One of the dimensions in which I categorize information environments is the degree to which they are open to re-configuration by their users. Some environments have structures designed to enable prescribed experiences while others have structures that encourage emergent (read: unscripted) experiences to happen.
For example, if you are designing a website to support the marketing of a product, you probably want people to eventually purchase the product. But to get to the place where they do so, they must first learn about the product. There is a causal relationship between these activities: one must come before the other. (This is what is called a “sales funnel.”) In this case, the website’s structure must allow people to learn about the product at their own pace, but it must also make it easy for them to find the checkout page. This calls for a top-down structure that encourages this behavior and remains invariant through the interaction.
Compare this with an environment designed for people to collaborate on multiple projects. Users in such a place must be able to create new project spaces and arrange them in different ways suited to the needs of each project. Some projects may require more interpersonal collaboration while others may require a greater focus on documentation. These uses call for different structural arrangements, so the designer of the underlying information environment should create structures that encourage end users to re-configure it for their particular needs in each project. (This can be done using user-configured navigation bars, tags, and other UI widgets.)
An environment such as this that was structured too rigidly would be stifling, just as an environment that allowed users to do whatever they wanted would not accommodate the sales funnel very well. Designers must understand the uses the environment is meant to accommodate so they can decide how flexible it should be.
It’s worth noting that in both cases designers are defining structure; determining the contents and location of a global navigation bar is as much a structural decision as deciding that the contents and location of the navigation bar will be defined by its users later. So is implementing a user-facing tagging system. The only difference is that the latter encourages the emergence of structural constructs the designers of the underlying environment can’t foresee. As with so many things, this can be an asset in some cases and a liability in others.