In many cases, taxonomies are an after-thought, added to applications to supply the necessary metadata or grafted into Web sites to bring order to the intranet chaos. But what if users took charge of taxonomy development? What if creating and assigning categories were almost as easy as sending an e-mail message? What if creating cross references were as easy as adding a link to a Web page?

User-driven taxonomy development is already happening, as indicated by the growing popularity of "social tagging" on Web sites like del.icio.us and Flickr. In this article we look at how the changing needs of Montague Institute staff, members, and course participants have driven the development of such tools as indexes, search functions, and subject hierarchies — all based on an expandable metadata repository structure.

The foundation: metadata storage
User-driven taxonomies are not feasible without a good way to store metadata. Typically, each source of information (e.g. a customer relationship system, a business intelligence database, an archive of sales proposals) has its own organization and format. Most are designed for a single, stand-alone purpose, not to be part of an integrated system. As a remedy, the IT function may try to create a standardized taxonomy for the whole enterprise — an effort that often turns out to be too complex, time consuming, and costly (see "Upstream knowledge management" and "Two taxonomies are better than one").

The most practical solution, according to many experts, is a metadata repository — a detailed catalog that describes what metadata is available, where it's located, and the relationship among data components (see "Quick study: enterprise information integration"). We have had a metadata repository for 10 years. In the examples below, we look at some of the ways we've used this basic taxonomy structure to serve the needs of members, staff, course participants, and the public.

Exposing metadata
Instead of creating the ideal metadata repository and then plugging it into applications, our taxonomy structure evolved gradually in tandem with application development based on our needs at the time (see "Taxonomy milestones: lessons learned"). Once the basic taxonomy structure is created and populated with metadata, its terms, categories, attributes, and relationships can be "exposed" or displayed to users in an endless variety of ways. In our case, this means database (fielded) search, A - Z index, site search, table of contents, and various lists (e.g. top 20 articles).

For internal staff, the most important of these tools is the database search (see below). That's because our staff members are familiar with the substance and format of our content and are looking for speed and precision. They know which fields to search and how to combine field values to get exactly what they want. They aren't casual browsers, although they occasionally use both the A - Z index and site search.

More ... (members only) How to become a member

Created on October 10, 2005