In Web Organization is not Like Book Organization, I said that the TOC does not work at a large scale. It is worth talking a little more about why it doesn’t work, because it has important consequences for many of the navigational schemes that people propose for the Web and other large information systems.
The reason a TOC does not work at a large scale is that a small scale TOC acts as a list that the reader can simply browse, but a large scale one becomes a classification scheme that users have to navigate. If you can’t take the TOC in at a glance, or at least read it through comfortably, then you have to trace down particular paths in the hierarchy of the TOC, and that means you have to figure out the classification scheme behind the TOC. And that does not work.
The reason it does not work is that people do not classify their experience. They do not say, “I have an ache in my second upper bicuspid,” they say, “I have a toothache.”
Experts classify things in the area of their expertise, in order to be precise. Often, they make up vocabulary to express their categories because common usage of common words does not divide the world neatly into categories (because people don’t classify their experiences). Thus biologists classify life into: kingdom, phylum/division, class, order, family, genus, and species.
The botanist says “Plantae/Angiosperms/Eudicots/Rosids/Malpighiales/Violaceae/Viola/Viola tricolor”; the common man says “Pansy”. The expert classifies; the common man simply names.
People call the pansy by a many different names, but seldom by its botanical classification.
Actually, the common man does sometimes classify to a certain extent, but often in a completely different way from the expert. If the common man is a gardener, he may look in the garden center for “bedding plants / perennials / Pansy”. But if he asks the clerk for assistance, he will just say “Where are the Pansies?” Or if he is reluctantly in tow behind his wife, he may say “the small purple ones”.
People do not classify their experiences; they name them. And they do not name them according to a universal taxonomy. Often they name them according to local custom, so that even “pansy” is not always “pansy”. As Wikipedia explains:
Many of these names play on the whimsical nature of love, including “Three Faces under a Hood,” “Flame Flower,” “Jump Up and Kiss Me,” “Flower of Jove,” and “Pink of my John.”
In Scandinavia, Scotland, and German-speaking countries, the pansy (or its wild parent Viola tricolor) is or was known as the Stepmother (Flower). This name rose out of stories about a selfish stepmother; the tale was told to children in various versions while the teller plucked off corresponding parts of the blossom to fit the plot.
In Italy the pansy is known as flammola (little flame), and in Hungary it is known as árvácska (small orphan). In Israel, the pansy is known as the “Amnon v’Tamar”, or Amnon and Tamar, after the biblical characters (II Samuel 13). In New York, pansies have been colloquially referred to as “football flowers” for reasons unknown. In some countries of Spanish language, the pansy is known as “Pensamiento” or “Trinitaria”.The name “heart’s-ease” came from the woman St. Euphrasia, whose name in Greek signifies cheerfulness of mind. The woman, who refused marriage and took the veil, was considered a pattern of humility, hence the name “humble violet”. The specific colors of the flower – purple, yellow, and white – are meant to symbolize memories, loving thoughts and souvenirs, respectively.
Another name for the pansy is that of “herb trinity,” with its three colorful petals acting as symbols for the Holy Trinity.
This is why people Google for information. Because if you Google “stepmother flower” you get a bunch of results about pansies. This isn’t foolproof, of course. If you Google “football flower” you get a bunch of results related to floral arrangements shaped like soccer balls or American footballs. On the other hand, if you attempt to find “stepmother flower” by navigating the botanical taxonomy, you are going to get no results at all.
The problem for the technical communicator is how to get the reader who does not classify their experience into the right place in a classified content set. Presenting them with a classification and asking them to navigate it is not going to work. No matter how good or how logical the classification may be, people do not classify their experience.
Teaching them the classification schema and then asking them to navigate the classified content is also not going to work. As content strategists frequently note, if you have to give the reader instructions for navigating your website, you have already failed. When all they want is a single item, people simply don’t care how that item fits in a classification; they just want the item. They expect, however unreasonably, that the answer should be as simple and straightforward as the question. And the question, as far as they are concerned, is always a simple one because it concerns one entirely concrete and discrete piece of their own experience.
The challenge, then, is to lead the reader to the right piece of content without in any way teaching them the classification scheme by which the content is organized.
This is not a new problem. While the TOC of a book presents an organizational schema for the content, the index presents an unclassified list of names. Readers can use the index to find something in a book without ever engaging with or understanding how the book is structured or organized. (The problem, of course, is that the content they find this way is often written in a way that assumes you are following along in the organizational schema of the book. This is why Every Page is Page One topics provide a better reader experience than books when the reader is not reading the whole book.)
For the writer, the information architect, and the content strategist, the lure of the classification schema is strong. All of these, to one extent or another has to concern themselves with the overall information set, and with its completeness, and these concerns are impossible to address without imposing some form of classification on the content. The writer, the information architect, and the content strategist have to create a classification schema, and it quickly becomes for them the easiest and most reliable way to navigate the content set.
No wonder, then, that they swiftly come to believe that what works so well for them must also work well for the reader. But it doesn’t. The writer, the information architect, and the content strategist classify their content because they work with it all day and think about it constantly. But the reader does not classify their experience. The hierarchy that is so lucid to the writer, the information architect, and the content strategist, is entirely opaque to the reader.
The reader does not classify their experience. We have to stop organizing content as if they did.
You are talking about my favorite subject, thank you, and I don’t agree! At least I don’t agree when the reader is a bit more ambitious than just getting a simple answer.
Intelligent people tries to create some order in life. Carl von Linné categorized plants and created a structure. You learn a about Spore plants and Seed plants in school. If someone says “This Red Algae is a Spore plant” you know a lot of things about the algae just by the category.
Then you forget and become less smart.
If you could find information in the context of a structure you would re-remember things once forgotten. When searching for the Red Algae I would love to find it as a grand child to Spore plants. I would then have the option of reading once more about the characteristics of Spare plants as compared to Seed plants.
No structure is perfect but any structure is better than none. Today living things are structured according to different biology branches (Molecular, Cellular, Population) so today I have the option of viewing my Read Algae in different contexts.
If I search for Caesar I would love to find him in a clear visible structure: The classical ages – The Roman Era – The Roman Republic (510 BC – 23 BC) – Caesar. By just looking at the structure I will discover that the Greek Era was before the Roman. That makes me smarter.
When the structure is present whenever searching you will start to navigate instead of searching. Navigation is much quicker when you are familiar with the structure. Searching always has the risk of missing the target. A hierarchy is an extremely compact navigational structure. I guess all knowledge in the world requires less than 10 hierarchical levels.
The reader does not classify their experience. The writer and the presentation system should help readers re-remember classification. There is a knowledge structure behind all the found fragments and the readers will love to discover it.
I would like to say that you are correct. But you aren’t.
Categorization helps experts on a specific topic to find things more easily. But it doesn’t help the others in any way.
People are simple minded, when things are unknown to them. They don’t know much about terminology, structure, classification or the general topic itself.
Let’s take your algae example (which fits perfectly, as I was studying biology). For me – even while I was an “expert” at the topic long time ago – the algae would be a small, maybe living thing at first glance.
While I’m familiar with typology and categorization of living things, I wouldn’t think of using the classification books first hand. I would try to search directly for the characteristics of that “thing”.
And this is, if I would try to find out something about an object in my former field of expertise.
If I change the subject to something more abstract, like sailboats, I’m completely lost. I’m a really stupid user in this field of expertise. I don’t have the slightest clue about sail ships. All I can categorize is: boat, mast, sails. That takes me as far as categorizing the thing as a sailboat and that’s about it.
Next logical categorization would be the type of the boat. I don’t have a clue about boat types and would have to guess wildly.
Then there might be a classification regarding the sails (number, type). Not the slightest idea.
You see where this is leading. I’d just take a look at the “navigation” categories, get lost immediately and start searching in Google. After I found what I was searching for, I might even become an expert and find a categorization useful – but only then.
I would like the categorization to sneak up on the readers!
If they search for help when using a product the manual structure should be discretely presented around the found fragment. The fragment should always be part of a context – the reader can ignore it or exploit it.
I am not competing with Google – searching will always be needed. The structure is a complement. After a while the structure becomes familiar and eventually the reader can benefit from it.
One of the design principles of Every Page is Page One topics is that they should establish their context. But context is not position in the book, it is position in the subject matter, which is a much more complicated and multi-faceted thing than position in a book.
I think you hit the nail on the head here about why categorization does not work as a finding mechanism.
That’s true. If you use the same resource often enough, you eventually get to know how it is organized, and that knowledge helps you find things. But its “eventually” that presents the problem. It takes many many repetitions for the structured to become apparent, and so the structure is no help unless you are using the resource pretty frequently.
Most of us don’t use tech docs often enough to learn their structure. So the structure is not helping anybody the first several times they use the resource, and is not helping most user ever, since they don’t use the resource often enough to learn its structure.
You are correct to say that the writer is not competing with search. But in a world where search dominates, the question is, are you cooperating with search. That, really, is what Every Page is Page One is about: writing in a way that cooperates with search, and with linking, as the reader’s modes of navigation.
Yes, I think that describes the situation exactly.
It is worth noting, also, that even when we set out to study a subject, its classification schema is not a useful starting point. We need classification when the individual objects we are trying to manage become too numerous to keep track of. We have to be deep enough into a study to have accumulated that many objects before we find classification either useful or comprehensible.
This, I think, is the flaw in so much teaching. The expert is so comfortable with the classification schema that they come to see it as the root of understanding, and then attempt to teach it to the novice as a starting point. It doesn’t work, because the novice does not have enough experience of the individual objects to make sense of, or see the value of, the classification schema.
And this, interestingly enough, is exactly the reason the tech writers give for why SMEs can’t write user docs: they are too close to the subject. They can’t translate it into terms the user will understand. But in imposing a classification schema on our content, and forcing the reader to learn that schema in order to navigate it, we are doing exactly the thing that we say disqualifies the SME from doing our job — casting the subject in expert terms that the user does not understand.
Thanks for the comment, Jan.
One of the great temptations for any writer is to write for the audience they want rather than the audience they have. For a novelist or an essayist, this is okay, because if the audience you want is out there, and if you actually write something that interests them, they may find you.
But the technical writer has no opportunity to find an audience other than the one that is handed to them. The audience is the user of the product, and very few user of the product are more ambitious than getting a simple answer.
Classifications are very useful for students of a subject who wish to study that subject in depth. Most users of most products are not students of that product and do not wish to study it in depth. Indeed, the whole arc of product design is to reduce the amount of knowledge the user needs in order to use a product, because the less intellectually demanding a product is, the more it will sell. Our job as technical writers is not to lead the reader into deep study of our product, but to enable them to use it successfully with as little intellectual effort as possible.
Easy to use mean, principally, requiring the least intellectual effort and background knowledge as possible.
Now, all that said, there are cases where a deep understanding of a system is necessary, and in these cases, learning the vocabulary of the system design (and its implicit categorization) is required. It is not the common case, but it does occur, and these cases require a different approach to technical documentation, and also to employee qualification and training. But in these cases, there is no finding of information involved. The curriculum is planned and laid out ahead of time and handed to the candidate as a whole package.
What do you think about about multiple indexes in books? Different ways in.
One reason I think print is still useful is that surrounding information, page headers, and the position of the material in a book provides some kind of context to what the material is about, and allows the reader to get the gist of it. Strict DITA and EPPO topics (I think) would lose this.
You can also remember where information is on a page (“table near the top”).
Even the (topic-based) UNIX man pages, included modern well-maintained ones like OpenBSD, retain numbered section numbers after the commands to give a sense of place and context, even though they seem a throwback to print.
I think the EPPO approach is excellent but is complementary to other approaches.
Thanks for the comment, John.
I think multiple indexes are fine if the differences between the indexes is based on a categorization schema that the reader understands. Thus separate indexes of actors, directors, and movies would work well for a movie review collection, because every reader will immediately understand that classification. But if the reader does not understand the classification schema, then a single index is preferable.
You are right about context. The necessity is providing context is one of the things I really stress in EPPO design. Topics formed by bursting book often lack context, which makes it hard for the reader. But the context provided by a book is actually pretty weak. It is not the place in the book that cements a topic in its context, but its place in the subject matter. A well designed EPPO topic will do far more than a book can to place a topic in its subject context, and to make that context navigable.
The memory of physical location is certainly real, though I find I have far less of that than I used to. I’m not sure if that is because I am learning to rely on other methods of remembering where things are — or simply not bothering to remember because I am confident I can find them again. But I think it does point out the importance of finding aids, and I don’t think we do nearly enough right now to provide Web-like finding aids for the content we put up on the Web.
Definitely, though, EPPO is not meant to replace books entirely. Long form narrative still has its place, and always will. But a lot of communication tasks that used to be handled by books are not being handled by the Web. EPPO is about designing content that is intended primarily for the Web (and other web-like systems).
I thought I could share some structures! Hope these links work.
Biology
History
Thanks for these examples, Jan. I think they illustrate two very important things.
1. Categorizations don’t work as TOCs. Imagine if you were an English student studying Shakespeare’s Julius Caesar and you wanted to look up information on the real Caesar. You would not have to knowledge that would enable you to find the information you needed by descending this hierarchy.
2. Categorizations can be very helpful in placing a particular topic in its context in its subject matter. The Caesar example certainly helps place Caesar in his historical context once you have arrived at the topic on Caesar. Note, though, that it only places him along one subject axis. Caesar actually exists on several subject axes. He has a place in Latin literature as an author, in English literature as a character, in political history as the model of a ruler (that, for instance, gave Russia the Tsars and Germany the Kaisers), etc.
And that very much illustrates the point of the post. The classification is very useful to the expert, but we should not make the mistake of using it as finding aid for the non-expert, because the non-expert does not categorize their experience, and therefore cannot navigate a category based hierarchy.
And also, because as subject can exist in multiple classification schemes, all of which may be of use to the more studious reader, it is better to place a topic in multiple classifications than in just one. The Wikipedia article on the Manicouagan Crater, which I have used to illustrate my last two posts, is a perfect example of this, placing the crater in its context in both astronomy and geology.
If you are looking for Caesar you might find a salad so narrowing the search is necessary in order to find Julius.
When the correct Caesar is found I would love to have him presented in a context. Yes, history can be structured in many different ways but that should not stop us from showing one of them – one is better than none.
We can let the reader choose among different structures. This is probably only relevant for experts. In school we usually (and hopefully) learn one structure so why not take the most common one?
I am not saying we should use classification as a finding aid (to start with). My point is that it does not hurt and it will give the reader a feeling that the information found is being part of something bigger which is not fragmented.
A structure is not only for navigation. It communicates something. Let’s take a Safety chapter. “Safety is important to us and it is built into the product. Here you will find everything regarding safety.”. The chapter contains different safety topics – the reader found one topic in his search. A structure gives you the “whole” picture.
The Roman Era, the Spore plants and the Safety chapter give us the skeleton to hang our fragmented knowledge and they make us wiser.
“If you are looking for Caesar you might find a salad so narrowing the search is necessary in order to find Julius.”
Actually, what you often need is a broad search, so that you get to see all the possible connotations of the word you are searching for. A search does not have to return only a single result, and the best ones return many for precisely this reason. You can’t narrow a search unless you have enough information to narrow it, and often you need to search in order to get that information. (Wikipedia’s disambiguation pages are a really good example of doing this well.)
I think it is important to make a distinction in user behavior between exploration and retrieval. Exploration is a quest for new information or new understanding in a field you know relatively little about. Classification is of little use for exploration, because you don’t yet know enough to understand the classification.
On the other hand, classification is essential to efficient and accurate retrieval. If you are looking to precisely retrieve a single piece of information, then you must classify it accurately so that you can be sure you get back the right data, and that, of course, requires a knowledge of the classification scheme on the part of the reader.
Both exploration and retrieval play a part in technical communication, but most of the end user information retrieval problems lie in the exploration area.
I agree, though, that classification communicates something, which is why I say it is useful within a topic to place its subject within one or more classification schemas. Is one better than none, for this purpose? Absolutely. But two or three or four are better still.
I would not go so far as to associate classification with wisdom, though. Classification is a necessary tool to impose some order on the chaos of reality so that we can manage it with out limited brains. But it is also always false to one extent or another, always involves choosing to make one characteristic more important than another, for the sake of a classification that we can get our heads around. Wisdom would consist in knowing the limits of classification as much as it uses. Placing a topic in multiple classifications is one step toward wisdom, in that it reminds us of how complex the world is, and how limited our classifications are as a tool to deal with that complexity.
You still have to have some way for the user to start.
Preferably, you need several such ways.
Guides, Indexes, Search, FAQ’s, a set of classification schemes: all are useful as a starting point. A list of tasks may be too big, but may be a small enough set to usefully apply a classification to.
Then you need to have cross-references to closely related material, and to “maybe you meant this” material.
Helping people to find the right information in a large documentation set requires a multi-faceted approach. A single classification of all your topics is not sufficient.
Having said that, you can still apply classification to provide experts with a useful structure to your material. It’s classification alone, or as the first view of your data, that doesn’t work.
Rob, agreed. You need to provide as many different ways into the content as you can reasonably afford to. For one thing, people come to your content with different background and different information seeking habits, and no one navigational technique is going to work equally well for all of them (which is part of why on one classification scheme is going to work for all of them).
Overwhelmingly, though, people’s first choice is search, which means that they land at an arbitrary page in your content set. The problem with most content sets today, especially those derived from books or constructed on book-like principles, is that they provide very poor support for contextualizing the reader when they arrive at that arbitrary page, and very poor facilities for onward navigation once they are there. Thus, as you say, “you need to have cross-references to closely related material, and to “maybe you meant this” material.” And I would go a little further and say that you need to support navigation along all the lines of subject affinity that converge on the current topic.
What I would really like to encourage is that people start paying a lot more attention to onward navigation from each individual topic and less to top-down navigation outside the topic. As the Manicouagan crater article demonstrates, placing a topic withing its classification within the topic itself can be a powerful way of providing rich onward navigation.
Classification can play several important roles in a content set, but none of them should be based on the assumption that readers classify their experience because, by and large, they don’t.
Pingback: Sometimes, Readers Do Classify their Experience - Every Page is Page One
Pingback: Can Help Content Have Recognizable Facets? | I'd Rather Be Writing
Pingback: Why simplicity is more important than functionality in content navigation - Every Page is Page One