Research On The Methods Of Java Code Clone Detecting

The successful large-scale systems are called as legacy systems. These systems have tremendous business value. But cause of the gigantic volumes and lacking of documents, the systems are not maintainable. One of the problems confused the software engineers is the code clone. A large-scale system may contain more than 10% codes which are similar. The code clone take more than 70% cost during the software maintenance.The code clones are the same or the similar code fragments. During the maintaining of software systems, the code clones may bring many problems. The dissertation expatiates on the reasons for producing code clones and hidden trouble of them.As the usage purposes are diverse, different tools use different standards to classify the code clones. After summarizing the classifications and the detecting methods of code clones, the dissertation divides the code clones into five classes, divides the detecting methods into four classes from a new perspective. One code clone set would have many classes of code clone. Different classes of code clone should use different methods to detect them.After expounding the detecting methods based on string,metrics,PDG, The dissertation also proposes the concepts and theories concerning the methods and introduces the design schemes and implement methods of them.