Research on knowledge-based descriptive cataloging of cartographic publications. An experimental advice-giving system: Mapper

The main objective of this study is to understand how an experimental, semi-automatic cataloging advisor (Mapper) could be designed to assist the intermittent user in the descriptive cataloging of certain U.S. produced single-sheet maps.;The first part of the study uses a multiple-observation approach to identify and analyze instances of expert judgment in solving a set of nontrivial cataloging tasks. Attention is on examining the questions of map authorship and the degree of significance to which various authorial responsibilities participate in the process of map-making.;The study empirically derived a working definition of a "map author" for the maps under investigation. With respect to the degree of significance of responsibilities that participate in the creation of the map, the study has suggested rank order of responsibilities that are represented in the content of the examined maps.;The second part formalizes experts' judgment into a set of rules which were then used to design Mapper's knowledge base. We examine the questions of user modeling and human-computer compatibility to propose a set of user interface characteristics related to Mapper's design. Mapper is implemented using the Apple HyperCard system.;In the third part we evaluated Mapper's performance on accuracy and affective measures in a controlled environment. This study has shown that library students can produce significantly better cataloging answers when helped by the Mapper than those who were not assisted by the Mapper. The term "better" is operationalized in terms of statistically significant higher mean scores on the cataloging tasks under study. Mapper students expressed favorable general opinion of the system and found it easy to understand and use.