I have posted the Folksonomy presentation delivered 30 November 2005 at the Online Information Conference (PDF at 1.3MB). This presentation was part of the New approaches to making taxonomies work panel.
The main focus of this presentation is not that folksonomies should be seen as a replacement to taxonomy, but as a means to augment taxonomies (if there is one in place). As was resoundingly echoed by others on the panel, taxonomies are hard work and expensive to build and maintain. The cost and effort are often reasons why taxonomies are not exhaustive nor emergent, as budgets and time constraints provide limits. Most often we follow the Pareto Principle (also know as the 80/20 rule) where we focus on 80 percent of the use with 20 percent of the resources (in reality we aim toward something more like a 90/40 rule), but we do have limitations. Taxonomies are also authoritative, but this is problematic for the people who have a vocabulary that is different than the authoritative vocabulary(or more correctly vocabularies). This means a taxonomy will most often have a limited view, which is not a reason to stop taxonomies, but a reason to augment them.
I also point out tagging as it was done in the past provided no certainty as to the vocabulary being correct in the perspective of all trying to use it. Many people found tagging to often be misleading. Tagging in and of its self is not the solution, but part of the solution.
In folksonomies we have three distinct data elements: the tag, a clear understanding of the object being tagged, and an identity of the person doing the tagging (this may be a real or cloaked identity). These three points allow real people to tag information and objects for their own recall. This is the most important aspect as it provides clarity that one's own vocabulary and use of the terms can provide. All people have authority over their own term choice if it is for their own recollection. The provision of identity allows us to give meaning to the individual tagging, which is not a property that general tagging provided prior. Being able to discern the person and their tags on distinct information and objects provides a great advantage as it the person and us to find community. The person can easily find other people who have tagged the same object with the same tag and through this they can find people of like mind and interest. This is an incredibly valuable tool.
This also helps us viewing from the outside find communities and usage of terms that are valid for individuals. We can use this to query real people as to their vocabulary choice as well as validate our taxonomy for use and identify gaps. This is all done with relatively little cost, particularly when comparing to the effort and cost of maintaining and updating a taxonomy.
This presentation is the one I presented, which was why it took a little longer to post than usual as it was not the presentation I still had on my hard drive. I had edited the presentation for clarity, highlight some points, and for the purpose of audience listening and not reading. The edited presentation did not make it into the presentation queue, so I presented from the unedited version.