Our series on DITA and localization continues, after our interview of Keith Schengili-Roberts at the beginning of July. This week we speak to another expert and key influencer in the DITA community, JoAnn Hackos.


Dr.  JoAnn Hackos is a true visionary. The founder of Comtech Services, she started talking about documentation usability in the eighties, when the thickness of software documentation rivaled with phone directories. In the early nineties she took part in introducing new concepts for technical writing: minimalism, structured authoring and single sourcing. The DITA standard, which she instituted alongside IBM, Arbortext, Nokia, and others, brought these concepts together and was released in 2004. Dr. Hackos’ daily routine consists of helping corporations define a strategy for their content management (including DITA), their information architecture and product interface design.

JoAnn, within your client base, who is implementing DITA today?

Many of our clients come from the telecommunications industry with references such as Ericsson, Huawei and Nokia. They understood early on the value of structured content and have been eager to adopt the DITA model. The big software developers came next, because the concept is easy to implement company-wide. Today additional sectors are migrating on a large scale: off-road machinery manufacturers, like Caterpillar and AGCO, as well as medical device companies, because of content reuse, lower translation costs and more efficient regulatory compliance.

Can you give us some examples of the achievements that come with DITA?

One of our medical device clients opted for DITA to reduce translation costs. Step one was to implement minimalism and publish only what the medical staff needed to run the equipment. They were thus able to reduce the size of their documents by 50%. At the same time they moved all of the content to DITA, which meant it could be reused in other equipment configurations. They focused on the caution and warning instructions, structuring and standardizing the content. The ROI was immediate: the very first translation of the new document earned back the entire cost of the project.

For another client, we migrated the employee handbook to DITA. The HR department used to produce eight different Word documents to cover all the employee profiles and levels of responsibility. The slightest update, like a change in the smoking regulation, took six months. Now all the employee manuals can be generated in six minutes.

150717_content management Hackos interview_v2

Can DITA become a universal standard?

It’s gaining momentum. Today at least 5% of companies appear to be using DITA. Others have opted for a proprietary or in-house solution or what seems to be on everyone’s desktop, like MSWord. DITA has a major advantage: there is a community of experts and users backing it and working on improving it, rather than just a single organization. However companies who have invested in a proprietary solution argue that it is too hard to change. Some companies have organized their content so badly that the short-term cost of fixing it would be prohibitive.

How do you help clients implement DITA?

DITA has been designed to serve a wide range of subject areas and industries and provides many options in its content model to support those industries. Of course, that means that organizations do not need the hundreds of elements that DITA supports. Through our information modeling process and training, we help an organization select the parts of DITA that it needs, thus simplifying the authoring process for everyone. We help groups create templates that guide authors through the development of a concept, a task or a topic type, usually using a simplified authoring tool that resembles writing in MSWord more than working with XML elements. To optimize the DITA authoring environment, we help organizations develop valid DITA constraints and remove the several removable parts of the Document Type Definitions (DTDs) by creating shell models. In some cases, we develop specializations that introduce elements and structures needed for special purposes. Finally, we use the decisions documented in the DITA-based Information Model to help authors make good decisions about content and avoid mistakes. Using Schematron, we might define for instance the number of words in a short description or make certain that a numbered list has more than one number.

How much time should you plan to set up DITA in the first place?

Adopting DITA is a bit like changing your ERP system. With a project roadmap, human resources and assistance in applying best practices, a DITA project can take six months. We’ve seen companies take two years because they don’t have the resources to tackle the new process while still maintaining the old material during the transition phase; we recommend that clients bring in temporary help to maintain the old content while the experienced people work on the conversions.

What lessons have you learnt from combining localization and DITA?

Do not take your LSP’s DITA skills for granted! One client was already working with a localization vendor, so we invited them to the DITA project set-up meeting. The company did not pay attention to the DITA process. They were supposed to present their first localized sample but came without it. We were expecting the usual 50% break in localization costs because of the removal of desktop publishing; instead they imposed extra fees to maintain their revenue levels. They had not understood that they would be getting much more translation work as a result of the drop in prices. They ended up losing the contract.

With another account, the LSP measured only a 55% match with the translation memory (TM) after the conversion of the source to DITA. I had a feeling that the match was not correct. I sent the file to a colleague who was a DITA and translation expert, and he came back with a 95% match. In addition the LSP did not understand how to set up their system to accommodate the DITA elements correctly. On the basis of the 55%, they quoted 400% more for the translation. A lot of LSPs say they understand DITA, when actually they don’t.

Which DITA-related trends are emerging in technical writing?

More information is minimalist; answers to questions are accessible through online search. On the client side, we strive to spread the reuse and structured content messages to all the departments of a company. We are currently carrying out a study for a client who would like Product Managers as well as Service, Training and Documentation teams to create content that they can all share. So far each team has been doing their own thing, outsourcing to different LSPs.

We are setting up Easy Authoring environments that enable authors to write topics without being XML literate: we simply hide the XML. We had the opportunity to set up Easy Authoring at one company. The sales people were tired of bringing the hundred-odd pages product brochure literature to their potential client. They wanted to create custom literature that pulled out only the relevant products. We set up a system whereby the salespeople themselves, without the intervention of the documentation team, can access Sharepoint, check off the products they want and generate the document they will take the following week to their customer. The system gathers the information in a pre-formatted brochure, with parameters such as layout and language. The result is truly fantastic!