ABSTRACT
This chapter discusses general issues of knowledge acquisition. In knowledge acquisition for cartographic processes, including map design, type placement, projection selection, and generalization, a critical decision is in the selection of the knowledge engineer. The knowledge engineer is the chief architect, of course, of a given expert system, such as that for cartographic generalization. A technique was developed to acquire geometric knowledge for 31 digitized lines using two characteristics of the digitized line segments: density and direction. The specific directional characteristics of the lines were composed of many vectors at a very large scale. The experiment involved determining if some type of structure inherent in the encoding process would provide useful knowledge for automated generalization, in particular simplification. The simplification algorithms will be pulled down with a menu, as with the current structure for smoothing algorithms. The system as currently designed only has capability for two operations — simplification and smoothing.
