Abstract:

The framework of aspect-oriented software reverse engineering is proposed to solve the comprehension and evolution problems of crosscutting attributes in the software legacy system. An approach to the formal concept analysis driven by use case is then presented based on the proposed framework.with the aim of c0nverselv recovering the crosscutting attributes on the level of requirement models.With this approach.the execution profiles of the objective system are analyzed using concept lattices.thus enabling the identification and extraction of the invoked sets of the crosscutting system use case models.Finally. the crosscutting methods are abstracted into the earlv aspects of a software system.As compared with the aspect-mining techniques. which are mainly applied to the code reconstruction on the program code level,the presented approach is more suitable for the program comprehension on a higher abstract level. Moreover,as a kind of semiautomatic approach with the strategy of incremental resuIt extraction,the proposed approach is more controllable and adaptive when the requirements in the reverse engineering are changed.

Key words: Reverse Engineering, crosscutting concern, extraction, form al concept analysis, Unified Modeling Language