Understanding comment practices in Scratch: A study of comments in a block-based visual programming language

Comments are vital for software documentation. They provide necessary insights and assist developers in understanding and maintaining the software. Due to their importance, comments have been extensively studied, and much has been learned about them. These existing studies have predominantly focused on text-based languages. Conversely, block-based visual programming languages, particularly Scratch, are becoming increasingly popular. Some studies regarding comments related to the Scratch online community focus on topics such as fostering online community and engagement, sentiment analysis, etc. However, they overlook the visual aspects and the qualitative analysis of comments within code in Scratch projects. This is a meaningful limitation, and this research project studies comments and their pattern in Scratch projects from both textual and visual perspectives. We examined comments collected from different Scratch projects. Each comment was manually annotated based on textual and visual attributes, producing a taxonomy model of comments for a visual programming language. The classification results were analyzed to understand better the practice of commenting in Scratch. Our result revealed that Scratch projects produced noisier(i.e., less understandable) comments than text-based programming languages like Java. In addition, the study also revealed several limitations and shortcomings that could be addressed to improve the commenting experience in Scratch.