Intelligent grading of programming exercises in a MOOC

MOOC or massive open online courses are very popular now. For instance, UCL has thirteen courses in edX platform in different area. In this thesis we will focus on the course Louv1.1x, Paradigms of Computer Programming – Fundamentals. And in particular, we will focus on exercises of this course. This exercise will be implement in Oz language, because this language has three fundamental paradigms that will be taught, namely functional, object oriented and dataflow concurrent programming. The teacher has no direct contact with students because they are too many, then we want to create a tool to help student to understand their errors without help of teacher. So, in this thesis, we will create an intelligent grader to this course. This grader will have as goal to show to user the execution, in details, of its program. Oz kernel language allows showing easily this thank to this abstract machine. The second goal will be to give feedback on a particular exercise regarding its source code. So the global goal is to improve the learning by Internet. The tool created here is able to understand a part of Oz kernel language with concurrency and to display all execution of a program. Also it can propose exercises for students, and to compare their version. Of course it able to detect errors in the provided source code and inform the user about that.