“The purpose of software engineering is to control complexity, not to create it.”
Pamela Zave’s line appears in Jon Bentley’s 1986 Programming Pearls and has endured because it gives software engineering a demanding, practical test. A system does not become engineered merely because it uses sophisticated tools, formal processes, or an impressive architecture. Engineering earns its place when it makes a difficult problem easier to reason about, verify, operate, and change. Software inevitably represents complicated domains, but its structure should prevent that necessary difficulty from spreading everywhere. The quote asks designers to notice whether each new layer clarifies responsibility or simply gives the team another set of rules to remember.
Controlling complexity is largely the work of creating boundaries that match the problem. A useful module gathers related decisions and exposes a smaller, stable interface. A clear data model makes important constraints visible. Consistent error handling prevents every caller from inventing its own recovery policy. Tests, types, documentation, and observability each reduce uncertainty in a different way. None of these practices is valuable merely because it is present. Their value lies in how much reasoning they remove from routine work. When developers can make a local change with confidence, the design has converted a large system into a sequence of manageable questions.
The warning against creating complexity is not an argument for the fewest files, the shortest program, or the simplest-looking diagram. A compact solution can conceal assumptions, couple unrelated concerns, or make failure behavior impossible to see. Conversely, an additional abstraction may lower total complexity if it captures a concept that otherwise leaks through dozens of places. The relevant measure is the burden placed on the people who must understand consequences. Teams should be suspicious of duplicated policy, exceptions that bypass normal paths, overlapping frameworks, and indirection that has no clear owner. Simplicity is not visual tidiness; it is predictable behavior supported by coherent choices.
The principle matters even more in systems assembled from cloud services, third-party packages, generated code, and machine-learning components. These tools can reduce the effort of producing features while increasing the number of interactions no one fully understands. Zave’s standard suggests evaluating additions by the complexity they retire as well as the capability they deliver. Does a platform consolidate operational knowledge or scatter it? Does automation make a decision inspectable or merely faster? Does an AI-generated change fit the system’s existing concepts? Software engineering remains successful when growth is accompanied by stronger explanations, safer boundaries, and fewer surprises. The goal is not to make the world uncomplicated; it is to keep the software from making an already complicated world harder to manage.
The line was attributed to Pamela Zave in Jon Bentley’s 1986 Programming Pearls, where concise engineering lessons were presented through practical examples.
It defines software engineering by its effect on understanding: tools and process matter only when they make a difficult system easier to reason about and change.
Controlling complexity means gathering decisions behind clear boundaries, exposing stable interfaces, and reducing the facts required for routine work.
It does not mean minimizing files or abstractions; an extra structure can lower total complexity when it captures a concept that would otherwise leak across the system.
Cloud services, packages, AI-generated code, and machine learning make it easy to add capability while multiplying interactions and hidden assumptions.
Zave’s standard asks whether each addition retires uncertainty, strengthens ownership, and makes consequences easier for the next engineer to predict.
Explore more "Quotes of The Day"
Discover more notable quotes from influential voices across politics, science, business, technology, sports, and culture. Each quote offers insight into how ideas, beliefs, and decisions shape the world around us.
