Autograding Overview¶
Autograding is an essential part of any scalable computation-centered course. It allows instructors to easily and efficiently grade swathes of student submissions, and also lets students have a way to check their progress and ensure they’re heading the right direction. There are many different options for autograding depending on the scale and requirements of your course, from those that run entirely on the instructor’s machine to those that utilize a server that grades students’ assignments on submission.
Autograding, as it is done at Berkeley, originated with the development of OkPy, a Python autograder originally designed for CS 61A that was later retrofitted with the tools to grade Jupyter Notebooks for use in Data 8. Since OkPy, Berkeley has released a few different autograding solutions based on the different requirements of certain course offerings, including Gofer Grader, which is used in the EdX MOOC offering of Data 8, and Otter Grader, a scalable solution that supports both local and server-based autograding.