EXERCISE 1  EB, Westlaw, and Medline are not the only examples of editorial enhancement. Find another example of an information resource on the Web. Describe the raw corpus and then the editorial enhancement used in this instance. How would you argue that it makes searching more efficient or effective?

EXERCISE 2  Clearly, having each user evaluate whether every document retrieved for every query is relevant is excessive. What approximations to this notion of “correct” answer might be useful?

EXERCISE 3  What other feedback might we reliably and easily be able to elicit? Can users reliably react that a retrieved document is too general or specific? Too theoretic or applied? How could such information be exploited by a learning system?

EXERCISE 4  Many new issues arise when classifications are organized hierarchically. Must all classifications occur only at the leaves? Is a document classified at a leaf an instance only of this subclass or of all ancestor classifications as well? Discuss the consequences to a classifier of each of these questions.