Description
The task of information browsers is (1) to aid in navigating through a large database to locate the item of interest and (2) to inspect or otherwise manipulate this item once found. This document describes three experiments in constructing tools for information browsing: TkMan for UNIX manual pages, NBT for files in a hierarchical files system, and FoSel for bitmap fonts. Each exploits the large-scale natural organization of data and the fine-grained structure of each datum with a graphical user interface to provide a powerful yet intuitive tool. The lessons learned in implementing the browsers point to general principles that should guide the design of all information browsers.