Findability is an intractable problem. This does not mean that we should not try to improve findability. World peace is an intractable problem, but it is still worthwhile to make a friend. Climate change is an intractable problem, but it is still worthwhile to plant a tree. Findability is an intractable problem, but is it still worthwhile to add a keyword.
Still it is important to recognize the that the problem is intractable, lest we waste too much of our time and energy for too little gain.
There is an abiding faith among many in the field that the key that will unlock findability is taxonomy. Searches often fail because readers use the wrong words, they argue, or because content uses different words for the same thing, with the result that the reader gets one set of results for one search and another set of results for another search. Establish a unified taxonomy, they argue, and the problem goes away.
The truth is that the problem is much more difficult than that. Language is not simply the assignment of symbols to objects. It is the assignment of symbols to experience, and so language is history. Experience is normative; language is contingent. A story to illustrate the point.
I was once engaged in a consulting assignment with a global organization that wanted to install a content management system to assist their agencies and partners around the world in finding information in their vast collection. The IT developers charged with this task were convinced that the backbone of the system would be a single global taxonomy which would make it easy for anyone in the world to find information created anywhere in the world. To begin the process of creating this taxonomy, they met with the organization’s librarians.
The developers explained to the librarians how the taxonomy was going to solve all their findability problems. The librarians sat there shaking their heads. They tried to explain to the developers that a farmer in Botswana does not simply use different words for things than a banker in Switzerland. The farmer and the banker think in entirely different categories, the librarians attempted to explain. Different things matter to them, they have different histories and different experiences, and they care about different distinctions. A single vocabulary that fit the banker’s world would completely fail to capture the farmer’s world; a single vocabulary that fit the farmer’s world would utterly fail to fit the banker’s world.
My chief role in this meeting was to explain to the programmers what the librarians were talking about. Because the point of this story is not that bankers in Switzerland and farmers in Botswana see the world in entirely different categories, but that programmers in a western capital, and librarians living in the same western capital, people who live in the same neighborhoods, who had attended the same universities, who watched the same TV programs, read the same newspapers, and worked for the same global organization dedicated to a single set of goals, saw the world in entirely different categories. Getting the librarians to understand what the programmers were proposing, and getting the programmers to understand why the librarians were saying it wouldn’t work, took the best part of a two hour meeting, and even then I am not sure that the message really go through.
For the developers, establishing a controlled set of symbols in a well defined namespace was not just the solution to this problem, it is a necessary part of the solution to every IT problem. Computers don’t deal well with ambiguity, therefore the first thing you do in designing any computer system is to establish an unambiguous set of symbols to represent the objects of interest in the system. Of course the developers saw this as the solution. Whenever we grow thirsty, we draw water from the same well.
Different people see the world too differently for us ever to solve findability. Finding the information you need will always be work. It will always be a skill that people have to learn. We can help, but we cannot solve. Our efforts in the field of findability ought to be focused not on achieving perfect effortless findability (which is unattainable) but at easing the work of finding that the user has to do for themselves.
Pingback: “One Day I’m Going to Figure Out the Solution to Help…” | I'd Rather Be Writing
I would like to agree with you that findability is intractable because ambiguity, which occurs in thinking and natural language, is the result of information reducing. Reducing lead to (often irrecoverable) loss of information, therefore some ambiguities are just irresolvable. But also I do agree that we may help to handle ambiguities as best as we can. You say that “establishing a controlled set of symbols in a well defined namespace” is a necessary part of any IT solution. And this is inevitable. We should establish as unambiguous set of symbols as possible but if possible. Which means we cannot ignore problem of ambiguity and we cannot fully resolve it. You gave the example of taxonomy, but it is more complex than just setting just a controlled set of symbols, it also usually involves hierarchies, which is different matter.
We need something which will join ambiguous natural language and specific computer identifiers. And this is human-friendly identification. It does not mean all symbols would mean exactly what their bringer meant. But it do mean ambiguities between symbols with similar values may be resolved. Moreover, sometimes ambiguities may be resolved quite simply: “Opera” hotel in Madrid is unique world-wide, whereas “I was in Opera hotel” is ambiguous. Identification may make ambiguities as much precise as we can establish. And this may alleviate findability to some degree, at least. Of course, it also means that such identification should not force to establish identity of all symbols, otherwise it may even distort the original meaning.
Different people see the world too differently. But, in any case, they successfully communicate and collaborate with the natural language. If we may make it more precise, why not? What’s more, it should be really human-friendly, which means we cannot ignore usability. Semantic Web does not have appropriate human-friendly representation at the moment. We should think broader. The convergence between usability and content (meaning) should change not only Web and Semantic Web itself but the way we interact with computers. The way we deal with user interface and even files are not ideal today unless we will have semantic ecosystem.
If you are interested, you can find more on this:
http://on-meaning.blogspot.com/2011/06/great-blunders-of-modern-it-and-their.html
I love this post. I’m excerpting a paragraph at the end for a conclusion to an upcoming presentation I’m giving on the topic of findability. I realize that my flaw early on was to assume that there was a single solution to it all.
Cool. Let me know how it goes.