Static Code Analysis for Software Quality Improvement: a Case Study in Bci Framework Development

摘要

This paper shows how the systematic approach in software testing using static code analysis method can be used for improving the software quality of a BCI framework. The method is best performed during the development phase of framework programs. In the proposed approach, we evaluate several software metrics which are based on the principles of object oriented design. Since such method is depending on the underlying programming language, we describe the method in term of C++ language programming whereas the Qt platform is also currently being used. One of the most important metric is so called software complexity. Applying the software complexity calculation using both McCabe and Halstead method for the BCI framework which consists of two important types of BCI, those are SSVEP and P300, we found that there are two classes in the framework which have very complex and prone to violation of cohesion principle in OOP. The other metrics are fit the criteria of the proposed framework aspects, such as: MPC is less than 20; average complexity is around value of 5; and the maximum depth is below 10 blocks. Such variables are considered very important when further developing the BCI framework in the future.