Author: | Schacherbauer, Martin Hubert |
|---|---|
| Supervisor: | Prof. Gudrun Klinker |
| Advisor: | Dyrda, Daniel (@ga67gub) |
| Submission Date: | [created] |
Abstract
In order to increase correct representation and inclusion of disabled people this paper addresses the current state of video games. We found that the most represented group of people in games is still the White, able-bodied, young man. In order to remind developers that there are other people than that aforementioned person, we analyzed the game development process and found that by hiring with accessibility and equity in mind, teams get more diverse and in turn more voices in a team get represented. Additionally, practices like the game development document or agile processes can be expanded to also take into account accessibility and representation. We also developed tools that can be used with GitHub, which ensure that changes in code, if relevant, need to be approved before being put into the game.
Tool Descriptions
While the thesis tackles the whole game engineering process and also investigates how a diverse workforce can be gained and retained in the game development industry, the main focus lies on the practical application of increasing diversity, equity, and inclusion (DEI) awareness of developers. In this section, the developed tools will be presented shortly.
Labeller puts, as the name suggests, labels to commits. In a configuration file developers can set which file changes lead to which label to be attached to the commit. The files can be both included or excluded by specific factors, such as combinations of folders, specific file names, or file extensions. The labels themselves do not have a big functionality except for filtering purposes in GitHub.
Building on the Labeller, the Automatic Request Reviewer (ARR) goes a step further. Just like with the Labeller, the user can set folders or files which trigger the execution of the ARR. When a merge commit to the main branch is made and the requirements for a review are fulfilled, the developer can not merge on the main branch without the review of an (or multiple) assigned reviewer(s). This is made to ensure that changes or commits that change files, for instance, related to the controls of the game, they need to be reviewed for their accessibility before they get pushed to the main branch.
Both of those tools, however, do only work on the micro level on the project. For the macro level of the project it is important to evaluate the project as a whole. Therefore, an automatic issuing system was used. Whenever there is a merge commit to main there is a random chance for an issue to be created. This issue can follow certain templates and can contain instructions about what needs to be evaluated.
Conclusion
The aim of this thesis was to help developers to have a better understand of representation of diverse people as well as increase the accessibility of their games. This is done in two ways. Whereas the engineering process itself could be adapted to make room for DEI in development, it is also possible for tools during the production phase to ensure certain goals that have been set in pre-production or by the company in their own guidelines for DEI. Furthermore, the adaptations of already commonly used techniques in the games industry could help to increase the awareness of DEI not just among development teams that are already dealing with DEI, but also among teams that have not yet thought about DEI in and for their games.
While there are always additional ways to include more steps to improve DEI awareness in a development process, it has to be said that tools that analyze diversity based on characters and categories they fall into still need more research. In "How to measure diversity when you must.", Budescu and Budescu talk about the different ways of measuring diversity with scores, however, the title itself suggests that measuring diversity has to be done with care and only be done if there are no other ways. Future research should be made in order to find out how to do that for games, especially. It may be better to measure diversity by having a list of traits that make people unique and how that is represented in the game. However, this might also lead to negative results and therefore needs research.