A new design defects classification: marrying detection and correction

  • Authors:

  • Rim Mahouachi;Marouane Kessentini;Khaled Ghedira

  • Affiliations:

  • SOIE, University of Tunis, Tunisia;CS, Missouri University of Science and Technology;SOIE, University of Tunis, Tunisia

  • Venue:
  • FASE'12 Proceedings of the 15th international conference on Fundamental Approaches to Software Engineering
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Previous work classify design defects based on symptoms (long methods, large classes, long parameter lists, etc.), and treat separately detection and correction steps. This paper introduces a new classification of defects using correction possibilities. Thus, correcting different code fragments appending to specific defect category need, approximately, the same refactoring operations to apply. To this end, we use genetic programming to generate new form of classification rules combining detection and correction steps. We report the results of our validation using different open-source systems. Our proposal achieved high precision and recall correction scores.