Software Reuse

Software reuse is the practice of using existing software components — libraries, modules, frameworks, or entire systems — in new contexts rather than building from scratch. It is one of the oldest promises of software engineering: write once, use everywhere, compound productivity gains across projects. The practice is now ubiquitous, from open-source package ecosystems like npm and PyPI to internal corporate codebases where modules are shared across product lines.

But reuse carries a hidden systems risk that the productivity narrative obscures. A component designed for one context carries assumptions about its environment — performance characteristics, safety architecture, error handling, trust boundaries — that may not transfer. The Therac-25 disaster was enabled in part by software reused from earlier machines without re-analysis of how the new hardware context invalidated old safety assumptions. The code was not wrong in its original context. It was wrong in its new one.

This is the fundamental epistemic problem of software reuse: the boundaries of a component's validity are not marked in the code. They exist in the history of its design, the constraints of its original deployment, and the culture of its maintenance. When these are lost — through documentation decay, personnel turnover, or organizational fragmentation — reuse becomes a form of institutional amnesia. We do not know why the component works, only that it appears to. And in safety-critical contexts, 'appears to' is not enough.