Improving automated programming assessments: user experience evaluation using fast-generator

Abstract

Automatic Programming Assessment (APA) has been known as a method used to automatically mark and grade students’ programming solutions. In order to realise APA as a tangible deliverable, a number of automated tools which are called Automated Programming Assessment Systems (APAS) have been developed and tested for decades. Basically, the need for decreasing the load of work among lecturers, timely feedback to students and accuracy on the grading results are the common reasons that motivate the need for APAS. In order to carry out a dynamic testing in APA, it is necessary to prepare an appropriate and adequate set of test data to judge the correctness quality of students’ programming solutions in terms of the aspects of functional and/or structural testing. Manual preparation of quality test data becomes a hard, time consuming, and feasible task in the practice of both software testing and APA. Thus, the generation of automated test data is highly desirable to alleviate the humans’ burden from performing repetitive tasks. This paper aims to describe the design, implementation and user experience when evaluating a tool developed to support APA as a test data generator that is called FaSt-generator. The tool plays an important role to furnish a test set that includes an adequate set of test data to execute both the functional and structural testing in APA. Results collected from the conducted user experience evaluation using FaSt-generator reveal that all the subjects had relatively positive opinions and greatly favour the criteria of User Perception and End-User Computing Satisfaction (EUCS).