-
Predicting Defective Visual Code Changes in a Multi-Language AAA Video Game Project
Authors:
Kalvin Eng,
Abram Hindle,
Alexander Senchenko
Abstract:
Video game development increasingly relies on using visual programming languages as the primary way to build video game features. The aim of using visual programming is to move game logic into the hands of game designers, who may not be as well versed in textual coding. In this paper, we empirically observe that there are more defect-inducing commits containing visual code than textual code in a A…
▽ More
Video game development increasingly relies on using visual programming languages as the primary way to build video game features. The aim of using visual programming is to move game logic into the hands of game designers, who may not be as well versed in textual coding. In this paper, we empirically observe that there are more defect-inducing commits containing visual code than textual code in a AAA video game project codebase. This indicates that the existing textual code Just-in-Time (JIT) defect prediction models under evaluation by Electronic Arts (EA) may be ineffective as they do not account for changes in visual code. Thus, we focus our research on constructing visual code defect prediction models that encompass visual code metrics and evaluate the models against defect prediction models that use language agnostic features, and textual code metrics. We test our models using features extracted from the historical codebase of a AAA video game project, as well as the historical codebases of 70 open source projects that use textual and visual code. We find that defect prediction models have better performance overall in terms of the area under the ROC curve (AUC), and Mathews Correlation Coefficient (MCC) when incorporating visual code features for projects that contain more commits with visual code than textual code.
△ Less
Submitted 6 September, 2023;
originally announced September 2023.
-
Identifying Defect-Inducing Changes in Visual Code
Authors:
Kalvin Eng,
Abram Hindle,
Alexander Senchenko
Abstract:
Defects, or bugs, often form during software development. Identifying the root cause of defects is essential to improve code quality, evaluate testing methods, and support defect prediction. Examples of defect-inducing changes can be found using the SZZ algorithm to trace the textual history of defect-fixing changes back to the defect-inducing changes that they fix in line-based code. The line-bas…
▽ More
Defects, or bugs, often form during software development. Identifying the root cause of defects is essential to improve code quality, evaluate testing methods, and support defect prediction. Examples of defect-inducing changes can be found using the SZZ algorithm to trace the textual history of defect-fixing changes back to the defect-inducing changes that they fix in line-based code. The line-based approach of the SZZ method is ineffective for visual code that represents source code graphically rather than textually. In this paper we adapt SZZ for visual code and present the "SZZ Visual Code" (SZZ-VC) algorithm, that finds changes in visual code based on the differences of graphical elements rather than differences of lines to detect defect-inducing changes. We validated the algorithm for an industry-made AAA video game and 20 music visual programming defects across 12 open source projects. Our results show that SZZ-VC is feasible for detecting defects in visual code for 3 different visual programming languages.
△ Less
Submitted 6 September, 2023;
originally announced September 2023.
-
SUPERNOVA: Automating Test Selection and Defect Prevention in AAA Video Games Using Risk Based Testing and Machine Learning
Authors:
Alexander Senchenko,
Naomi Patterson,
Hamman Samuel,
Dan Isper
Abstract:
Testing video games is an increasingly difficult task as traditional methods fail to scale with growing software systems. Manual testing is a very labor-intensive process, and therefore quickly becomes cost prohibitive. Using scripts for automated testing is affordable, however scripts are ineffective in non-deterministic environments, and knowing when to run each test is another problem altogethe…
▽ More
Testing video games is an increasingly difficult task as traditional methods fail to scale with growing software systems. Manual testing is a very labor-intensive process, and therefore quickly becomes cost prohibitive. Using scripts for automated testing is affordable, however scripts are ineffective in non-deterministic environments, and knowing when to run each test is another problem altogether. The modern game's complexity, scope, and player expectations are rapidly increasing where quality control is a big portion of the production cost and delivery risk. Reducing this risk and making production happen is a big challenge for the industry currently. To keep production costs realistic up-to and after release, we are focusing on preventive quality assurance tactics alongside testing and data analysis automation. We present SUPERNOVA (Selection of tests and Universal defect Prevention in External Repositories for Novel Objective Verification of software Anomalies), a system responsible for test selection and defect prevention while also functioning as an automation hub. By integrating data analysis functionality with machine and deep learning capability, SUPERNOVA assists quality assurance testers in finding bugs and developers in reducing defects, which improves stability during the production cycle and keeps testing costs under control. The direct impact of this has been observed to be a reduction in 55% or more testing hours for an undisclosed sports game title that has shipped, which was using these test selection optimizations. Furthermore, using risk scores generated by a semi-supervised machine learning model, we are able to detect with 71% precision and 77% recall the probability of a change-list being bug inducing, and provide a detailed breakdown of this inference to developers. These efforts improve workflow and reduce testing hours required on game titles in development.
△ Less
Submitted 28 June, 2023; v1 submitted 9 March, 2022;
originally announced March 2022.
