In the authors’ alternative approach to rule-based systems development, the logic programming language Prolog is used to implement the rules of name placement strategy. They describe the NAMEX logic programming system, which has been used for initial name selection, the implementation of rules governing label positions and configurations, and conflict resolution. The high-level rules are interfaced to a low-level program, which uses primitives to access a specially designed name placement database and low-level name placement algorithms. Three types of name placement primitives are considered in the paper: computation of label position, determination of labels in potential overlap, and label conflict detection.

The strategy for label conflict resolution involves systematic search of possible positions until a solution is found in which no conflicts exist between labels. Due to the vast number of permutations that could result using this “generate and test” approach, heuristics are presented to reduce the investigation of unacceptable label positions. The efficiency of these heuristics is evaluated in order to ensure a rapid convergence of the name placement problem to a solution.

To provide a better understanding of the strategy for conflict resolution, the paper extensively describes an application of the NAMEX system to the creation of road maps based on the Ordnance Survey’s 1:625,000 digital dataset for Great Britain. Due to the ease with which rules can be plugged in or removed using Prolog, other examples shown in the paper include an administrative area map, a geophysical shot point map, and a lunar crater map; the NAMEX system can also handle other forms of annotation, such as for network diagrams and engineering drawings for mechanical designs.

The style is concise, clear, and easy to understand. As a result, readers other than those interested in cartography (to whom the paper is obviously addressed) will find an attractive use of Prolog to solve a difficult CAD problem.