Single Source Publishing Part 3 : Is Automation the Answer?
Obviously it would be awesome if we could push a button in our imaginary publishing system and all the outputs would roll out automagically – fully formed, perfect and beautiful.
Is such automation possible? Let’s have a closer look at the types of conversion involved.
There are two types (roughly) of structural conversion – upconversion, and downconversion.
To create a simpler structure (eg plain text) from any source format, the transformation mostly means removing information (structure).
Removing structure is easy (usually), we just throw stuff away.
To get from our source file to a more complex structure, we need to add information (structure).
Adding structure is doable but not as easy.
A simple diagram to help us understand up and down conversion would be as follows:
Down-conversion is generally an easier target for automation. Up-conversion is more likely to require human intervention.
Another category of conversion is typesetting. Typesetting means we require a change in the look and feel of the output (design).
There are two approaches to typesetting – ‘automatic typesetting’ (which is done by a machine) and manual typesetting (done by a human).
Automatic typesetting is when a designer sets up a bunch of rules (templates) and then ‘a machine’ applies those rules to the content.
Automatic typesetting is possible to achieve in some Publishing contexts but the more complex your design requirements, the harder it is to achieve.
When a designer uses a design tool such as InDesign to produce the required design, this is manual typesetting.
Again, the more sophisticated the design, the more likely it is to require human intervention to produce the desired result.
Where does that leave us?
If we can achieve all of our conversions – structural conversions and typesetting – automatically, then we are in a winning situation. We can start with a source file and then press a button and all of our outputs will be generated automagically. This is single source publishing achieved through automation.
Automatic conversion is most likely to be achievable where we have either:
- very simple conversions
- moderate expectations from our conversions.
While often true in ‘web publishing’, these two conditions are seldom true for publishers. Publishers, generally speaking, have high expectations for all their conversions and their outputs are often complex.
What does this mean? In most publishing contexts, this level of automation is not achievable – we will require manual processing, and manual processing inevitably leads to the broken workflows we were trying to avoid. If, for example, we want to introduce a manual design tool such as InDesign, then we are changing the people, tools, and source as discussed in Part 1 – and we are deciding against a single source workflow.
If we are to achieve single source publishing, but we can’t do it by automation alone, then we need another strategy.
To achieve all the efficiencies of the single source workflow that we have discussed before (see Part 1), we need to work out how humans (content producers, designers, production people) and machines can utilize different tools, respecting their tool-cultures, while working on the same source to create all the desired outputs. How is this possible?
More to come…
Part 4 now available here. Thanks to Henrik van Leeuwen for the images and Raewyn Whyte for copy editing.