Since 2003, we have strived to determine the most appropriate method for improving students' programming skills. One reason for this research is that the evaluation of programs requires sufficient time and careful consideration of correctness, fairness, and appropriateness. Another reason is that programming should always be evaluated on the grounds of not only compilation and testing but also correctness of the algorithm and validity of the model. Due to various factors, even if students submit their programs electronically, it does not contribute to improving the correctness of the evaluation in terms of marking their programs. Moreover, a lack of concrete assessment items and criteria for the programs implies that the marked results will depend on the instructor's assessment. To reduce the extent to which assessment results depend on the instructor, we propose the introduction of subject design in our programming exercises. We have developed a marking support tool that enables us to evaluate programs on one screen where we can readily compile and test the program, browse the code, record the student's marks, and analyze assessment results. This research sought to improve the assessment of programming exercises using support design and the marking support tool that we developed.