Of course users have to understand what a controlled vocabulary is and to understand taxonomies before using them. However, this is a ‘one-off’ cost. Once the user has understood the need for such tools (via an information literacy session, for instance) they are better placed to take advantage of them. The price of resource discovery for the user henceforth declines.

Controlled vocabularies (or taxonomies) are tools; a means to an end. For a tool to be useful, one has to understand how it operates in order to take advantage of what it offers. Tagging harbours few rules and therefore its use as a ‘tool’ is limited. Indeed, in stark contrast to a controlled vocabulary, an information literacy session with tagging would not enable a user to improve his/her chances of discovering relevant resources, since the rules of discovery are not predictable enough nor are they learnable. Tagging therefore contrasts with controlled vocabularies by imposing a ‘perpetual discovery cost’ on the user, rather than a ‘one-off’ cost. That is not to say that tagging can be useful in particular contexts, but high precision searching will never be its forte.

It is true that many users lack the skills necessary to take advantage of advanced searching / browsing options. However, the exponential growth in information literacy instruction and information literacy research reflects the international need for users to understand these tools and, as the IL literature reveals, understanding such tools need not be long or particularly arduous. There are many dedicated librarians and information professionals working extremely hard to ensure their user groups have the best IL skills possible.

It’s also worth noting that Ian’s contention is not merely theoretical, as Jim suggests, but is actually borne out by much empirical evidence and research within the fields of the library, information and computing sciences.

The problem is that discovery of information with professionally classified content in large information retreival systems (or even small ones) is not necessarily made easier for the average user. At LexisNexis we see only a small percentage of users interacting directly with our taxonomies. These are mostly information professionals (librarians, info brokers, researchers, etc). Why is that? The average user doesn’t know what a controlled vocabulary is and how it can help, even if we put it under their noses. The learning curve to understand taxonomies in IR is high.

How do you present a complex classification system to the average searcher so he or she can interact and use it effectively without much (or any) thought?

This is where the information science community, in my opinion, has failed to innovate and where IAs must excel. Take a look at any large online library catelogue, for instance. Terrible interaction and usability (usually). You need training (e.g. bibliographic instruction) and practice in order to understand it. In other words, the price of discovery of classified information is high for the end user. The contention that it is not is theoretical. Or do you have a practical or concrete example of easy-to-use taxonomies or controlled vocabularies?

informat [...]