An Informal Information-Seeking Environment

When an opportunistic searcher encounters an over-deor ‘‘evolving,’’ ‘‘contracting’’ and ‘‘expanding’’—is termined information system, less than ideal search available, then searching is potentially richer because the strategies often ensue. The mismatch can be addressed chances improve that the searcher will be prompted to by reducing the determinacy of the system, thereby makreformulate and abandon plans, backtrack to points of ing it more amenable to informal problem-solving practask deferral or significant decision, and compare things tices. This claim is investigated by designing an information-seeking environment, where search techniques are side-by-side. When aspects of the search process are exrepresented with a data-flow notation and where the ternalized, both prospective memory, plans for the future, searcher has control of layout; thus, to seek information, and retrospective memory, tasks completed, are supis to manage space. Search plans can be represented ported. In programming environments, support for these on the display, and perceptual cues about search progbasic cognitive tasks is considered essential because they ress are captured while searching. With elements of search activity visible, opportunistic problem-solving allow people to work the way they want to—opportunisshould be well supported. The interface is designed to tically (Visser, 1994). There is no a priori reason to think be extensible so a wide range of search techniques can they are any less important in environments for informabe represented, and emphasizes search material, such tion seeking; indeed, planning, backtracking, and comparas queries, results, and notes, rather than system conison are at the heart of search tactics and strategies (Bates, trols. 1979a, 1979b). When assessing the match between users and informaIntroduction tion systems, reviewers often point to learning difficulties and working memory failures, technical incompatibilities, A major characteristic of any problem solving situation and large and complex feature sets. In a penetrating analyis the ‘‘conversation’’ that the user has with his or her sis, Bates (1990) argues that the goal should not be to materials, whether computerized or not. Schön (1983) replace the searcher with a knowledged-based system, but characterizes how pencil and paper support design this rather to design the interface to support the strategic, way: opportunistic behavior of searchers, and argues that there