In Tag mania sweeps the Web, Jon Udell asks if tagging is ad or a real breakthrough in information management and answers, “both.” I’d have to say, “fad.” Tagging just means adding labels, such as categories, to data. For example, you can label a photo with the names of the people in it as well as the date and the event type. The value is that you can later search on these tags to pull together groups dynamically. In contrast, you otherwise would put your photos into folders–if you didn’t put them in the right folders, they’ll be hard to find or reorganize later.
The problem is that tagging is essentially foldering–it’s just putting the pictures or data into multiple folders. You still have the same problem: if you didn’t choose the right folders or labels, you can’t find the data later. There are many, many, possible labels you can put on photos–if you don’t put them all, you can’t search for them later. Want to pull all of the photos containing someone wearing a red hat? Can’t do it if you didn’t label all of the red hat photos…
The online sharing of tagged information, a la flickr and del.ico.us, is cool because it leverages the efforts of others in a “harmony of the commons” way. But it only works if people use tags consistently. I suspect the apparent value of the new online sites is that they filter so much information that they appear to be successfully organizing information. But do we really know? Yes, we can construct a feed of C# stories. But we have no way to know how many useful C# stories have been missed. There’s value to seeing the C# stories identified as interesting by others, unless you don’t share their interests and they skipped the story you really would want to see.
The danger is that tagging is like a huge flea market. At first, you’re impressed; wow, they have everything! But after a while, you leave disappointed. Wow, there was a lot of other people’s junk. An occasional gem doesn’t make the system efficient. Ultimately, we need systems that find information you need. Tagging may be an short-term proxy that provides human collaboration, messy and inconsistent though it is, to provide some occasional improvement over manual web searches and surfing.