I have been reading for a while various post on tags and bookmarking sites. As far as I can tell there are two main objections to tagging and in particular to the bookmarking sites that rely on them: the first one concerns the folksonomy method of content organization and the other the structure of incentives to add and share tags (think about bookmarking services like delicious, magnolia, et al). Let’s start with the first.
Folksonomy is a non hierarchical organization system based on common usage of words. A taxonomy made by the regular folks like you and me. The main advantage of using tags is that people can organize content in a way that is both personalized as well as constructed on popular and common usage. The problem is that tags might be too loose for an organizational format. Among other things, content ordered by loose tags often has problems of polysemy, synonymy, and depth (specificity) of tagging.
Lets imagine someone tags a specific content with “apple”: are they talking about the fruit, the company or The Beatles record label? In this case we face a problem of polysemy. But even if there are a clear univocal definitions available, different people may use different synonymous to tag content. For example, I might use “car” while someone else use “automobile”. Finally, levels of specificity may very across tags. I might tag something as “technology” while others might tag it “microprocessor”. Other challenges include acronyms as well as plural/singular tags.
The other main issue is the structure of incentives to collect (e.g., bookmark) and tag content. In short, when bookmarking sites become popular, people bookmark and tag content not according to their surfing habits using relevant tags but rather with the purpose of “gaming” the bookmarking system in order to provide relevance to links they want to drive traffic to. This renders the bookmarking system irrelevant.
More to the point, even if people where not gaming the system and even if we could agree on a level of hierarchy so that some sort of pre define structure can be provided to tags, tags and bookmarks still remain sob optimal methods to deal with information overload. At some point, we are going to get more content per tag that we can digest.
Clearly dealing with the challenges tags and bookmarks face can go a long way, but certainly at some point there we are going to face more content that that we could possible digest.
As far as I can see, if we can get tags to deal seamlessly with the challenges I just described then they are bound to add great value in the efforts to deal with information overload. But the story won´t end there. We need a more fundamental approach to the problem in order to provide a more sustainable solution.
