Bash Scripting: How Efficient Scripts Can Improve Development Productivity; Automation, Error, and Breakdown: An Actor-Network Analysis of the Knight Capital Trading Failure 5 views

My technical report and STS research paper are connected through the common theme of automated software in the context of professional work. In my technical project, I designed Bash scripts that automate the creation of local testing environments for a software firm, allowing developers to reproduce production issues more quickly and reliably. In my STS research, I examine the case of the Knight Capital trading failure to understand how automated software can cause complex systems to break down when human and nonhuman actors are poorly aligned. Although one project focuses on building a reliable system and the other on analyzing a catastrophic failure, both works center on the same topic of automated software systems. My technical project addressed a practical problem faced by a Virginia Beach software firm. Developers were spending a lot of time manually recreating client environments before they could begin troubleshooting reported bugs. To solve this, I created Bash scripts that automatically transformed blank Linux virtual machines (VMs) into mirrors of the production environment. The scripts configured both a database VM and a server VM, installed dependencies from a secure local repository, set up an Oracle database and JBoss server, imported application data, and applied security configurations. Because the project had to run in an air-gapped environment accessible only through the command line, the design required careful adaptation to strict security and infrastructure limitations. The resulting product reduced setup time by over an hour per instance while also improving reliability by reducing human error. My STS research paper also explores an automated system, but one that failed dramatically. Using Actor-Network Theory (ANT) and the concept of translation, I argue that the 2012 Knight Capital trading disaster was not simply caused by a coding mistake. Instead, it resulted from a breakdown in a heterogeneous actor-network that included software, servers, developers, managers, regulatory deadlines, and market pressures. My paper shows that misalignment between human oversight and nonhuman execution created structural vulnerabilities, while competitive pressure and a culture of reactive governance reinforced those weaknesses rather than correcting them. By applying ANT, I reframe the Knight Capital failure as a sociotechnical breakdown rather than an isolated technical malfunction. Working on these projects together strengthened my understanding of engineering as both technical and social. My professional project showed me how much software reliability depends on documentation and careful coordination across many dynamic components. This strengthened my understanding of the technical background for my research paper. My STS research helped me see a similar software project on a larger scale from a broader social perspective by showing how hidden assumptions and weak governance can destabilize an entire software network. Together, these projects reinforced that effective software engineering is not only about writing functional code, but also about building systems where human, technical, and organizational actors remain aligned.

BS (Bachelor of Science)

software; automation; actor-network theory

School of Engineering and Applied Science Bachelor of Science in Computer Science Technical Advisor: Rosanne Vrugtman STS Advisor: Benjamin Laugelli

All rights reserved by the author (no additional license for public reuse)

2026-05-09