Mining temporal rules for software maintenance

  • Authors:

  • David Lo;Siau-Cheng Khoo;Chao Liu

  • Affiliations:

  • School of Information Systems, Singapore Management University, Singapore;Department of Computer Science, National University of Singapore, Singapore;Microsoft Research, Redmond, WA, U.S.A.

  • Venue:
  • Journal of Software Maintenance and Evolution: Research and Practice - Special Issue on Program Comprehension through Dynamic Analysis (PCODA)
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Software evolution incurs difficulties in program comprehensionand software verification, and hence it increases the cost ofsoftware maintenance. In this study, we propose a novel techniqueto mine from program execution traces a sound and complete set ofstatistically significant temporal rules of arbitrary lengths. Theextracted temporal rules reveal invariants that the programobserves, and will consequently guide developers to understand theprogram behaviors, and facilitate all downstream applications suchas verification and debugging. Different from previous studies thatwere restricted to mining two-event rules (e.g., (lock)→(unlock)), our algorithm discovers rules of arbitrarylengths. In order to facilitate downstream applications, werepresent the mined rules as temporal logic expressions, so thatexisting model checkers or other formal analysis toolkit canreadily consume our mining results. Performance studies onbenchmark data sets and a case study on an industrial system havebeen performed to show the scalability and utility of our approach.We performed case studies on JBoss application server and a buggyconcurrent versions system application, and the result clearlydemonstrates the usefulness of our technique in recoveringunderlying program designs and detecting bugs. Copyright ©2008 John Wiley & Sons, Ltd.This work was done while the author was with School ofComputing, National University of Singapore.