Visualizing inference process of a rule engine

In this paper, we introduce an approach to visualize the inference process in a rule engine -- Drools, which employs Rete as its pattern matching algorithm. As a software visualization work, our approach is focused on both static structure of the Rete network and dynamic behavior of the inference process. Since logic programming is distinct from other traditional programming paradigms, our approach is also different from traditional program/algorithm visualization methods. In this paper, we first introduce the target we choose to visualize, and then provide a description of the problem and our visualization approach. Finally, with an implementation and an interesting case -- sudoku solving, we show that the visualization work is helpful to understanding not only the Rete algorithm, but also the rules used in the inference. Besides, our work supports debugging, tracing and analyzing the rule engine, which is useful in finding errors and optimization.