A hybrid program knowledge base for static program analyzers

Static program analyzers (SPA) are interactive tools that enhance program understanding by answering queries about programs. An SPA extracts relevant information from input programs and stores it in a program knowledge base (PKB). In this paper, we present a hybrid PKB design model that integrates a relational database with attributed syntax trees. In the hybrid PKB, global properties of programs are stored in a relational database and detailed program structures are stored as attributed syntax trees. The hybrid PKB approach simplifies the structure of the PKB and provides a flexible mechanism for analysis of complex structured objects such as syntax trees and control/data flow graphs. The model reduces the size of the database, and hence program queries can be answered efficiently