Autograding and Feedback for Snap!
Michael Ball, Daniel D. Garcia · 2016
While text based languages have been (relatively) easy to grade automatically, visual programming languages have been largely left out of the equation. However, as the demand for introductory CS courses has recently grown, so too has the interest in CS courses that use visual programming languages, and thus the need for a scalable solution to providing student feedback. In this poster we present a system design for an autograder for Snap!, a visual programming language designed at UC Berkeley, inspired by Scratch. Creativity and open-ended responses are an important part of many CS0 courses, such as the Beauty and Joy of Computing, our autograder includes tools that allow instructors to build robust feedback mechanisms that handle a wide variety of student submissions. We also present our initial work to integrate this system into edX for a series of MOOCs as well as a prototype of integrating this tool into a traditional LMS on campus.