<![CDATA[Clean, well-structured code is one competency demanded by Indonesia's National Work Competency Standard for vocational programming graduates, yet coding classes commonly still focus only on whether a program runs, without giving automatic feedback on code quality. This study examines the effectiveness of CLING (Clean Learning Integrated Gateway), a website-based Python compiler integrated with Pylint for automatic clean code assessment. The method used was Research and Development with the ADDIE model, combined with a Quasi-Experimental Nonequivalent Control Group Design involving tenth-grade Software Engineering students at SMK Antartika 2 Sidoarjo: an experimental class (25 students) taught with CLING and a control class (25 students) taught with Google Colab without linter integration. Material and media expert validation scored 88.2% and 87.7% (Very Valid). The Shapiro-Wilk test showed the control class posttest data were not normally distributed (p = 0.002), so the Mann-Whitney U Test was used, yielding U = 604.0 (p < 0.001), indicating a significant difference in clean code skills between the two classes. The experimental class posttest average (65.77) was far higher than the control class (32.51), with an N-Gain of 0.488 (moderate category). The Technology Acceptance Model questionnaire recorded a practicality score of 87.55% (Very Practical). These results indicate that CLING effectively improves vocational students' clean code skills. ]]>
Copyrights © 2026
