Design of an Automated Pseudo-Code Grading Tool
Patrick M. McDowell · International Journal of Engineering Research and · 2019
The intent of this paper is to present a method for the automated grading of language-independent code segments in computer science classes.The proposed grading tool focuses on the overall structure and logic of the code segment rather than on the syntax or formatting constraints that exist within specific programming languages.For that purpose, when a student inputs code, that code is first processed into a standardized pseudo-code before analysis of the code is performed and a grade is outputted.Currently, a sliding correlation is used to measure the accuracy of the inputted code compared to an "answer key" (key code).The correlation is performed on the entire block of code using condition statements and loops as peaks in the processing phase.The output of the grading tool will yield a value between 0 and 1 that can be converted to a point or percentage value depending on an instructor's preference.