Calibration of source-code similarity detection tools for objective comparisons

Today there are many source-code similarity detection tools. These tools are used for many purposes and one of them is plagiarism detection, in which context this paper is written. Every time a new tool is developed authors want to show that it is better than existing ones, and so they perform comparisons. Often these comparisons tend to be unfair towards the existing tools, for which there can be multiple reasons, such as the lack of calibration of existing tools. Almost all tools have configuration parameters, but often they are not calibrated before the comparison. The paper presents a way of calibrating the tools to keep the comparison more objective.