Mining Early Aspects based on Syntactical and Dependency Analyses

• José M. Conejero ^[1] ; Juan Hernández ^[1] ; Elena Jurado ^[1] ; Klaas van den Berg ^[2]
  1. [1] Universidad de Extremadura

     Universidad de Extremadura

     Badajoz, España

     
  2. [2] University of Twente

     University of Twente

     Países Bajos

     
• Localización: Libro Actas de las Jornadas de Ingeniería del Software y Bases de Datos (JISBD´11), 2011, ISBN 978-84-9749-486-1, págs. 685-686
• Idioma: inglés
• Enlaces
  • Texto completo (pdf)
• Resumen

  • There are different approaches in the requirements engineering community to deal with crosscutting concerns, introducing the benefits of the application of aspect- orientation at these early stages of development. However, these approaches lack a formal definition of crosscutting to be based on. Most of these approaches rely on the use of Natural Language Processing techniques for aspect identification in textual documents and thus, they lack a unified process that generalises its application to other requirements artefacts such as use case diagrams or viewpoints. In this paper, we propose a process for mining early aspects, i.e. identifying crosscutting concerns at the requirements level.First, the aspect mining process is based on a conceptual framework that is independent of specific requirements artefacts and that provides a formal definition of crosscutting based on the trace relations or mappings that exist between two different domains, source and target. The aspect mining process relies on syntactical and dependency analyses at the requirements level that allow the process to be automated since the existing mappings between the two domains are automatically obtained. Second, early aspect refactoring is given for UML use cases diagrams. The refactoring process is driven by a set of concern-oriented metrics that provide modularity assessments (i.e. quantification of the degree of scattering, tangling and crosscutting). Third, our process is both validated by the utilisation of the concern metrics at requirements level and compared with other early aspect mining proposals.