knuths-optimization-principlelisted

# Knuth's Optimization Principle > "Premature optimization is the root of all evil." > — Donald Knuth, 1974 (often attributed; the full quote is more nuanced) ## The core idea Optimizing code before you know it's a bottleneck is one of the most common and costly mistakes in software development. It makes code harder to read, harder to change, and harder to debug — all in service of a performance problem that may not exist, or that may not be in the place you're optimizing. ## The full Knuth quote The complete passage is important and often missed: > "We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. Yet we should not pass up our opportunities in that critical 3%." The principle is **not** "never optimize." It's: - Don't optimize before you've measured. - Don't optimize the 97% that doesn't matter. - Do optimize the 3% that actually is the bottleneck — carefully and deliberately. ## Why premature optimization is harmful **It optimizes the wrong things.** Human intuition about where code is slow is notoriously unreliable. Profilers routinely reveal that the bottleneck is somewhere nobody expected. Optimizing your hunch is usually wasted effort. **It adds complexity without benefit.** Optimization techniques — caching, pooling, bitwise tricks, manual memory management, loop unrolling — make code harder to read, maintain, and debug. If there's no performance problem to justify them, you've paid the cost