But what are “tags”? At the most basic level, they’re just keywords — categorization without a master table of categories. Categorization that people can make up on the fly without having to “create” the tag in advance or fit it into some taxonomy. It’s categorization from the bottom up, exactly like I theorized on here but never implemented.
What’s the drawback? Mainly, that tags won’t get standardized. I may tag my articles with “automotive” while you may tag them with “cars.” Or that you may mispell a tag name once in a while and not notice it.
But this apparently doesn’t matter so much to anyone. Tagging is the Next Big Thing, even though it’s a technical step backwards from the button-downed, tightly normalized way of doing things.
I think it’s refreshing. It’s nice to let our a deep breath and just use something that works without having to make Poindexter the Perfect Programmer happy.
Jon Udell gave a great screencast about a month ago about del.icio.us and how he uses it. It’s worth watching if you haven’t grasped the significance of the tagging phenomenon. Towards the end, he gets into the “soft” topic of how languages are formed and how tagging is a way of essentially writing another language which isn’t perfect but comes about though informal usage, negotiation, and acceptance. (This is technically called a Creole.)
For those more interested in the technical ramifications of all this, here’s a page with links to resources on how to make SQL understand tag concepts, regardless of application.