Towards a semantic comparison system of source code for support in learning programming

Meta-cognitive skills and problem solving skills are weak among learners in learning programming and trainers focus much more on the syntax of the languages. These skills are fundamental in learning programming and are one of the causes of dropout, but they can be developed during learning. Several strategies can be used to develop these skills, including feedback on code semantics. We propose a semantic source code comparison system that can be used for semantic feedback in learning programming. This system is based on the properties of instructions and through a process of initialization of the semantic value of the input data and chaining/substitution variables in the instructions, and a process of formal computation, determines the final semantic value of a computer program.