Distribution rule based bad smell detection and refactoring scheme

Bad smells are signs of potential problems in code, which may reduce the design quality of software. Detecting bad smells, however, remains time consuming for software engineers despite proposals on bad smell detection and refactoring tools. Large Class is a kind of bad smells whose specific characteristics are hard to determine, and the detection are hard to achieve automatically. In this paper, a Large Class bad smell detection approach based on class length distribution model and cohesion metrics is proposed. In programs, the lengths of classes are confirmed according to the certain distributions. The class length distribution model is generalized to detect programs after grouping. Meanwhile, cohesion metrics are analyzed for bad smell detection. The bad smell detection experiments of open source programs show that Large Class bad smell can be detected effectively and accurately with this approach, and refactoring scheme can be proposed for design quality improvements of programs.