Think Info

Exploring the information space

Deblobbling in the Real World (CS Forum 2013)

What must have been – at least from my perspective – one of the two best talks at CS Forum 2013 was delivered by one of the most turned-on people I know – Jeff Eaton – tackling the beautifully thorny challenge of structured content.


Content is not just a tool for self-promotion. Content is an asset, with intrinsic value; for many, it is a core product.

In order to leverage the value of content, it needs to be eminently reusable; repurposable. This is not just that it can be shared socially. It needs to be sufficiently fluid that is can move easily between formats and platforms, remoulding itself to suit the situation. Content that is baked into a particular format is about as reusable as an ingredient in a served cake. The only solution is to store content in a structured format.

Of course, even a well-conceived structure is not infallible. To paraphrase Field Marshal Helmuth von Moltke, “No model survives contact with real content.”

The main problem is that most content modelling is derived from an existing presentational paradigm; the modelling process is skewed, prejudiced by existing behaviour and needs. This results in a model that is good for nothing but recreating the current output. To successfully model content, we need to move beyond the present, into the realm of potential: how might the content be repurposed? What metadata will it then need? How granular does it need to be? Are there alternate ways of looking at the underlying information that would allow it to be resurfaced in unpredictable new ways?

In order to manage and maintain a powerful structured content environment, the author experience needs to be optimised for governance of that content. The content management system must facilitate the management of the entire content lifecycle. There are several key rules that are the starting point in ensuring this:

  • The editor should never be exposed to the underlying technology.
  • Provide control of meaning, emphasis and association, not presentation (that is handled by rules and transformations).
  • Sequence-modelling needs to be available, to create narrative flow. This requires structure to be captured within the body of content.
  • Speak the language of your authors; use terminology appropriate for the content.

While in a perfect world, we would model every piece and type of content, there are limits imposed by the reality of limited resources. When faced with a one-off scenario, something utterly unique, it is rarely worth the effort to model it fully. But on the other hand, the tools for unique content should not be made too readily available – some people will have trouble containing their creative juices.

Raw notes:

  • Content isn’t only about self-promotion. For many, content itself is the product. It is the business core.
  • We want to be able to remix and repurpose our content. But it’s stuck in jello.
  • Social re-usable content is not the same as structured content.
  • No model survives contact with real content.
  • The largest failure of content modelling: design coupling – the structure is derived from a prior presentation format.
  • Your content structure is not only about the presentation. There’s also metadata. Loads, and it’s valid.
  • Modelling by designs is a jigsaw. Swap bits around, but there’s only one “correct” picture.
  • Compare your model to use cases. See how the model works in the real world.
  • A new type of content: the platypus: one-time, oddball content, that is very unique.
  • Modelling the platypus is not usually a good investment.
  • In a world with infinite time and infinite resources, we would model all content.
  • Exposing the technology underlying content presentation to content editors – to design – is bad.
  • Layout itself can be a type of content in its own right.
  • Don’t expose editors to the presentation of content. Allow them to define relative emphasis. Image? Pull-quote? Etc.
  • The hard part of structuring rich content: sequential asset linking to narrative flow.
  • If you are treating a big piece of content as a one-off (do it manually), it will become the dominant content type.
  • Html is terrible for capturing meaning in content
  • Transform. Content is structured. Have your associations. Make the system know how to render it.
  • Speak the language of your content. (And authors)
  • Model meaning, not mock-ups!
  • Narrative requires structure in the body of the content.


Jeff’s presentation slides on Slideshare


One response to “Deblobbling in the Real World (CS Forum 2013)

  1. Jeff Eaton 2013/09/21 at 18:07

    Thanks for the kind words, Rick! It’s really encouraging that others are thinking along the same lines with the structured content shift. The “narrative flow” problem that I touched on in the talk is (as we’ve discussed) one of the big challenges I want to chew on over the next year. There’s a LOT of work that’s gone on in the document/schema-oriented world (DITA and similar standards, for example) but so much of it is inaccessible to the web-heavy world that a lot of us work in.

    Glad you enjoyed the talk, and I hope there will more more interesting developments coming soon…

%d bloggers like this: