Information Architecture for the World Wide Web

Non-Representational Labeling Systems

This chapter emphasizes the need for labels to be familiar for users, and also that consistency and representation are the foundations for building that familiarity. Now that we have belabored that point, we’ll counter it with another: labeling systems should not necessarily be representational.

What? Would you make up your mind already?

Well, let’s put it this way: non-representational labeling is not something that we’d recommend using regularly. In fact, it’s difficult to determine when it should be applied. Following are two examples where we think it succeeds.

Good Head-Scratching

Head-scratching is usually a Bad Thing. It means that some aspect of a site has confused a user and is in the way of achieving the site’s main goal, namely, conveying a message. But, like everything else, even cognitive confusion has a good side: Mystery.

Consider the main page shown in Figure 5.11. What the heck is going on here? If you come to this site, you may already have a little context, knowing in advance that it’s a personal site. If not, you might figure this out fairly quickly, as this text uses the first person and seems to describe a personal quest. Beyond that, this page tells nothing about what you’ll find in this site.

Figure 5-11. Is it obvious where these links lead you?

But you might want to know more. The radical aspect of this page involves its use ...