Thinking about design patterns for tagging

As tagging is a relatively new phenomenon, it is not surprising that there is not much existing literature yet. But when i was googling yesterday for tagging pattern design it was disappointing though. I didn’t find a collection of examples how adding/removing of tags and other metadata should be designed. So i have to start by myself?

delicious is the most sophisticated tagging machine, as far is i know. So i start with this.

The input field for the tagging metadata is a standard HTML form element, labeled „tags” on the left. On the right, there is a special annotation „space separated” explaining advice how to enter multiple tags. (i remember a blog post of the pros and cons of space separation vs. comma separation a couple of months ago, but i can’t find it anymore).

Below on the left hand side, there is the „save” button which closes the dialog of the four input fields beginning with „url”. The really helpful part of the interface is not part of the dialog itself. The „recommended tags”, „your tags” and „popular tags” areas are not necessarily needed for editing and therefore it is a good idea keeping them outside.

„popular tags” are the most frequently used tags for this bookmark by the folks. „your tags” is the (complete?) tag set of the current user. Own tags matching popular tags are marked green. „recommended tags” aggregate own tags and popular tags in a list.

When using delicious, it is a lot easier clicking on the recommended tags than typing in someting new. Unless you do not have a reason, you would not „invent” a new tag. So what is the sense behind this?

I found a good description in an hp research paper. The aggregation of tags by other users leads to a stable usage of the same tags for a specific item. Users agree on a tag set on the long run, and this agreement can be descibed as „shared knowledge”.