Component-based experimental testbed to faciltiate code clone detection research

摘要

Over the past few decades clone detection has become a major area of study among software researchers and software practitioners. Clone detection experiments present a number of challenges such as accurate data collection, data cleaning and selection of proper code detection algorithms. This urges the need for a systematic and unambiguous approach to conduct clone detection experiments. As a solution, this paper presents an experimental testbed, which consists of a collection of “clone detection components (CDCs)”. CDCs are concrete representations of tasks associated with clone detection experiments such as data extraction, pre-processing and detection of clones. These CDCs could be used in isolation to represent a simple task or could be composed to represent a complex task. The usefulness of the experimental testbed is evaluated with an important clone detection experiment for three open source projects.