This brief, optional pre-conference session is for novices who want to learn the step-by-step fundamentals of how to build a taxonomy. Taxonomy Boot Camp sessions assume a basic familiarity with taxonomy construction.

One of the biggest challenges of any taxonomy project is actually “getting” the taxonomy. Can you buy an existing taxonomy? Do you have to build one from the ground up? Or can you use an automatic categorizer to create one for you? This lively debate explores what you need to make the decision to buy or build your taxonomy, or if automatic taxonomy generation will work for you, as each of the speakers presents and defends an option.

While having their own complexities beyond fielded metadata, taxonomies need to be approached as part of your full metadata set. This session will focus on how taxonomies fit into an overall metadata strategy, including how to identify different types of metadata for your organization, how to manage content and data modeling issues for fielded metadata, and how to identify and implement work flow to handle both standard metadata and taxonomies.

Learn how automatic metadata generation solutions can be used cost-efficiently and effectively for large-scale projects without a major investment in human resources. In this case study, Wolters Kluwer, a major publisher, will illustrate how it has applied Teragram’s automatic metadata solution to its vast repositories of published materials.

Continue the day’s discussions with new colleagues and old friends over drinks and hors d’oeuvres. Meet and talk with the speakers and the conference sponsors. Then wrap up the day, if you wish, by joining a Networking Dinner Group for a fun and casual evening at a nearby restaurant. (Watch the Taxonomy Boot Camp Web site for signup forms for these Dutch-treat dinners.)

Your taxonomy will not be perfect or complete and will need to be modified based on changing content, user needs, and other practical considerations. Developing a taxonomy incrementally requires measuring how well it is working in order to plan how to modify it. In this session, you will learn qualitative and quantitative taxonomy testing methods including:

• Tagging representative content to see if it works and determining how much content is good enough for validation
• Card-sorting, use-based scenario testing, and focus groups to determine if the taxonomy makes sense to your target audiences and to provide clues about how to fix it
• Benchmarks and metrics to evaluate usability test results, identify coverage gaps, and provide guidance for changes.

One major taxonomy challenge is how to easily integrate a taxonomy into multiple content management and content delivery environments in a way that is responsive, scalable, and cost-effective. Using an XML solution, Richard Beatch has implemented a solution at Allstate that has one stored taxonomy that can be consumed by all affected applications. The result is vastly improved turnaround times and ease-of-use for taxonomy changes and maintenance.

Once you’ve developed a taxonomy, the work has only begun. Keeping it current, meaningful, and accurate over time as business needs evolve is an ongoing effort. Hear best practices and innovative approaches for maintaining your carefully constructed taxonomy. As a result, you will learn how to derive maximum value from your information assets and keep end users coming back for more. A case study from Harvard Business School illustrates these best practices in action.

Faceted navigation for information collections is gaining wide acceptance. However, a considerable impediment to the wider adoption of faceted interfaces is the creation of the faceted hierarchies and the assignments of terms from the hierarchies to the information items. Marti Hearst and her colleague, Emilia Stoica have designed an algorithm called Castanet that semiautomatically
generates hierarchical faceted metadata from textual description of items. Using an existing lexical database (such as WordNet), the algorithm carves out a structure that reflects the contents of the target information collection. Learn how the algorithm has been successfully applied to collections as diverse as recipes, biomedical journal titles, and art history image descriptions. The resulting category hierarchies require only small adjustments to achieve intuitive results with good coverage.

As illustrated in the prior session, the rapid emergence of social tagging has had a ripple effect in corporate settings due to the potential positive impact on the cost and quality of human-generated metadata. This closing session discusses forward-thinking work on systems that are beginning to merge social tagging with more formal approaches to metadata creation and management. The goal is to allow folksonomies to make taxonomies more responsive to change, while allowing taxonomies to make folksonomies more responsible in the context of information governance.