2020/12/02 by Ivona Bezáková, Kimberly Fluet, Bezáková, Ivona +7
Computer Science · #Teaching and Learning Programming #Software Engineering Techniques and Practices #Software Engineering Research
paper · pdf · doi:10.48550/arxiv.2012.01546
Computing theory analyzes abstract computational models to rigorously study\nthe computational difficulty of various problems. Introductory computing theory\ncan be challenging for undergraduate students, and the main goal of our\nresearch is to help students learn these computational models. The most common\npedagogical tool for interacting with these models is the Java Formal Languages\nand Automata Package (JFLAP). We developed a JFLAP server extension, which\naccepts homework submissions from students, evaluates the submission as correct\nor incorrect, and provides a witness string when the submission is incorrect.\nOur extension currently provides witness feedback for deterministic finite\nautomata, nondeterministic finite automata, regular expressions, context-free\ngrammars, and pushdown automata.\n In Fall 2019, we ran a preliminary investigation on two sections (Control and\nStudy) of the required undergraduate course Introduction to Computer Science\nTheory. The Study section used our extension for five targeted homework\nquestions, and the Control section solved and submitted these problems using\ntraditional means. Our results show that on these five questions, the Study\nsection performed better on average than the Control section. Moreover, the\nStudy section persisted in submitting attempts until correct, and from this\nfinding, our preliminary conclusion is that minimal (not detailed or\ngrade-based) witness feedback helps students to truly learn the concepts. We\ndescribe the results that support this conclusion as well as a related\nhypothesis conjecturing that with witness feedback and unlimited number of\nsubmissions, partial credit is both unnecessary and ineffective.\n