Tuesday, January 30, 2007

Jump-starting a folksonomic community?

Joshua Porter at Bokardo keeps the taxonomic vs folksonomic discussion going. He raises this point about folksonomies that I have been giving a lot of thought to of late:

Folksonomies suffer from the Cold Start Problem (CSP). You have to build up tagging datasets over time, so at the beginning there is really no navigation to build on top of them to speak of.
I've been wondering if there aren't some strategies for jump-starting a folksonomy? I was considering a transitional strategy that might include creating a concordance for each piece of content, and then using that concordance to suggest tags. Here's a concordance that I grabbed from Amazon for the book Social Network Analysis: A Handbook. If an entire book can be condensed into this single graphic then I‘m guessing most content would produce a far smaller set of choices to guide people.


My idea is that as people begin to build up their own set of tags you could gradually wean them from the “suggested“ words. I‘m not sure what the algorithm looks like to determine the optimal timing, but that‘s the beauty of models, you‘re allowed to tweak them until you get it right.

So, just like I asked the question for examples of working taxonomies for dealing with the deluge, I‘m now asking for anyone‘s experience in transitioning a “community” who have been using a taxonomy, which is on the verge of collapsing under its own weight, to a folksonomy. I‘m looking for anyone‘s experience in dealing with the CSP. Like I said, I‘ve been giving this a lot of thought and failure is not an option.

I‘ll look forward to hearing some suggestions, and thank you Joshua for keeping the conversation going.

No comments: