As I say, I’m not really a taxonomist/ontologist. I try to simplify complex ideas so that they are more widely digestible. I think, though, that your last point (i.e., [paraphrase] that one should focus more on site structure and navigation than on taxonomy) can only take one so far. If you have a small site, then creating an ontology, I agree, is a lot of work for very little gain, but navigation and site organization quickly become too complex to reveal the site’s content after a few areas and levels (across and down). Browsing becomes impossible at that point, and site search remains unreliable.

In addition, I think it’s impossible to talk about “metadata” on its own because the risk is too great to overlook the importance of defining a formal schema. Tagging becomes inconsistent without a solid mapping of terms. So in a particular content domain where people have many ways of finding what they need, the metadata must be applied consistently according to the structure to ensure that everything falls into its rightful “place,” regardless of how the visitor defines that “place.” My own goal for content is to have it arrange itself wherever the visitor expects to find it, rather than defining a website structure to hold it.

What do you think?

Stephen

In the case of taxonomy in the world of software or web design it’s more difficult because in some projects the traditional meaning and the new meaning will sometimes clash.

From conversations with people, it seems to me that most people see the hierarchal structure of content on a website they make the leap from hierarchy to taxonomy, but of course just because something is structured hierarchically doesn’t mean that it has been structured by the rules of a taxonomy.

Because of this I’ve avoided using any new meanings for taxonomy, and instead discuss metadata or site structure.

Overall I think you’ve got a great article here, but there were a couple of points I think are important to discuss:

>> “…Amazon and other e-commerce giants have made such ubiquitous and successful use of taxonomy to sell related things, but it’s really quite difficult to establish those kinds of relationships in your content without taxonomy.”

I don’t (and have never) worked for Amazon, but my understanding of how their recommendations work are mostly based on customer buying behavior (so relations are made by patterns found between what users are browsing, purchasing or adding to wish lists).

>> “…Whereas in a biological taxonomy, we’re dealing with only one dimension of relationship, the ultimate relationship of one species to another through its name…”

Living species could be classified by any number of ways (habitat, colour, type of noise they make… but like you mention earlier, the power of a taxonomy comes from it being widely adopted so that different groups of people can classify, analyze, share, discover and learn around a common set of information.

When more complex relationships start to form, instead of using multiple taxonomies, an easier approach is to create an ontology which has must more flexible classification rules (think of a network or a graph instead of a tree).

Traditionally, taxonomy has been a preferred approach to classification because in the past it was the only practical way to maintain a complex set of records (we were physically constrained in the past with paper based records – think back to the old stacks found in libraries).

The potential of developing and using a taxonomy (or ontology) in software design is huge. Especially in the area of search – but from my experience, unless you’re working in document/record management, or in the area of content governance then you probably don’t need to think about taxonomy at all.

Instead people should focus on a structure that allows people to find what they are looking for – a great place to start here might be some card sorting activities with people from your target audiences.