David Weinberger
KMWorld Archive
This column is part of an archive of David Weinberger's columns for KMWorld. Used with permission. Thanks, KMWorld!


Link to Original at KMWorld  Index

David's home page | Bio | Speaking | Everyday Chaos

Does information need architects?

26 April 2006

At the Information Architecture Summit in March in Vancouver, it was surprisingly hard to find someone willing to call herself an information architect. "I don't like to call myself that because ... " was a common phrase used as people introduced themselves. Similarly, the question of the precise definition of information architecture was frequently raised and almost as frequently skirted. This is surprising at a conference whose attendance is booming and in a profession the value of which is clear and clearly increasing.

What information architects do is actually pretty straightforward. Gene Smith, a noted information architect who doesn't call himself an information architect, began one session by stating (approximately) that information architects structure shared information spaces. Their oft-cited prototype of a well-architected information space is Amazon.com: Not only can you easily navigate to what you want, you can find what you didn't know you wanted. It's also clear what information architecture is distinct from: It isn't Web site design and layout, which should instantiate the design laid out by the information architects. And it's not traditional library science, which has focused on the rational organization of information more than on our electronic interaction with it.

So why the confusion and hesitancy?

In part it's because the constituency of information architects is diverse. Some come out of library science. Others come out of anthropology and sociology. And others have a webbier background. Jesse James Garrett, author of The Elements of User Experience and one of the most respected of the information architects, says that one of the fault lines divides those who think that any imprecision is noise from those who think the difference between signal and noise is context- and use-dependent. If you think there is a best answer for every question, then a system that delivers anything other than that answer is flawed. But if some questions don't have single best answers, or if including multiple answers spurs creativity and the connectivity of ideas, then the "noise" becomes valuable. Of course, it depends on the type of question: "Who wrote Moby Dick?" has a different type of answer than "Can you recommend a book I might like?" Nevertheless, the dividing line among information architects is real. One group tends more toward control and using expertise to create a structure that should work the same way for every user. The other tends more toward flexibility and enabling user interaction to determine the structure of the site and the content of the answers.

Tagging and folksonomy, the subject of much discussion at the conference, look like the triumph of noise to one group and like the will of the people to the other. In a tagging system, users are able to attach brief descriptions (tags) to online resources such as URLs or uploaded photos. Those tags are visible to others, and can be aggregated so that you can see photos anyone tagged, say, "pants." A folksonomy is a bottom-up ordering of tags. So, if users started tagging items at eBay, items that in eBay's official taxonomy are called "pants" might turn out to be tagged as "slacks" by a high percentage of taggers.

In many instances, tagging is a useful way of navigating an "information space," but many people are excited about it because it makes a political statement: "They" no longer get to control how we organize and categorize ideas and information. "We" are going to do it for ourselves, and we'll do it better than "they" ever could. Information no longer has to be shoved into pigeonholes created by faceless information bureaucrats. Tags let us burst out of the pigeonholes, or, more accurately, create our own system of pigeonholes, and fly (relatively) free, dude.

Information architects are aware that some of the loudest expressions of enthusiasm about tagging are in fact directed at those who would structure information spaces for us, i.e., information architects. Thus, their uneasiness with the concept--a greater uneasiness than is warranted by their well-founded critical assessment of tagging as a navigational tool.

But information architects need not worry. They are not going to be displaced by bottom-up tagging systems. Tagging systems will be used with top-down information structures, except perhaps in a few exceptional instances. Information architects, in fact, will be the leaders in figuring out when and how tagging systems make sense. I would think it's a good time to proudly state "I am an information architect, dammit!"