King's College Researcher Challenges Foundational Model of Software Abstraction

Stephen Kell, a researcher in the Software Systems group at King's College London's Department of Informatics, has published a research essay arguing that the foundational methods used in computer science and language implementation are fundamentally limited. He suggests that the industry relies too heavily on a specific, narrow approach to how software abstractions are realized.

In his essay, "Reversing abstractions: an existential crisis," published on the humprog.org blog, Kell asserts that computer science is built upon what he terms "forwards" and "existential" realizations of abstractions. According to Kell, the standard theory of abstraction follows a rigid sequence: a developer first selects a desired abstraction—such as a specific programming language or a data type—and then determines the necessary realization, which includes technical details like calling conventions and memory layouts.

The Tension of Realization

This top-down approach is a cornerstone of modern software engineering, where high-level intent dictates low-level implementation. Kell's work, part of a broader series of "Rambles around computer science," contrasts this common "existential" approach with a "universal" or "reversed" perspective. By examining the tension between these two models, Kell questions whether the current paradigm of picking an abstraction and then figuring out how to make it work is the only viable path for system design.

Implications for Infrastructure

The shift toward "reversing" this process matters because the current reliance on existential realizations may constrain the development of more efficient or flexible programming infrastructure. If developers are bound by the traditional top-down model, they may overlook alternative ways to implement languages and systems. Kell suggests that by reversing the flow—moving from the realization back toward the abstraction—engineers might discover new architectural possibilities that are not bound by existing conventions.

Future Directions

Kell's critique invites a broader conversation on how language designers approach the creation of software. While the standard model remains the industry norm, the proposal to move toward universal abstractions suggests a potential path for evolving how the next generation of programming tools is conceived. It remains to be seen how these theoretical shifts will translate into practical, real-world implementation tools, but the challenge to the status quo highlights a critical juncture in the evolution of software systems.