Joy joy - Victor seems
Joy joy - Victor seems to be blogging again, and asks: "It seems to me one of the big holes in our knowledge of information architecture, one of the main holes in fact, is how taxonomies become navigation. We're starting to develop very good methods for arriving at taxonomies for modern websites, and we're also getting better at determining what characteristics are apparent in successful navigation. But that junction of taxonomy and navigation still seems to be part of the black art of IA, the challenge of marrying the bottom-up to the top-down."
Here's my take: the disconnect between a taxonomy and a navigation on a site happens when taxonomies are designed without thinking through how they will be used. This ties in with the myth that you can design a taxonomy for something that is not influenced by its use (or its audience). Or even sillier: that there would be a taxonomy without bias. The idea of the objective taxonomy. It's wrong. There is no such thing. So if you are designing a taxonomy without imagining how you will use it in the navigation, yes, it will be hard to turn it into navigation.
Another take: if you look at taxonomies as organization schemes (for example: Geographical, or By Product, or by Task), a general rule when turning taxonomies into navigation is: one navigation section per taxonomy. Avoid mixing them, except in the global navigation, where mixing navigation schemes is often a good thing. If you visually separate organization schemes it will be easier for the user to find the right one for them.
If you look at the structure of taxonomies (ontology vs. topicmap vs. simple hierarchy vs. facets vs. controlled vocab vs. ...), you are thinking about how to design the actual navigation widgets. We know how to best navigate a simple hierarchy. Yahoo. We are learning how to best navigate a faceted structure. Flamenco. The more complex/rich the structure of the taxonomy gets, the more possibilities for creating complex navigation structures.