Taxonomies and Tags

I found this post the other day by David Weinberger, best known as co-author of The Cluetrain Manifesto. It’s entitled Taxonomies and Tags: From Trees to Piles of Leaves. I knew I had to blog on it as soon I encoutered the pithy and nearly impenetrable phrase “webogeny recapitulates ontogeny.”

Anyway, I’d read the post because it includes a nice history of taxonomy, briefly argues that we are entering a third intellectual order (where there is no line between data and meta-data — which conveniently is exactly the way XML works), and it provides a nice introduction to meta-tagging.

As a comment, I’d observe that there has been some major, seemingly unnoticed, parallel evolution occuring between the data and the content worlds. During the past few years data people have been excited about OLAP dimensions and multidimensional navigation. During those same years, content people have been exicted about taxonomy and faceted navigation.

Guess what? They’re the same thing.

Data people want to slice-and-data using hierarchical dimenisons and hypercubes: show me sales of sports watches by brand in NYC in the month of April. Content people want to see ethnic recipes using chicken with less than 500 calories per serving and prepared in 30 minutes.

In both cases you are talking about organizing information along multiple dimensions. (A dimension might be geography, product, time period, or subject area.) In each case these dimensions are hierarchically structured:

• Time: Years contain months which contains weeks which contain days
• Product: Fashion include watches, jewelry, and shoes which each in turn contain numerous subcategories
• Subject: Cancer contains lymphoma which splits into non-Hodgkins vs. Hodgkins lymphoma, each of which has numerous subtypes

When you combine multiple hierarchically structured dimenisons as a way to classify information, you are, in effect talking about an OLAP hypercube. Most people don’t think of it that way. Few content systems actually implement it that way, but the observation is nevetheless true.

I’m sure that this idea, combined with $3.00, will get you a Latte at most Starbucks. But I had to point it out it anyway, if only because so few people seem to have noticed.

Perhaps, ultimately, it’s just another datum in favor of the data/content divide about which I so often write. Data people and content people are inventing the same ideas — and they don’t even seem to know it.