A rough stab at an "interest web"

Knowledge maps are one of those knowledge management tools that seems perfect in theory but often ends up as a disappointment or a failure. The idea is to collect all of the experience and skills in an organization into a single map, which can be an online database, a directory, or some other repository. In most maps people rate themselves on the skills they posses. The maps usually take the form of a matrix or table listing skills along one dimension and people along the other. The use for these maps is supposed to arise when someone needs to find an expert in a particular area but doesn’t know anyone directly. They look at the map and find who the experts are.

Unfortunately there are all kinds of problems with such maps. Self-rating is rarely very accurate, experts usually downplay their knowledge while novices will overestimate their skill. The skill lists may not match what people perceive to be their actual skills, but if people submit their own skill names there is a problem with the vocabulary getting out of control. In addition there are the value and incentive problems, just who is this data really good for, management, employees, HR? The initial investment of time to enter the skill information may be larger than the perceived benefit, resulting in an adoption problem.

What I’d like to see built is a social interest map instead of a knowledge map. Change the map into a map of interests instead of skills, and use an interface similar to popular social tagging sites such as delicious. So, for example, a person would login and tag themselves with different interests such as “chess”, “python”, or “web-design.” On another page you would aggregate all the people who have used a tag to describe themselves, thus showing you all the people who declare an interest in chess. People can then browse other profiles and other interest tags.

Some other possible features might be:

  • Let a user can add information feeds to an interest tag page, say from delicious, a blog, or flickr. Then each interest page becomes an aggregator for information about the topic.
  • On each profile pages there could be an aggregator that collects the union of information from all of the users interest tags.
  • Further variations could be to create a recommender system that would look for close neighbors with similar interests.
  • Another variation would be to use some kind of filtering system on the information feeds, letting people rate items or delete feeds.

I think the idea has a lot of merit for use in a particular community that might already have something in common, such as a school, neighborhood, or alumni association.