A Formal Concept Analysis-based Statement-level Aspect Mining Method

The topic of this paper is a kind of aspect mining method, aspectmining means identify candidate aspect in existing OOP systems andisolating them from the system into separately descried aspects. The goalof migrating an OOP system into an aspect-oriented one includes, amongstothers, improving the comprehensibility of the system, and therebyimproving it's maintainability and extensibility. Current aspect miningtechnologies usually focus on mining of method codes in classes. Theymake great efforts of restructure software systems. However, they cannotdirectly solve the two statement-level problems that are important inaspect-oriented programming named "code tangling "and "codespreading". In order to solve these problems, this paper presents a formalconcept analysis-based statement-level aspect mining method. The methoduses formal concept analysis to identify concerns and achieves astatement-level aspect mining. This method has many features, such asautomation, statement-level and high efficiency. It can be used in rapidreconstruction of legacy systems in an aspect-oriented style. At last, thismethod is validated by a prototype in the paper.