Systematic literature review on automatic generation of help in programming exercises

Worldwide, computer programming is one of the most demanded skills in the labor market and is an essential component of the curriculum in any university systems engineering program. One way to help students who have difficulties in solving exercises is the generation of automatic hints, which consist of providing personalized hints during the solution process. One of the main challenges associated with the generation of programming hints is the automatic modeling of the solution steps from a large number of correct solutions, due to the diversity of possible solutions that a student can write. The objective of this paper was to present a systematic literature review of existing algorithms for generating automatic hints from a set of correct solutions. This paper concludes that, in spite of the fact that different researches have demonstrated the effectiveness of this type of hints, their massive employability is just beginning to be implemented in Latin American universities.