Balancing Autonomy and Persona: Investigating Developer Preferences for Effective Human-Bot Interaction
Date
2023-12-07
Authors
Ghorbani, Amir
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Software bots play a pivotal role in collective software development, promising enhanced productivity. While prior research has highlighted that excessive bot communication can lead to developer irritation, the broader array of human-bot collaboration attributes influencing developer preferences and their consequences remain less clear. This thesis delves into the key characteristics that shape developers' preferences for interactions between humans and bots, focusing on the context of GitHub pull requests.
Employing an exploratory sequential approach, we conducted interviews in Phase I, followed by a vignette-based survey in Phase II. The current thesis primarily reports on Phase II and its findings. A custom-designed vignette-based instrument was employed to survey open-source developers, recruiting participants from third-year software engineering students and the Prolific platform. Rigorous screening procedures ensured data collection from eligible participants only.
The study's results reveal a prevalent inclination among developers towards personable bots that demonstrate limited autonomy. Interestingly, the preferences appear to be influenced by developers' experience levels, with more seasoned developers exhibiting a preference for bots possessing greater autonomy. These empirical insights advocate for bot developers to enhance configurability options, allowing developers and projects to tailor bot behaviors according to individual preferences and project contexts.
Description
Keywords
software engineering, software, bot, interaction, human-bot interaction, github, open-source, survey