Model Checking to Improve Precision of Design Pattern Instances Identification in OO Systems

Abstract

In the last two decades some methods and tools have been proposed to identify the Design Pattern (DP) instances implemented in an existing Object Oriented (OO) software system. This allows to know which OO components are involved in each DP instance. Such a knowledge is useful to better understand the system thus reducing the effort to modify and evolve it. The results obtained by the existing methods and tools can suffer a lack of completeness or precision due to the presence of false positive/negative. Model Checking (MC) algorithms can be used to improve the precision of DP's instances detected by a tool by automatically refining the results it produces. In this paper a MC based technique is defined and applied to the results of an existing DPs mining tool, called Design Pattern Finder (DPF), to improve the precision by verifying automatically the DPs instances it detects. To verify and assess the feasibility and the effectiveness of the proposed technique, we carried out a case study where it was applied on some open source OO systems. The results showed that the proposed technique allowed to improve the precision of the DPs instances detected by the DPF tool.