examples and apply them to his local situation.

You mention taxonomy. Can you give me an example of how you roll it out?

One of our customers, an industrial gas vacuum pump manufacturer whose products are built to order, has high requirements in terms of content management: internally they need to track the content and who has written it, how the topics relate to each other and which are the root topics, what name to give each topic. They need to tailor the content to its reader, whether it is a Service Technician, an Engineer or a Product Manager.

What drives your customers to implementing DITA?

Customers come to us with a DITA project to generate content faster, to increase content quality – for legal purposes for example – to reduce costs, or all of the above.

What is your greatest challenge with DITA projects?

It’s the people, getting them to embrace change, to understand and “happily” adopt the DITA concepts. Within a company’s technical writing team there are many casual DITA users, such as engineers who contribute their expertise on a specific product. They use DITA occasionally and aren’t familiar enough with it to be productive. In some cases, these users are actually better off creating content in Word, which is then automatically transformed to DITA for a core team to edit later. This is not the sales pitch given by vendors, who sell the DITA system as being enterprise-wide, a common denominator among all technical writers, whatever their level of contribution.

How large are the companies coming to Mekon to implement DITA?

It’s not how big the company is. It’s not about the sales turnover or the volume of staff. A good indicator is the number of professional technical writers, which should be of three at the least. The ideal number is five, because John, Bill and Jane need to know who is writing what in order to work in unison and avoid content duplication. If the company doesn’t employ any technical writer and relies on hundreds of engineers to write about their product or technology, DITA may not meet their expectations, because their level of pain and interest isn’t sufficient. You do nonetheless run into exceptions with one single technical writer in a start-up; he is a real DITA evangelist and he will make it happen.

Which industries are most keen on adopting DITA?

I’d like to come back on what Keith Schengili-Roberts said in your first interview and how software companies were the best fit for the DITA processes. From my perspective there are millions and millions of small software companies out there, with one to no technical writers; they are potentially the worst customer. All they need is online help for their software and Madcap and Robohelp do a perfectly good job. Their requirements are simple: they want it context-sensitive; they don’t print it and deliver it within the control of their own software. In this scenario, they are doing a perfectly good job managing their technical content, even though it is not structured or XML-based. The DITA software trend-setters, such as IBM, HP and SAP are large corporations who have well outgrown basic online help tools and whose software solutions are so large and complex that they need a whole new paradigm of infrastructure to support them.

Other than that, we serve the semi-conductor and medical device industries, as well as manufacturers of modular products with hundreds of possible combinations, like the vacuum pump manufacturer we were talking about earlier. Institutes and professional organizations, whose business model is to provide information or training material at a premium, are looking for innovative ways to deliver information to justify their existence, in a world where a lot of information is free and only a click away.

Another of Mekon’s specialties is dynamic publishing. What is it?

Dynamic publishingIt is publishing content that can guide a user, based on the intentions of the originator. Content can be adapted on-the-fly, according to the reader’s profile, context and device. It requires to be fed metadata, which can be extracted from the CRM. The connections between the CRM’s user profiles and the content are established during the taxonomy set-up. In the case of an institute for example, you can adapt a report by selecting case studies that are relevant to the trainee’s background, such as his geographical location or his field of expertise. Dynamic publishing is commonplace on B to C e-commerce sites: if you want to buy a car, you will naturally zoom in on BMW/convertible/red to browse the six available models, you won’t scan through the hundreds of unsorted pages. B to B websites are lagging behind. Visitors at work must struggle with monolithic PDFs and catch-all homepages. They want to bring professional websites up to speed, turning them into viewer-specific portals.

What are the benefits of dynamic publishing?

The number one benefit, one of today’s buzz words, is customer experience: delivering the information the person needs to fulfill a task. The second one is increasing the richness of the information. Today cross-references are the preferred method to guide a reader in his navigation (Wikipedia hyperlinks within each page are a good example). They are done manually, which means they are time-consuming, subjective and linked to content that already exists. Dynamic publishing gives the ability to use the metadata words and link to information that doesn’t even exist yet.  We can feed readers with a regular stream of current and future documents, based on their keywords and related topics, exactly like Amazon, which suggests related products based on your search.

Can you give us an example of a dynamic publishing project?

Altera is a US chip manufacturer. In the chip industry, the engineers like to read the technical documentation both in their language and in English, the source language. The documents were published in PDF format, which meant the exercise was technically challenging: the engineer would for instance want to find page 310 of the Japanese manual in the English one. Dynamic publishing has enabled Altera to switch languages of their technical documentation seamlessly. We are also working with a train manufacturer, who wants to publish content dynamically for its design teams and make the information easier to find, comment and upload to the rest of the system. We integrated a Wiki style system using DITA that allows the engineers to add fault codes and diagnostics on-the-fly during the prototype phase, thus building up the knowledge base.

How do you tie localization into your projects?

It’s vital to integrate localization into the DITA or dynamic publishing project, whether it is executed internally or with an LSP, even if you don’t implement everything at once. It’s not something you can add on at the end. Your language experts must be in the loop when designing the new content process: how is the translated content injected back in your system? Does the LSP understand the content snippets? For example you may  have a variable corresponding to the customer name. The translator will enter the customer name and be confused because it is not the name they are actually seeing in the preview. Knowing XML doesn’t mean you’re proficient in DITA.

Do you see differences in how companies value and create technical content around the globe?

German companies are clearly best-in-class in terms of technical publication. They have a higher regard for a piece of equipment and the information that supports it. Technical writers occupy a strategic position, and their expectations and qualifications (some go as high as PhDs) reflect this. In the United States translation is more often than not set aside because the domestic market is sufficient, so the cost reduction induced by localization is not a selling point when promoting structured content. This brings us to put forward other structured content benefits: semantic content, searchability and richness for example. In Asia, with the exception of Japan, a common attitude among manufacturers is that technical documentation is a mere box to tick and is disconnected from the purchasing decision. Products in China and to a certain degree Taiwan are commodities, the focus is still on industrialization and little effort is put into branding and customer experience. DITA is not widespread in Asia yet, so there are opportunities to seize in the mid term.