Imposing Order Versus Observing Order [Organizing Content 4]
May 18th, 2010 | Posted in blog 12 Comments »
It’s easy to postpone organization. We begin writing discrete help topics, hundreds of them, and then try to group them together in a logical way. But here’s where the problem starts. What does it mean for a system of organization to be “logical”? And how does the user navigate this logic we create?
Our system of organization partly determines the findability of the content. Without findability, we might as well not even write help at all. It’s exactly this lack of organization/findability that turns so many users off with help content. When you click the help button and see an immense amount of folders and subfolders and sub-subfolders, all organized in seemingly impenetrable ways, it overwhelms users. They give up within seconds of their foray into the help.
As help authors, it’s so easy to be seduced by technical distractions and overlook the content. We get drawn into styles and design, the look and feel of the help content, the single source rendering between print and online, the parallelism of titles, and the exacting conformity to our style guide. These technical bells and whistles distract us from more fundamental matters of content organization.
But content organization is fascinating. The way a help author lays out the help topics in a table of contents shows you more than simply a list of topics. It shows you how the author has wrapped his or her mind around the content, how he or she has chosen to shape order from chaos. It shows you how the author understands the user. And it shows you one perspective on the structure of the content.
Organizing the Periodic Table of Elements
Organizing a jumble of help topics into a natural, beautiful order that strikes clarity in the mind of the user is no different from the scientific urge to classify, to organize, label, and categorize. Nothing illustrates this better than the story of how Dmitri Mendeleev organized the Periodic Table of Elements.
Dmitri Mendeleev
Before Dmitri Mendeleev organized the table, he traveled extensively in Russia on trains flipping around a stack of cards that had the elements and their behaviors written on them. Sorting and resorting the cards, he kept searching over and over in his mind for a pattern, for a principle he could observe that would lend itself as the organizing pattern for the elements.
Sure, the elements had different atomic masses. The atoms of some elements were heavy; others were light. But the atoms also displayed other behaviors. Some elements were friendly with other elements; other elements were shy, like loners.
This seemingly random array of heavy, light, shy, friendly characteristics made it tricky to organize the content in a logical way. What was the right way to organize these 90+ elements? What underlying logic explained the element’s natural classification and hierarchy?
The right system of organization finally came to Mendeleev in a dream. A sudden flash of insight, not an order he created, but an order inherent in the content he observed. He finally saw the elemental grid.
He not only arranged the elements by the degree of atomic heaviness, but also found that the elements had other patterns that repeated periodically. Hence Mendeleev called his table a periodic table, because the patterns repeated with specific periodicity. Through the repetition of this pattern, Mendeleev also predicted the existence of elements not yet discovered.
Radiolab
I listened to this story of Mendeleev organizing the elements on Radiolab, with hosts Jad Abumrad and Robert Krulwich. They ended the story with the following question: Was this organization an invention that Mendeleev imposed on this otherwise random set of elements? Or did Mendeleev see the underlying law that naturally organized the elements? In other words, is organization a human fabrication we impose on content, or do we observe an organization that is naturally inherent in the content?
By the way, I excerpted a short, one-minute clip from the Radiolab podcast here. Listen to the way Oliver Sacks, Abumrad, and Krulwich tell the story of Mendeleev:
Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.
Mendeleev apparently rode these trains for years before finally reaching his organizing vision. Here Abumrad and Krulwich elaborate on the principle that Mendeleev rationalized:
Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.
And the final result today:
Periodic Table
The Problem with Scientific Classification
After listening to the story of Mendeleev organizing the elements, I was convinced I too could find, if I just looked carefully enough, an underling pattern that would serve as the perfect organization for my 200 help topics with Project Swordfish.
But the more I looked, and the harder I thought, I could not come to any conclusion except that a software application is man-made. It does not conform to any unseen natural laws in universe, and finding a natural pattern is an unachievable illusion.
The only system of organization that approximates a natural law is to organize help according to the psychology of the user. In other words, rather than looking at an underling organizational structure in the content, look at the way the user’s mind organizes and structures the content instead, and use that mental principle as the pattern.
Having thought through all of that, somewhat fruitlessly, I scrapped the other content organizations and created a standard topic-based TOC that looked somewhat like the following:
Exchanges
Organizing Exchange Locations
Evaluating Suitable Environments
Reading Informant Microgestures
Burn Notices
About Burn Notices
Managing Your Life After the Burn Notice
Dealing with Fake Burn Notices for Double Agents
Sting Operations
Finding and Using Citizen Operatives
Recycling Burned Agents for New Operations
Evaluating Environments in High Risk Situations
Deciphering Informant Trust
Calculating Risks Based on Informant Credibility and Environment Security
This approach to the TOC follows the conventional topic-based approach, attempting to group like-minded content into its logical topic container.
Limits of Logic
This standard seems like a good approach, except that topics don’t always fit neatly into their “logical” containers. You can see that in the last section, Sting Operations, the tasks include evaluation of informants and assessment of environments — similar to topics under Exchanges. The remaining help topics also contained similar overlap.
The lack of a clean, neat grouping of content into topic containers is not just an exception, like the egg-laying platypus is an exception to its mammalian classification. Content in real projects is messy. It overlaps. It duplicates. It stretches outside of its row like an overgrown tangle of weeds. It could easily be classified in several different containers. When you look at the logic of the organization, it often tells you more about the organizer than the actual content. In the end, it isn’t really logical at all.
According to Abe Crystal in the Tech Comm Journal, “the idea of a single, unified hierarchy,” in other words, the standard topic-based containers organized in a hierarchical, logical arrangement, is “the fundamental weakness of standard IA [information architecture] frameworks.”
Yes, this single, unified hierarchy, the traditional topic-based approach to classifying content, which I struggled to grasp and implement with my 200 topics, is an old, tired framework. Instead, Crystal suggests moving in a more fruitful direction: faceted classification.
—————-
Crystal, Abe. “Facets Are Fundamental: Rethinking Information Architecture Frameworks.” Technical Communication Journal, Vol 54, No 1, Feb 2007
Sponsors
Tags: abe crystal, dmitri mendeleev, faceted navigation, facets, hierarchy, order, organization, periodic table, radiolab, structure, table of contents, taxonomy, Technical Writing
If you liked this post, keep updated with new content: Subscribe to I'd Rather Be Writing.
Both comments and pings are currently closed.
Mark Baker (5)
DiSc (3)
Ellis Pratt (3)
Jen (3)
Neal (3)
Richard Hamilton (3)
Jenny (2)
Sarah Maddox (2)
Sarah O'Keefe (2)
Stefan Kleineiken... (2)
One way to keep the topics from being so messy — from not fitting neatly into their containers — is to make them succinct and focused. For example, a topic that describes a task and also explains the underlying concepts, or a topic with two similar but distinct tasks, ought to be broken into two topics.
You’ll still have some messy topics that defy classification, but not as many.
If I know what you mean by faceted classification, and I think I do, then this succinct and focused approach will facilitate that too.
Excellent point, Larry. I agree that brevity increases focus, which makes topics fit more accurately into topic containers. However, too much chunking creates an abundance of topics, which makes topical navigation more onerous for the user, right? For example, suppose I have a topic called Forming an Operations Team. In the topic, I have an explanation of why one would need an operations team and the criteria for agents to be on the team. Then I have three tasks in this same topic — adding agents to the team, removing agents, and managing agents in bulk on a team. If I split this into 4 topics, and I did that for every topic, my number of topics would be 4 times as big as creating the TOC the other way. Is that something to worry about?
Is that something to worry about? Could be. This a big part of the whole “imposing order” conundrum, isn’t it?
The best approach here would be to configure the TOC so that it only displays two, or at most three, levels down from the top. The reader can still navigate easily to the main “Forming a Team” topic without being overwhelmed by the size of the TOC.
Topic-based and hierarchical approaches are far from “tired.” They are crucial in complex disciplines such as the Life Sciences. Facets are intended to enhance those approaches–not supplant them.
I don’t think we should discard topic-based navigation altogether. But I do think that part of the frustration users have with help, the reason so many people dismiss help as being useless, stms from an over-emphasis on the topic-based navigation. It makes no sense to the majority of the users.
Many years ago, Dave Gash and I did a presentation at a conference where we dissected an existing Help file and shared our thoughts and opinions. Two of them were:
- A TOC is ONE person’s interpretation of how users might find information. (As Tom has pointed out in this series, not everyone will find that interpretation useful.)
- Just because a HAT automatically adds every topic to a TOC doesn’t mean that you have to keep them there. In fact, a TOC should never include every topic in the project.
In the case of Tom’s earlier example, “Forming an Operations Team” could be the only entry on the TOC, especially if users understand (and if the Help author actually provides) that conceptual and task information related to that topic will be found “nearby”. As Larry said, don’t mix conceptual, reference, and task information in the same topic (this is the model that I first learned in 1995 with WinHelp 4 and that DITA follows). And provide links between the topics so that users can easily navigate between them.
Tom, this is why I think doing things like card-sorting exercises may help. I haven’t read the next post yet, but at this point, it seems like you are trying to get in the heads of the users without talking to them. (Though I know you’ve provided training and support to a lot of users and know a lot about them as a result.) This is why I’m looking forward to trying some card-sorting. I may find out that the users categorize information much differently than I do.
I’d like to say that it’s not all bad to write a bunch of content and then organize it later after performing research such as card sorting. This may be possible with a waterfall approach to product development. However, Agile development for software doesn’t really allow for that. We need to produce topics specific to an iteration’s work and jam it into some sort of organization in the limited time available. It seems that this leads to the situation we’re now talking about: trying to rework our content organization into something that makes more sense than the way it’s currently organized.
Any of the older folks reading this ever find a physical book in a physical library? Back in ye olden days, I used to spend most my days in that now archaic activity and my experience was that it works beautifully and very consistently even though it makes use of an arbitrary hierarchical single catalog system. In 30 years of doing it, I never failed to find the book I was looking for, unless it wasn’t there or someone misfiled it. Even then, the system usually led me to find the errant content. That’s pretty impressive, but admittedly it works because the content is itself structured and then tagged based on its place in the scheme. We clearly need something more flexible today. But flexible in what ways exactly? That’s the question.
I guess I start by looking at this backwards. We are looking at a “mess,” an arbitrarily collected content pile that happens to cling together for some historical reason (perhaps just recency), and then trying to impose a taxonomy on it. Clearly that’s a whole different activity than what Mendeleev was attempting. We can still do that through card sorting techniques, etc., with a defined audience, but of course it is artificial and the only meaningful organization is the psychology of the reader themself as they ask questions to be answered by the content.
As an alternative (the backwards part), I imagine producing content that is meaningfully organized. The taxonomy then reflects the cognitive structure of the content, the thinking processes of its authors and the thinking processes engaged by readers as they successfully navigate the content. Sort of like what we used to do when books rather than tweets and blog pages were considered good models of content. A good annotated table of contents in that situation is an amazing map of the content, telling you where everything is clearly and unambiguously and placing it into context. This gives you places to put other content as well.
Of course there is no final taxonomy and no one magic table of contents for all books on the same subject, but why should we limit ourselves to a single one? There are different ways of thinking about a subject, and each suggests its own taxonomy. The taxonomy simply (well not so simply, this is a bit tongue in cheek) has to translate from the question the reader is asking to the right sort of taxonomy to address the question.
I suspect that with a little thought we can come up with dynamic taxonomy matching based on the nature of the questions being asked. Ok, the current search engines do this in a rather silly way, but that doesn’t mean it can’t be done for real.
My point is I think we need to preserve the intellectual power of content structure as a guide to thinking, the natural underlying structure of the topic (in cases like natural science where it actually matters!) and also allow for the dyanmic and more arbitrary nature that seems to permeate so much of other sorts of content. If some variation of “faceting” accomplishes this, wonderful! If not, then time for another idea that doesn’t throw out the meaning and relevance struture of human thought in the name of good hit rates and popular results.
I’ve been creating portal taxonomies for years for clients, and some work better than others, but the best by far are the ones that map content that is actually organized in a thoughtful way. If we combine a way of mapping queries to dynamic taxonomy with some recognition of the structure of the content, we meet in the middle by generating a meaningful organization for the content that shifts with the perspective of the query being asked. Would that make more sense than a one-size fits all taxonomy or an arbitrary classification scheme?
Todd, thanks for your reflective comment. I agree that multiple taxonomies, such as combining “a way of mapping queries to dynamic taxonomy with some recognition of the structure of the content,” as you point out, would work well. The one-size-fits-all model fails to be useful for all users (however thoughtful it might have been). Implementing a dynamic system of organization, or providing focused entry points generated from user queries and metrics, would make a great complement to a topic-based hierarchy.
The problem for most help authors is the lack of a toolset to create this dynamic content organization. We see sites on the web that routinely show us the most popular content based on hits or references, or we see a list of the most common keywords users are searching for, but most HATs can’t provide this dynamic delivery of information. As such, technical writers get stuck in old-school mode.
Re libraries in general, I agree that they often do a good job at making information findable. However, when there’s a lot of books about a specialized topic, it’s more challenging. Maybe a user can find all tech comm books relatively easily when there’s just a couple of shelves of tech comm books. But when an entire floor consists of tech comm books (similar to the abundance of information online), it’s a whole other game, and that’s where the topic-based hierarchy really breaks down for me.
Great points, Tom. Let’s get to work on that toolset!
Hi,
I wanted to let you know that I have been following for a a couple of months on and off and I would like to sign up for the daily feed. I am not to computer smart so I’ll give it a try but I might need some help. This is a terrific find and I would hate to lose contact, and maybe never discover it again.
Anyway, thanks again and I look forward to posting again sometime!
When you get your personal ipad time, maybe you also would like to have your personal iPad Case and iPad Accessories to better protect the iPad. When you come to our website, the 1st you have to know is that you may buy the best top quality and most expensive iPad case, also your favorite iPad cases and iPad accessories. There are hundreds kinds of ipad case