Authors: Fairouz Dahi; Nora Bounour
Addresses: LISCO Laboratory, Department of Computer Science, Badji Mokhtar – Annaba University, P.O. Box 12, 23000 Annaba, Algeria ' LISCO Laboratory, Department of Computer Science, Badji Mokhtar – Annaba University, P.O. Box 12, 23000 Annaba, Algeria
Abstract: The existence of crosscutting concerns scattered or tangled in the source code complicates the software comprehension, maintenance, evolution and reuse. To improve the modularity of software systems, several approaches aimed to identify the crosscutting concerns to allow their modelling by aspects, in order to take advantage of the benefits of the aspect-oriented paradigm. In the absence of early detection of these crosscutting concerns, they tend to be overlooked and become closely linked, not allowing developers to identify an optimal design. This paper proposes a new approach for crosscutting concerns identification at design level. This latter is materialised by class and sequence diagrams. Sequence diagram carries pertinent information which consists of interactions between objects of the software system, and the chronological order of its tasks.
Keywords: aspect mining; crosscutting concerns; optimal design; modelling; reverse engineering; sequence diagrams; UML; source code; software development; aspect-oriented programming; AOP.
International Journal of Computer Applications in Technology, 2016 Vol.54 No.4, pp.334 - 349
Received: 26 Feb 2015
Accepted: 17 Mar 2015
Published online: 17 Nov 2016 *