Tags are the latest social computing effort sweeping the blogosphere. The idea is great; with each piece of information you share you apply a simple text tag to it. For instance, if you're uploading pictures of your trip to the Luxor Hotel in Las Vegas you might use the luxor or luxor hotel tag. Neat idea, but doomed to fail.

There were two articles today that made me think about tags and the future of tags. The first was from Paolo Valdemarin and the second from John Dvorak. Each of these articles examines the future of tagging.

Paolo's article talks about the nebulous metaphor of tagging. He talks about a couple of examples of what he thought were logical tags only to stumble upon other tags. He holds Wikipedia as a success in the tagging effort.

John's article is more scorched earth. He talks about the future of tags that are spam ridden and useless. He also ties it into the accepted definition for tagging which is "public graffiti."

Both articles point out the issues with tagging as it exists today. Paolo's article looks at the openness of defining tags and the loss of importance that follows. John's looks at the negatives that will happen with tagging.

Both are right.

Tagging at this stage is based on a community's ability to support it and police itself. Right now the disparate tagging services take different approaches to each responsibility. Flickr handles it one way, while Technorati handles it another.

For tagging to be successful the following needs to occur:

Centralized tag definition: There needs to be one organization responsible to the taxonomy of tags. If I post something with a FIREFOX tag does that mean it relates to the whole browser or was I too lazy to enter FIREFOX HACKS. The DMOZ open directory project uses specific sections under which to classify websites and it works well. Paolo's comment concerning Wikipedia is right on in that they centralize the tags for their system.

Centralized tag maintenance: Along with centralized definition you need centralized maintenance. There needs to be a way to move information that clearly doesn't belong under a certain tag to another tag. The best way to do this is through peer review. When a piece of information is submitted the group has the chance to suggest changes to the tag and after a few days the movement of the information is automated.

Global acceptance: This is the killer. This removes the ability for people to free form add their own tags to the main taxonomy, which flies in the face of the tagging philosophy right now. Also, it's critical that the taxonomy be agreed upon by all vendors who adopt tagging.

The history of libraries is a good place to start in looking for a solution. Melvil Dewey had this same exact issue in the 19th century and came up with a solution that was adopted by most of the world. That solution is known as the Dewey Decimal System.

There are two issues with adopting a system like this:

1. It's based on numbers: Yep. So are IP addresses, but that's why we have DNS. Once the infrastructure is fleshed out the software used to interface with tags can handle the same translation through a service call. The end user would only see FIREFOX and not 561.44.4 or whatever the numbering system shakes out to be.
2. It has to be centralized: Yep. Covered above, but once the initial taxonomy is fleshed out the addition of new tags shouldn't be an issue.

Are tags worth saving? I think so, but only if there is a coherent, centralized control mechanism.