Search

WALT (Washington University's Approach to Lots of Text), is a prototype interface designed to support information retrieval research. The WALT interface serves as a "front end" to a wide array of retrieval engines including those based on Boolean retrieval, latent semantic indexing, term frequency--inverse document frequency, and Bayesian inference techniques. The WALT interface is composed of seven distinct components: a document examination component known as the Document Browsing Area; four navigation components called the Book Shelf, the Book Spine, the Table of Contents, and the Path Clipboard; a term-based information retrieval component called Control Panel; and a relevance feedback component known as the Reader Feedback Panel. WALT's most unique feature may be it's use of "book shelf" and "book spine" metaphors both to facilitate navigation and to provide a histogram-based display showing documents deemed appropriate for answering user queries.

A required course for second year medical students explores an integrated approach to the patient with cancer. Lectures detail the role of each medical discipline in the management of cancer. Panel discussions illustrate the cooperation necessary between clinical specialties in the diagnosis and treatment of these diseases. A small group session provides a problem-solving exercise around a patient with cancer, and laboratory sessions present the opportunity for students to receive hands-on experience in detecting tumor masses in anatomical models as well as the opportunity for students to test their stools for occult blood. Evaluation of the course has shown that students increase significantly their knowledge about cancer, that "practice-effect" is not a factor from pre- to post-test, and that although students improve in their capacity to solve problems they are less proficient in this than retention of facts. Student surveys have indicated generally good acceptance of most aspects of the course.