I think I don’t have to explain to anyone what tags are, or? But to make sure, Wikipedia states that tags allow anyone on the Web to annotate and categorize content and then roam around and discover similarly tagged content. Sounds like a blast to a lot of people, all one has to do is find some content one finds interesting (a lucky search will do), tag it accordingly and then sit back and enjoy the thrill of discovery via his or her RSS reader.
Now, suppose we stumble upon one of those Microsoft-Yahoo merger articles (well, you couldn’t avoid running into one even if you wanted to) and want to tag it. It’s surely about.. Microsoft.. and Yahoo.. and business deals/mergers/acquisitions.. the “Future of the Web as we know it”.. then technology.. media.. ads.. revenues.. oh and the ubiquitous Google.. I just crawled these from 5 blog entries or so, if I took a week off and went through the other thousand, I’m sure the number of tags used on this one subject would reach a century.
What is then the problem with tagging? Is that different people use different words for the same topics? Or that everyone categorizes differently? Or that people read in different topics into stories? Obviously it’s all of the above and this is what makes tagging, a practice targeted to be used by communities, so hard to actually employ this way. There is a great post by Ms. Rashmi Sinha on the cognitive processes behind concept building and tagging, which shows that it’s a very tough task to actually employ tags that would make the tagged content be easily discovered by others.
So what are the possible solutions to this here mess? The “Tagging” blog describes the approach of ‘tag description’ where users can make a note for themselves, what they actually understand under a selected tag. “Tag gardening” is a great definition for this approach, and I assume that most of you won’t be to keen on having to spend hours on growing your favorite tag plants. Another attempt to make tags more generalized is the idea of semantic tagging, e.g. as provided by Zigtag. I’m sure it will improve the tagging experience by a country mile, but it still doesn’t reduce the number of tags nor the number of words we can theoretically use.
Another solution would be to actually extract the main topics from a piece of content so that the story is naturally categorized. Then, depending on if you want to read a story about just one of these topics or explore in similar topics, you can either search directly or use natural language semantic relationships to find similar stuff. This might take the social component out of inventing tags that are understood by a possibly large part of the community, but it would bring in more related content discovery and allow people with similar thoughts to find each other and start talking, rather than tagging. Or?