Simplifying Software Design

Simplifying Software Design: The Genius of Bureaucracies, or How Not-My-Job Sharpens Your Design addresses a deceptively simple question at the heart of software development:

• Where should this line of code go?

Every system is the accumulation of answers to that question. Each decision about responsibility, knowledge, and communication shapes whether a system remains understandable, maintainable, and adaptable-or becomes tangled and fragile.

In this book, international software design guru Dr Alistair Cockburn offers a practical answer using an unexpected metaphor: design your software the way you would design a bureaucracy!

While bureaucracies are often criticized, they excel at defining clear responsibilities, limiting who needs to know what, and structuring communication between roles. These same properties are essential for effective software design.

Two familiar human reflexes become precise design tools:

• "Not my job" - keeps responsibilities sharply defined
• "No need to know" - limits unnecessary dependencies

Using these principles, developers can decide where behavior belongs, reduce coupling, and create systems that remain clear as they grow.

Just as relevant at this time, they question arises: "How do we train AI systems to produce maintainable code?" It turns out that AI have read all our complaints about bureaucracies and understand "Not my job" and "No need to know" very well. These questions make for excellent dialog with them.

Building on responsibility-driven design (Ward Cunningham, Kent Beck, Rebecca Wirfs-Brock), the book presents a compact starter kit for software design. It introduces responsibility statements, scenario-based evaluation, and interaction diagrams as practical techniques for exploring and improving designs.

Cockburn also presents six design tests-Abstraction, Responsibility Alignment, Evolution,Communication Patterns, Data Connectedness, Data Variations-to help compare alternatives and reason about how a system will respond to change. Rather than defining a single "correct" design, the book focuses on how different designs support different futures.

These ideas are connected to familiar architectural styles, including Model-View-Controller and hexagonal architecture (ports and adapters), showing how responsibility and boundary management apply at every scale-from subsystems to individual classes.

Written as part of the Simplifying series, this book focuses on the small set of techniques that deliver the greatest practical value. It is accessible to newcomers while offering a clear framework for experienced developers and architects.

In an era of AI-assisted programming, where code can be generated rapidly, clear boundaries and responsibilities remain essential. These techniques help teams-and the tools they use-produce systems that are modular, understandable, and resilient over time.