Design Thinking for Implementation: Leveraging Design Systems for Sustainable and Reproducible Research Software

Scientific research values reproducibility, transparency, and collaboration—principles that should extend to the software that supports it. Yet research scientific software is often developed without systematic design approaches or dedicated UX expertise, leading to tools that may be functional but hard to use, scale, or sustain. At the National Center for Supercomputing Applications (NCSA), we’ve seen how intentional design practices can make a meaningful difference in developing research software. We value design thinking and emphasize that UX design is not merely aesthetic or post hoc, but a critical part of implementation that shapes how users interact with data, models, and tools. Our goal is to build scientific software that is more accessible, sustainable, and supportive of the reproducibility of both user workflows and development processes, and choosing the right design system as a foundation is essential to that process. A well-chosen and well-maintained design system enables faster collaboration, clearer communication, and enhances user experience through more consistent and usable interfaces. In this talk, we introduce how design systems and component libraries—when thoughtfully selected and adapted—can bridge the gap between design and development. We’ll share insights from our work supporting interdisciplinary teams at NCSA, where software must adapt to long project lifecycles, domain-specific needs, and diverse user expertise. We’ll present a practical framework for evaluating and customizing existing design systems in both design (e.g. Figma) and development (e.g. React/Vue libraries) contexts, helping teams make informed decisions based on their goals, user needs, and capacity. We’ll also highlight how early design choices influence not just usability, but reproducibility, maintainability, and cross-team collaboration over time. By integrating design thinking into implementation workflows, research software teams can build more intuitive, trustworthy, and enduring tools—software that, like good science, is built to be tested, shared, and built upon.