Variant Analysis: Set-Based Similarity Visualization for Cloned Software Systems

Software product lines are frequently created using an extractive approach, in which a group of existing software products is reengineered to extract their reusable core. To direct that effort, it is necessary to analyze the reuse potential and the code similarity across the products. We present Variant Analysis, a tool visualizing code similarity across a group of software systems. We represent the systems as intersecting sets of content elements, and place the elements similar between any n systems into the intersection of the respective n sets. Using the resulting set model and the system structure hierarchy, we provide similarity visualizations scaling for tens of compared software systems and millions lines of code. The current Variant Analysis tool analyzes similarity of text files such as source code. However, the underlying models and visualizations can also be used for other types of data, even beyond the software domain.