Mark Marron

Talk: Making Software (Artifacts) Boring

Date and Time

April 8, 2025 at 12:30pm

Location

PLSE Lab (CSE2 253)

Abstract

Creating software to solve a problem or address a business need is a creative and dynamic process. If all goes well the outcome is a dependable program that meets a user’s needs and addresses the target problem. However, traditional software development struggles with the creation of stable, low-maintenance (boring) artifacts – leading to persistent bugs, security vulnerabilities, dependency tacking/management effort, and operational overhead. Thus, instead of a stable solution to a problem, each of these artifacts becomes an ongoing maintenance task!

The Bosque project addresses this challenge by advocating for a fundamental shift in the development paradigm, aiming to produce “boring software” characterized by high assurance and predictable behavior. This talk will outline Bosque’s research agenda, which centers on four core areas: the design and implementation of a correct-by-construction programming language, the development of advanced static and dynamic analysis tools based on the language design, the formulation of a practical theory for modular software construction, and the creation of a stable and resilient runtime environment. This talk will cover the project’s vision, current research findings, and future research trajectories.

Bio

Mark Marron is a faculty member at the University of Kentucky, specializing in programming languages and software engineering. Prior to that he spent 10 years at Microsoft Research, consulted on data-integrity and formal-compliance issues at Morgan Stanley, and received a PhD from the University of New Mexico. His work includes training large-language model coding assistants (Copilot), data-quality specifications, programming system design and implementation (Bosque), automating CI merges (Edge/Chromium), creating an extreme low-overhead time-travel debugger for Node.js, and program synthesizers for NLP-to-code (Excel).