“Software architecture is the study of the structures that guide the organization of software components and subsystems into complete systems.”
Mary Shaw wrote this definition in a 1998 position paper about computing with coalitions of distributed resources. By then, she had helped establish software architecture as a distinct level of design concerned with systems larger than individual algorithms or modules. The sentence centers structure: how components are selected, connected, constrained, and composed so that the whole system has useful properties. Code quality inside a component matters, but it cannot by itself guarantee that services communicate reliably, failures remain contained, performance meets demand, or future changes can be made without destabilizing everything around them.
The word “guide” distinguishes architecture from a static diagram. A box-and-arrow picture can document names and connections while saying little about the rules governing them. Architectural structures express decisions about communication, ownership, state, deployment, trust, and change. They make some behaviors easier and others harder. A layered system may constrain which components can call one another; an event-driven system may separate producers from consumers while adding challenges in ordering and observability. Architecture is therefore a set of consequential design commitments, not merely the visual summary created after implementation.
Shaw’s paper considered coalitions of resources that were independently created, controlled, and sometimes transient—conditions that now sound familiar in cloud and platform ecosystems. A team may depend on external APIs, open-source libraries, managed services, and data sources it does not own. The complete system still has to operate when a dependency changes, becomes slow, or disappears. Architectural work identifies those boundaries and creates mechanisms for discovery, substitution, monitoring, and degraded operation. It also acknowledges organizational reality: ownership and incentives influence technical structure because components are maintained by people and institutions with different goals.
The definition remains useful in an era of microservices and AI components, where decomposing a system can be mistaken for designing one. More components create more interfaces, operational states, and failure paths. Architecture earns its value by making those relationships coherent and aligning them with qualities that matter to users: security, accessibility, reliability, performance, and evolvability. It should be reviewed against running evidence, not protected as an original blueprint. Shaw’s sentence reminds teams that a complete system is more than a collection. It is an organized whole whose structure determines what can be understood, trusted, and changed. Architecture records are useful when they preserve the reasons for a decision, enabling future maintainers to revisit the tradeoff instead of treating old structure as unexplained law forever.
Shaw used the definition in a 1998 paper about systems formed from distributed, independently controlled resources. The paper examined architectural challenges beyond closed, centrally managed software.
Her broader research helped establish software architecture as the design level concerned with system organization, interactions, and qualities rather than only local code.
Architecture guides how components communicate, share state, fail, and evolve. It is expressed through constraints and responsibilities, not merely through a diagram of boxes.
No architecture guarantees quality by itself. The chosen structure must be evaluated against requirements and revised when operational evidence shows that its tradeoffs no longer fit.
Cloud applications depend on many services and organizations outside a single team’s control. Their interfaces and failure modes determine the behavior of the complete product.
AI and microservice systems add components rapidly, making coherent boundaries, observability, ownership, and degraded operation more important than decomposition for its own sake.
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.
