Headless DITA is real and relevant

These technical specialisms and their tribes each cluster around certain real-world uses. For example, some technologies are good for banking applications; others for content management. It gets more niche, though – content management has its own subgroups with their preferred tech. A recent one is “headless delivery”.

Some years ago, large news providers realized that they wanted to give readers different experiences of the same basic content, the same articles. In some contexts such as a scrolling mobile feed, just a snippet and picture was needed to tease readers to read more. On a desktop site, a whole article could be good, with maybe videos in a sidebar. Elsewhere, maybe just the video and the article title would be best. For a content management system to deliver these different experiences, it wouldn’t be enough to just push one whole article everywhere. Rather, a flipped model was needed – one where the actual user apps and sites took the lead, requesting just the needed snippets from a CMS without its own presentation layer. A headless CMS.

 

As a marketing concept, that didn’t exist at the time. Some CMSs (such as the original Tridion) were already “decoupled”, with headless capabilities, but they didn’t shout about it. The big news broadcasters built their own headless systems to manage content, though at great expense over time. But other publishers and marketers liked what they saw, and a new breed of tool emerged to do just headless delivery. Well, the delivery part but also the data entry part. These simple CMSs did provide forms to collect content in the snippets needed for querying and later recombining in various output channels. Thus, “structured content”.

 

If you come from the technical content or publishing worlds, you’ll recognize that that basic kind of structured content lacks the possibilities to capture rich semantic documents of standards like DITA. The reliable ways to repurpose, relate, and govern critical content that DITA supports. Coming from this world, we understand what’s really needed to cope with the challenges of scaling and automating data while ensuring that authors give users the context they need. 

 

Some folk from headless tool vendors misunderstand what’s needed in our world. For example, they see a requirement for tables, and they think “we can do that”. They dismiss a requirement for PDF as “there are libraries for that”. They hear about content reuse and rich semantic relationships between documents, and think “it’s all just code”. Theoretical possibilities, that in practice are expensive and unmanageable — but a hammer can be machined to become a screwdriver, I suppose. 

 

To be fair, on the structured side, us publishing folk have sometimes dismissed headless too quickly. “Doesn’t DITA give you multichannel output”? Well, yes, but is that massively scalable, queryable on demand and in a structure that modern web developers will instantly understand? Don’t try to hammer nails with a screwdriver. Painful. 

 

Wouldn’t it be great if there were a well-governed, proper publishing system, intuitive to experts in technical domains like manufacturing, life sciences, and finance, that also had scalable, powerful headless delivery built-in? Not just delivery bolted on, glitching when you unpublish or edit content. Not just light integrations, but a real end-to-end publishing system, designed as one?

With Tridion, we actually do, and it’s the best-kept secret in content management. (Maybe we need to do some of that headless marketing I keep seeing). Tridion Docs with DXD does proper headless delivery with no need to “ditch DITA”. It’s not some dated system with a passing nod to APIs, but a serious web-scale solution that serves rich, well-governed structured content in formats that feel natural to frontend developers. It’s proven in practice with some of the biggest and most complex multi-regional content delivery environments there are. Yet it’s affordable and easy to get started with. Our upcoming release ensures that web developers need to write even less code to target exactly the content they need. (And we keep innovating — our next release recommends relevant content using semantic AI so you don’t even need user data to start getting useful results.)

 

Sometimes you need a matching toolset, not just a hammer.