A faceted classification (of some set of objects) is built upon the fact that the objects in that set can have multiple characteristics or facets, e.g. a Frisbee, which can have facets of weight, size, stability, density, flexibility and so on.

Thinking of objects in this way enables the classifications of those objects to be ordered in multiple ways, rather than in a single, pre-determined, taxonomic order.

In a faceted browsing navigation model, the notion of browsing - of traversing a taxonomy - begins to blur with the notion of conducting a search - of filtering possibilities down to just a few plausible candidates.

This is because the idea of a single, pre-determined, taxonomic order - as per a traditional navigation model - is replaced, at least in part, by a menu of facets. Using concomitant filtering mechanisms, a user can quickly and progressively refine a search down to a few likly candidates very quickly.

In my experience, there are no purely or wholly faceted navigation models: there is always some hybridisation with traditional taxonomic models. Faceted browsing models are beginning to crop up more and more, esepcially in e-commerce and other retail contexts with a large portfolio of products.