Although there's been much talk about the evolution of KM and the new ways of collaborating--through social networking, virtual worlds, and the like--it strikes me as interesting that, not only have some long-running KM issues not been solved, but they've resurfaced as important topics. From data points like our monthly APQC KM community calls, request for quotes we receive, and recent conversations with APQC members, four issues seem to stand out: taxonomy, expertise location, best practices, and lessons learned.
Recently, I learned of some taxonomy development projects at a few large-scale organizations, including one very significant software provider. The fact that these organizations are still considering developing enterprise taxonomy at a time when everyone seems to just "google" everything is a bit surprising to me. What, if anything, does this say about folksonomies?
Interest in expertise location at a time when many "acknowledged" experts may be leaving the work force is not a surprise. However, many organizations seem to be revisiting issues such as how to determine who is an expert, what constitutes expertise, and how that expertise is to be used, particularly in view of the ease of access to all "expertise."
Best practices? There doesn't seem to be an end to the number of organizations struggling to define best practices, make them available, and finally use them as part of a learning cycle. The same thing applies to lessons learned. Some time ago, a friend and colleague of mine, Wes Vestal, wrote about lessons learned often being nothing more than "lessons captured" and certainly a long way from the "lessons applied" that they should be. I would argue that most organizations are still stuck on "lessons documented and forgotten."
If there's any commonality among the issues above, I think that at least three of them (and arguably expertise location as well) focus on explicit, rather than tacit, knowledge. That is, "document it and they will come." I leave it at that.

The reason that companies continue to invest in building Enterprise taxonomies is that there is a great value in having controlled vocabularies. In regards to Taxonomies and Folksonomies in the enterprise, i like to advocate a hybrid approach that is proving to be successful in many organizations.
I recently published an ebook with Dow Jones that helps start the conversation about that approach that you can read more and download from here:
http://danielabarbosa.blogspot.com/2008/06/folksonomies-and-taxonomies-in.html