Last month I wrote that Every Page is Page One Topics are Everywhere, and I listed the principle characteristics of such topic as I see them. I said I would write in more depth about each of these characteristics. This is the first of those posts, on the standalone property of EPPO topics.
The word standalone is open to many interpretations. I have written before on what standalone means from the point of view of an authoring system and made a distinction between a chunk of content being able to exist alone (inside a CMS, for instance) and it being able to function alone for the reader. Here I want to look at what it means for a topic to function alone.
My purpose here is not to prescribe some new form of information design, but rather to explore the properties of the millions of EPPO topics that already exist in the world and are being written every day.
So what does it mean to say that millions of technical articles, blog postings, recipes, encyclopedia articles, used car reviews, etc. are standalone. Firstly, and most basically, it means that they are individually publishable. It means that readers are willing to accept them as discrete entities and regard them as useful and complete.
This does not mean that the reader will always find all they seek in a single topic. Clearly they will not. Sometimes they will blame the topic when it does not meet all their needs, but if it is a good topic, they will usually credit it with answering one need, and will then go on to seek other topics to answer other needs.
Suppose the reader is shopping for a used car. A review of a 2003-2008 Subaru Forrester will clearly meet a need for them. It will help them decide if a 2003-2008 Forrester is a car they should consider. But it won’t fulfill all their needs. They will want to read other reviews of other vehicles before choosing the one to buy. A fair and honest reader will judge the Forrester review complete if it covers all the salient points of the Forrester, and incomplete if it does not. They will not blame the Forrester review for not containing all the salient information on all the other cars on their list.
Here’s one way to look at it: there is a difference between a topic having a dependency and a reader having a dependency. Fair readers will blame a topic for having unmet dependencies. They will not blame the topic for their own dependencies.
Reader’s dependencies can be of various kinds. In the case of the used car shopper, their dependency is their particular list of car requirements. If the reader is looking only for vehicles with permanent all-wheel drive and manual transmissions, no single car review can be expected to anticipate that particular dependency. A comparison article on small SUVs, for instance, would be of little use to this reader since most small SUVs don’t match those criteria.
Another reader dependency is knowledge of a field. In the case of a recipe, for instance a particular reader may not know how to boil water. If a recipe says to boil potatoes for 12 minutes, for instance, it is not the fault of the recipe that the reader does not know how to boil water. Individual recipes are not responsible for teaching people basic cooking terms, or telling them how basic kitchen appliances work. If the reader does not know how to work their food processor, for instance, that is a reader dependency, not a dependency of the recipe topic. On the other hand, if it fails to tell you to melt the butter before adding it to the mix, that is a dependency in the recipe itself, not in the reader.
So, when we say that a topic is standalone, we mean that it is free of topic dependencies. We do not mean, and cannot reasonably demand, that it is free of reader dependencies. Most readers will have dependencies. To meet those dependencies, the reader is sometimes going to need to consult other topics.
Possible reader dependencies include: this is not the right topic for their task, they don’t have the requisite understanding to read this task, they don’t have all the components needed to perform this task (and must perform other tasks to obtain them), they don’t have all the specific data required to complete the task (and must consult reference information to obtain that data).
To be considered standalone, therefore, a topic needs to meet the reader’s reasonable expectation of a topic of this sort. It does not have to satisfy all the reader’s personal dependencies. To be sure, it may not always be obvious when a dependency is a reader dependency and when it is a topic dependency. Examining the other properties of a Every Page is Page One topic may help us with making the distinction.
To me, this discussion of reader dependencies goes back to something you wrote earlier about self-sufficiency of topics. That is, a stand-alone topic functions well on its own but also relies on the information set that it is a part of, whether that information set is the entire web or a more constrained set of topics. This I think is one of the crucial goals of topic-based authoring: that information sets constitute a well-defined web of topics that are related by their links and other metadata.
Standalone web articles that link to a hodgepodge of external sources don’t succeed as well by this measure, because of the uneven quality of their link destinations and the inability of the destination pages to truly support readers’ dependencies. And where the information set is huge or ill-defined (think Wikipedia) the same problem often arises.
HI Mark,
I agree, a standalone topic on the web relies on the entire web, which is a rich, but not entirely satisfactory source.
Tech pubs, on the other hand, often fails to create satisfactory standalone topics, and as a consequence, I fear, we are tending to lose our audience to the web. We need to figure out how to create good standalone topics, and to link them more reliably than the miscellaneous topics on the web are linked. Alas, linking often takes a back seat in topic-based writing systems.