Automated Assessment in Mobile Programming Courses: Leveraging GitHub Classroom and Flutter for Enhanced Student Outcomes

Pedro Alves, Bruno Pereira Cipriano · 2025

Read Automated Assessment in Mobile Programming Courses: Leveraging GitHub Classroom and Flutter for Enhanced Student Outcomes by Pedro Alves, Bruno Pereira Cipriano (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-04230.txt.

Introduction

Pedro Alves and Bruno Cipriano explore, in a 2025 paper, auto grading for mobile app courses with GitHub Classroom and the Flutter frame. Demand for app builders keeps such courses core in computing degrees. Yet app work is hard to grade by hand: set-ups are complex and apps are visual and live. Their path joins GitHub Classroom for task hand-out with GitHub Actions for checks that run on each turn-in. Flutter's strong test tools then grade the app's acts on their own. That mix aims at a low-cost auto grade that still tests what counts.

The team ran the path in a real mobile course as a trial. A student survey read how the class took to the model. Marks were sound, the team reports, and students took to the model well. The paper gives steps for like courses to copy.

Catalog record

CREATORPedro Alves, Bruno Pereira Cipriano
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE55 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-04230

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Automated Assessment in Mobile Programming Courses: Leveraging GitHub Classroom and Flutter for Enhanced Student Outcomes Pedro Alves, Bruno Pereira Cipriano Abstract The growing demand for skilled mobile developers has made mobile programming courses an essential component of computer science curricula. However, these courses face unique challenges due to the complexity of mobile development environments and the graphical, interactive nature of mobile applications. This paper explores the potential of using GitHub Classroom, combined with the Flutter framework, for the automated assessment of mobile programming assignments. By leveraging GitHub Actions for continuous integration and Flutter’s robust support for test automation, the proposed approach enables an auto-grading cost-effective solution. We evaluate the feasibility of integrating these tools through an experiment in a Mobile Programming course and present findings from a student survey that assesses their perceptions of the proposed evaluation model. The results are encouraging, showing that the approach is well-received by students. keywords mobile programming, automated assessment, github classroom, flutter, integration testing 1 Introduction The increasing prevalence of mobile applications has led to a growing demand for skilled mobile developers. Consequently, mobile programming courses have become a crucial component of computer science curricula. However, these courses present unique challenges due to the inherent complexity of mobile development, stemming not only from the intricate design and implementation of mobile applications ( Francese et al., (2015) ) but also from the need to set up a complex development environment (IDEs, SDKs, emulators, etc…) ( Modesti, (2021) ) . The difficulty of automating the build and testing processes further complicates the teaching and assessment of mobile programming. Traditional methods of code validation, such as output matching and unit tests, are often inadequate for mobile applications, which are predominantly graphical and interactive. For example, to test the navigation within the application, tests must be executed within a graphical emulator which typically take a long time to run and are difficult to automate. This complexity is magnified in advanced courses where students must develop applications that interface with external sources such as remote APIs, sensors, and local databases ( Paiva et al., (2022) ) . Automated assessment tools 1 1 1 Also known as auto-graders (AATs) have been successfully implemented in many programming courses to increase student autonomy and motivation ( Enström et al., (2011) ) while reducing the teacher’s workload ( Ihantola et al., (2010) ) . These tools can validate the correctness of student submissions and, in some cases, also assess its code quality ( Cipriano et al., (2022) ; Heckman & King, (2018) ) . Even though most AATs were designed for simpler programming exercises ( Cipriano et al., (2024) ) , some are well-prepared for project-based assignments, involving multiple files and integration with Git ( Cipriano et al., (2022) ) . However, the application of AATs in mobile programming courses remains limited due to the aforementioned complexities. The graphical nature of mobile applications and the intricate connections between the user interface (UI) and the underlying model pose significant challenges for automated testing. As a result, there is a notable gap in the research and practice of using AATs for mobile programming education. In recent years, the advent of multi-platform mobile development frameworks such as React Native 2 2 2 https://reactnative.dev/ and Flutter 3 3 3 https://flutter.dev/ has provided new opportunities for overcoming these challenges. Flutter, in particular, offers robust support for test automation, including both unit, widget and integration (instrumented) tests that can easily be executed from the command line. In particular, widget tests enable UI testing without the need for an emulator. This capabilities suggests that it might be feasible to implement automated assessment in mobile programming courses using Flutter. Moreover, by leveraging GitHub Classroom to manage student projects and GitHub Actions to automate the execution of tests, educators can create a more streamlined and effective assessment process. This approach not only prepares students for modern development practices but also enhances their learning experience by enabling rapid feedback through features like hot-reload. Given these developments, this study seeks to explore the potential of using Flutter and GitHub Classroom for automated assessment in mobile programming courses. Specifically, it addresses the following research questions: • RQ1: To what extent can Flutter, GitHub Classroom and GitHub Actions be integrated to support the automated assessment of mobile programming assignments? • RQ2: What are the students’ perceptions of the proposed evaluation model? This study makes the following contributions: • Presents a novel approach for assessing programming assignments focused on developing mobile applications • Presents the results of a student survey evaluating the proposed assessment method This paper is structured as follows. Section 1 introduces the challenges of mobile programming education and the need for automated assessment tools. Section 2 provides background on mobile development, testing frameworks, and the use of GitHub Classroom for managing and assessing programming assignments. Section 3 reviews related work on automated assessment in mobile programming. Section 4 presents the ”awesome quotes” exercise, outlining its academic context, design, test implementation, automation process, and the challenges faced. Section 5 summarizes the results of the student assessment and survey, highlighting their perceptions of the model. Finally, Section 6 concludes with a discussion of the findings, the model’s effectiveness, and potential improvements for future research. 2 Background 2.1 Mobile development Currently, developing mobile applications involves targeting the two dominant operating systems (OSs): iOS and Android, which together account for over 99% of the market share ( Hayat, (2024) ) . When these OSs first emerged, the only viable model was native development, which required using the native SDK of each OS. For iOS, this involves developing the application in Objective-C or Swift, and the developer must have a computer running macOS since the SDK only runs on this operating system. For Android, this involves developing in Java or Kotlin, but the SDK runs on any operating system (Windows, macOS, or Linux). Given the significant differences between the SDKs, the only solution is to develop two completely separate applications, one for iOS and another for Android, with little opportunity for code reuse, resulting in increased costs. Consequently, some companies opt to initially develop only for one OS. Meanwhile, three alternative models have emerged that allow development for both OSs from a single codebase: the so-called cross-platform models. The simplest model is to develop the application in HTML5, which allows responsive screens to be designed using CSS3, adapting to different screen sizes to ensure good usability even on smaller devices. Additionally, HTML5 capabilities have evolved, providing limited access to sensors, geo-location, local storage, etc. 4 4 4 See https://whatwebcando.today/ . The main advantage of this model is its simplicity and ease of learning. However, it requires connectivity when the application is launched and, most importantly, does not allow the application to be installed from app stores, which users are accustomed to and expect. Next, hybrid models emerged: models that allow the creation of installable applications for both OSs from a single codebase using an intermediate language. These hybrid models are divided into two types: hybrid-web and hybrid-native models. Hybrid-web models, with the Ionic 5 5 5 https://ionicframework.com/ framework being the most popular representative, are essentially a ”trick” because although they are installable applications, they are actually web applications embedded

Read the complete text