TASKGRADER an automated grader to assess code submissions at scale

One of the primary challenges when developing automated grading for large scale MOOCs is avoiding binary student grading and feedback. Feedback which only grades a student as “correct” or “incorrect” can frustrate learners who do not submit a correct program on their first try. Taskgrader allows instructors to grade a student’s submission within an automated grading script, providing students with partial credit for code that they have completed even if their submission is not entirely correct as well as feedback indicating what students may have missed. This partial credit and detailed feedback can help motivate students and assist them in completing complex assignments.

Main characteristics

  • hands-on coding activities.
  • large scale grading system that uses Linux Containers to grade source codes securely.
  • 3 steps: unit testing, source code parsing, performance metrics (Memory, CPU).
  • integrated in Open edX / edX or other LMSs using the LTI protocol.
  • instructor can set up customized feedback for learners based on common errors.

Test it

This tool has been developed by France ioi. You can test this tool by enrolling the C Programming with Linux Professional Certificate.