An Architecture for the Automated Assessment of Web Programming Tasks

Abstract

Automatically assessing students' solutions to programming tasks in the domain of web programming requires special means due to the characteristics of web applications. This paper proposes an architecture for a web-based learning application tailored to this domain. For the implementation of the automated assessment of programming tasks, we make use of end-to-end testing and container virtualization. This allows, in contrast to other popular approaches, the coverage of tasks that include special operations like DOM manipulations, which alter the user interface of web applications, in a way that is convenient for both students and instructors. We demonstrate the capabilities and limitations of the architecture based on two common usage scenarios.