The other day I was lying in the bath thinking about the use of enforced textual hierarchies in tagging (like most people do in the bath I’m sure…). Tagging is a method for assigning words (or sometimes phrases) to an object. Most often users are presented with an empty box within which to type these words in which are then associated with that object. My favourite example of the use of tagging is del.icio.us, a social bookmarking site. The beauty of this style of “categorising” is that the user is left to assign to it what they want and not what a committee has decided are the categories.

However what can sometimes happen is a hierarchy starts to develop. An example would be getting users to prepend customers with a key word so those relating to “Joe Bloggs Inc.” would look like this: “client_JoeBloggsInc”. This is very helpful if you are trying to scan through a load of data and if you’ve ordered a list alphabetically you’ll get all the clients grouped together. However, what you’ve done is reduce the opportunities for unintended but positive consequences.

You want to make it so that stumbling across the object you are tagging is as likely as possible. That someone who doesn’t know your self-created hierarchy can locate it with minimum effort so that they can add value to it.