Over the course of my doctoral research program from 2023 to 2026, the software industry has undergone a period of profound change with the widespread adoption of AI. Foundational assumptions that have held for decades—how software is constructed by writing code, and how developers come to understand that software by exploring code in IDEs—have both been uprooted during this time. As a software engineer for over 20 years, skilled in this craft, I have personally felt the disorientation. I have also had the privilege of building long-term relationships with many experienced software engineers through my research, giving me a close-up view of how the cognitive experience of developing software has shifted over this period.

While research shows productivity benefits from AI, the developer's experience is also drastically shifting with new agentic workflows and there are open questions as to whether the experience is getting worse [1]. Is the flow state going away? Is fatigue increasing? Are we losing our ability to understand our software? Developers can feel the shifts happening in their experience, but without theoretical models that help us understand or articulate what exactly is changing about the experience, there is little that we can do to respond intelligently.

I'm excited to announce our Theory of Developer Momentum that provides a new theoretical lens for understanding the developer's experience of progress toward the goal of a task and the craft through which progress unfolds. The theory disentangles the related constructs of flow state, momentum, and motivation, so we can more clearly see the relationships between them and how they interweave to support progress. To support the activity of developing software, we need to consider momentum at multiple scales and resolutions without collapsing one into the other—the progress within the activity experienced by the individual, and the project-level goals and coordination with the team. We also need to consider how we sustain our living capacities for care and skilled judgement that support this motion over a longer term. The theory helps us shift from seeing software through a manufacturing lens that centers the outputs of production, to an ecological lens of seeing and supporting software development as a living activity in motion.

Flow state has long been recognized as important for effective software work across Developer Experience, productivity, and satisfaction research. Yet beyond giving developers large blocks of uninterrupted time to focus, we have relatively few ways to operationalize this insight in practice. One of the primary distinctions in momentum theory is the difference between the experience of flow and the experience of progress and momentum. Rather than assuming developers in our study were using Csikszentmihalyi’s definition [2], we made space for developers to characterize their own experiences of flow. This gave us an opportunity to illuminate a richer phenomenology of the developer's experience of flow over time and how it differs from the experience of progress. The developers in our study often described their progress through a task using motion metaphors: going uphill, hitting a wall, feeling sticky and slow, and getting over the hump. They crafted strategies specifically to maintain forward motion in their work across breaks, meetings, and days. By contrast, the notion of being "in a flow state" corresponded to a state of full immersion, where attention was stabilized on the task and competing distractions fell away. By distinguishing momentum from flow, and illuminating the craft developers already use to maintain motion, our theory expands the operational surface area for developers to exercise skilled judgment and make improvements.