When integrating performance considerations into code standards, which approach is recommended?

• A. Prematurely optimize all code paths.
• B. Ignore performance until after deployment.
• C. Favor clear, efficient algorithms, measure performance where it matters, and avoid premature optimization.
• D. Use only microbenchmarks after every change.

Answer: (C)

Focusing on performance means building on solid, clear algorithms first and measuring how they actually behave in real use. The best approach is to favor clear, efficient algorithms, measure performance where it matters, and avoid premature optimization. Start with correct, maintainable code using straightforward ideas. Choose data structures and algorithms that fit the problem and typical data sizes, rather than chasing clever optimizations that may not help in practice. Then profile the program on representative workloads to find real bottlenecks—the places where time or memory use actually hurts users. Only optimize those hot paths, and do so with maintainability and correctness in mind. This ensures improvements are meaningful and scalable, instead of spending energy on speculative tweaks that add complexity.

Prematurely optimizing every path tends to complicate the codebase and wastes effort without guaranteed gains. Ignoring performance until after deployment risks unaddressed bottlenecks turning into user-visible problems. Relying on microbenchmarks after every change can mislead, because small, isolated measurements don’t reflect how the system behaves under real usage, with all its interactions and workloads.