factor-of-safety

• SKILL.mdcore · 13,917 bytes
• references/petroski-and-engineering-history.mddoc · 5,905 bytes

# Factor of Safety

Factor of Safety is the engineering practice of designing systems with capacity exceeding what's strictly required, to absorb unknown variations and prevent failure. A bridge rated for 5,000 lbs is built to hold 25,000 — a safety factor of 5. The "extra" capacity isn't waste; it's insurance against the things you didn't predict: corrosion, fatigue, unusual load combinations, materials that didn't quite meet spec.

For software design, the same logic applies: design with margin in performance, capacity, content handling, error tolerance, and edge cases. The unpredicted always shows up; the question is whether the system absorbs it gracefully or fails loudly.

## Definition (in our own words)

A factor of safety is the ratio of a system's actual capacity to the capacity it was designed to handle under normal use. The size of the factor depends on how much you don't know about the conditions the system will face — the more uncertainty, the larger the factor needs to be. The book's alternative name, "factor of ignorance," is illuminating: a well-understood, well-tested system can have a small safety factor; a novel system facing unknown conditions needs a large one.

## Origins and research lineage

- **Engineering safety practice** going back centuries. The Pyramids of Giza, by some analyses, were over-engineered with safety factors of 20+ — the builders couldn't predict load conditions over millennia, so they built radically conservatively.
- **Henry Petroski**, *To Engineer Is Human: The Role of Failure in Successful Design* (St. Martin's Press, 1985) and *Design Paradigms: Case Histories of Error and Judgment in Engineering* (Cambridge University Press, 1994). Petroski's central thesis: engineering progresses by understanding past failures, and safety factors are the explicit acknowledgment that the next failure will look like something not yet seen.