I posted earlier on a meeting of KM Forum, “To Tag or Not to Tag! Should We be Structuring our Knowledge Assets or is Free-text Search Enough?” Here are my notes from the session. There were a number of experts in taxonomy development in the audience and I learned a great deal. The term “tagging” was used for categorization for this session and not for social book marking as in del.icio.us. Of course, the multiple meanings of the word tagging points out some of the problems with tagging. We hope to have tagging as social book marking to be the topic of a future session.
This session was organized around several questions and these were the main one:
1. Can you fully exploit Knowledge Assets in the form of documents without categorizing them? – The answer is that categorization is usually needed. If the contents of a category get too large or if the content is very small, they are less useful. Inside the firewall a controlled vocabulary helps with this categorization but it needs to be continuously refined.
2. How best to give context to content? We said that this is best done by aligning the context with the goals and processes of the organization so it supports work flows. This is something we did in the first knowledge management system I was involved with in the early 90s. In this case we aligned the content with the insurance underwriting and claims processes we were supporting.
3. Is automatic categorizing working for some (or all) of your content? Several people experienced with these tools said they are getting better but still not ready for prime time in many areas. They do get you started but they people still have to do a lot of refinement. They work best in certain well defined fields with a specialized vocabulary. They also work best with hard facts like “age” and less well with inferences.
4. Does taxonomy or navigated search have a place in your organization? Is it available and used? The answer is again yes. Navigated search usually refers to guided search where you are offered alternatives to select from. Siderean Software is one vendor with Seamark Navigator.
5. When is key-word, free-text or Google style searching good enough? Of course, the answer it depends. If you find what you need the answer is yes. Google does have a lot of advanced features that allow you to narrow your search. They are generally not used by most people, including me.
6. How generally pays for categorization and taxonomy development within an organization. This effort is generally seen as “soft” and it is often difficult to get it funded and difficult to determine who should pay for it.
Denham Grey also provided some useful comments on classification and the issues in we discussed that go beyond what I mention above. He said that tagging (in the social book marking sense) “often fits best where the content is user owned and generated (blogs, images), a bottom up emergent categorization works when the group and domain is small (<150 people), strongly bounded and the participants can annotate, edit and refactor the content directly (without a strong editorial process).” There is more in his post, Classification - does it work? enterprise tagging
Bill, CIO magazine just posted well rounded summary of enterprise tagging including IBM dogear. FYI
http://www.cio.com/archive/040106/et_main.html?action=print
Posted by: Tomoaki Sawada | April 04, 2006 at 09:48 PM