Dynamic Analysis and Association Rules for Aspects Identification

Esteban S. AbaitSantiago A. VidalClaudia A. Marcos

One of the main problems with current object­oriented systems is the presence of crosscutting concerns, which makes software comprehension and modularization difficult. Aspect mining is the task of identifying the crosscutting concerns present in legacy systems. In this paper, a semi­ automatic four­step approach for aspect mining is proposed based on dynamic analysis and association rules. In a preliminary evaluation in two systems, the proposed approach successfully discovered a number of crosscutting concerns (candidate aspects) already identified for these systems by existing approaches.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: