Bitesize DITA: Relationship tables – what are they good for?

During a DITA training course, relationship tables seem to be a real sticking point for many students. The process of setting them up seems almost deliberately clunky and eccentric.

A novice DITA author’s mental model tends, quite reasonably, to equate the DITAmap with a publication’s TOC, so placing a table in it just seems a strange thing to do. Reltables themselves can be hard for the human eye to ‘parse’ – without careful reading it can be hard to look at a relationship table and understand what is going on.

As I take students through the process of creating the relationship table, processing the map and viewing the different results, there are usually questions like: ‘I can see what’s happening here – but why are these useful?’ or ‘Why not use the related links element?’

Relationship tables are the way they are for several reasons:

  • They act as a central point to manage relationships between topics in map. Depending on your content model, a publication could consist of more than a hundred individual topics. If something changes, do you have the time and the energy to open every topic and check that the links are correct?
  • They promote reuse by abstracting linking information away from the topic itself. If you build the links into individual topics, those links may not be universally relevant when you reuse those topics in other maps.
  • They enable you to do more sophisticated processing of links – for example by grouping links into families or by setting up bidirectional or unidirectional links between topics.

If you remain unconvinced that relationship tables are worth the effort, don’t worry. New ways of delivering DITA content mean that there are new ways of making relationships between topics. Clever use of metadata can create meaningful links between publications at publishing time – meaning the topic itself stays context-free and reusable, while the user is presented with appropriate links. Another advantage of creating metadata-based relationships between topics is that by using various kinds of metadata, you can offer personalized links: different links depending on, for example, the reader’s job title, level of expertise or locale. This approach requires more than just the DITA Open Toolkit, of course, but with a dynamic delivery platform such as Mekon DITAweb, you can tailor web-based content — and the links between topics — to your users’ specific needs.