Tagging – it’s where the action is right now. At it’s heart, it’s like all great technology ideas; so simple that you wonder why it hasn’t been a mainstream concept for years.

If you’re new to it, tagging is nothing more than assigning keywords to a piece of content, usually at the point of publication. A good example is the photo site flickr. When I upload a photo to the site I choose some words which are appropriate to describe the picture. These usually take the form of place names, people or objects found in the photo.

This is all about simple yet rich metadata, and it’s suprisingly effective. On flickr, for example, it’s now trivial for me to go back and view all my photos of wood, or those depicting Half Dome in Yosemite.

Apply the concept across a diverse group of photographers, and you get a collaborative picture of Half Dome which is built from the combined micro-efforts of each photographer to apply the tag.

Tagging, of course, doesn’t just have to be about photos. delicious uses the same concept to categorise bookmarks, and we’re even starting to see blogging tools that allow users to apply tags to their posts.

As a descriptive mechanism and a search tool, tagging is incredibly powerful. It is not, however, a “folksonomy”. Here’s why.

The beauty of tagging systems which build up a pool of associated concepts is that they’re flat, rather than hierarchical. Whilst it’s possible to infer from the way they’re used that the tag “Half Dome” fits with the tag “Yosemite”, there’s no defined relationship between Yosemite, Yosemite Geographical Features and Half Dome.

This is why “folksonomy” is a misnomer when applied to tag systems. It’s consciously implying a root in “taxonomy”, but misusing that root on the way.

What is a taxonomy?

The classification of organisms in an ordered system that indicates natural relationships.

Tagging is almost the opposite of that – an unordered system which implies natural relationships.

The logical step on from this is to envisage a system where we had a true “folksonomy”, introducing the concept of hierarchical relationships into the extant “tag soup”. I’m pretty confident that human nature would preclude the success of this right now.

The main advantage of “flat” classification is that it’s pretty non-confrontational. Sure, people might object a little if I insisted on tagging every photo I ever took as “Half Dome”, but it’s not in my interests to break the system in that way.

On flickr, you can also add tags to other folks’ photos, which takes the distributed effort mechanism one step further. Even here, there’s little need for confrontation, and in fact the results can be cool. I’m still amused that a friend added a tag to a picture of my lounge to denote the name of my Ikea bookshelves – “billy”.

“Billy” bookshelves, however, neatly illustrate the two big problems a true folksonomy would face.

1) Namespace Collision

Looking at the range of photos on flickr tagged with the word billy there are people, pet dogs and cats, bars and, well, my furniture.

It’s fairly easy using flickr’s tag system to specify, say, a search for all photos tagged with “billy” and “cat”, so I can easily view an implied set of photos depicting cats called Billy.

The beauty part is that there was no huge effort on the part of the person tagging their photo. They didn’t have to somehow pick a tag which said “Animals -> Domestic Pets -> Cats -> Billy” out of a giant tree in order to successfully tag the photo. They just put “cat” and then “billy”.

A hierarchy would necessitate lots of “billy” entries – one for bars, one for people, one for cats, one for furniture… Instantly, we’ve massively increased the complexity of both the tagging application, and the process of tagging each piece of information.

2) Agreeing Relationships

Say we did try to build a hierarchy. I dread to think of the arguments that would ensue. Humans aren’t terribly good at coming up with consistent hierarchical systems. For an example, look at the valiant attempt now being made by WikiMedia to bring together existing biological taxonomies on Wikispecies. The reason a project like this even exists is that up until now the science community has managed to create several fragmented hierarchies which don’t quite agree with each other.

Let’s go back to our pet cat Billy. How would I successfully place him in a hierarchy?

Earlier I suggested “Animals -> Pets -> Cats -> Billy”. But that’s not necessarily the best.

How about this:

“Eukarya -> Animalia -> Mammalia -> Felis -> Felis Domesticus -> Persian -> Billy”

After all, Billy might be a Lion (“Eukarya -> Animalia -> Mammalia -> Felis -> Panthera Leo -> Billy”)

…and so on. By classing my cat under strict species, I can’t easily show that he’s a pet, unless I choose both the classification above and my initial pet idea.

Mind you, there’s an argument with those biological classifications above – I’ve left out several important layers of the biological taxonomy. Some sticklers for accuracy might create a new tree which ensures they’re represented.

This, in short, is a bloody nightmare.

So, a true “folksonomy” is unworkable. Why care?

Simply put, words matter, and this one is a mistake for two reasons.

1) Inaccuracy is bad

By allying itself with Taxonomy, “Folksonomy” misleads the mind to expect the mess illustrated above. This is bad both because some poor fool might try to actually implement such a system, but also because it potentially puts people off getting involved in the useful, simple, pleasurable act of tagging their stuff.

2) Elitism is bad

I don’t know who coined “folksonomy”, but I can guess that it was probably an idea which seemed cute around a beer-laden table in a pub somewhere (see warchalking).

In confusing the complications of taxonomy with tagging and coining a new multi-syllable word, however, the “geek” community who are popularising it are building a wall around it. Whether intentional or not, the word has the effect of making the process sound complicated.

Let’s face it, which description would sound better to your parents, your grandparents or the little old lady who sells you your morning paper?

• Say you have a photo of your dog? You just attach the tag word “dog” to the photo by typing it into a box. Then when you want to find it again, you just search for “dog”… that’s what tagging is.

• Say you have a photo of your dog? You just attach the folksonomic category “dog” to the photo by typing it into a box… Then your photo is represented within the folksonomy as depicting a dog…

C’mon people. When we use flickr or delicious we’re “tagging”. The combination of those tags doesn’t lead to anything high-falutin or organised – it leads to a “tag soup”.

Doesn’t that sound more inviting, warming and comforting?

This article first arrived on the interwebnet on Wednesday, March 16th, 2005 at 5:54 pm .
It's arbitrarily categorised as being about: Geeking, Trends and culture.
Supernerds with RSS readers will probably wet their pants over the RSS 2.0 feed.
You can skip to the end and leave a response. We don't do trackbacks here because they're a horrendous spam-trap. Sorry.

This article proudly sponsored by BEER! - BEER! Ask for it by name.