agentic-rob-pike

Premature optimization happens when engineers improve code speed before measuring actual bottlenecks. Avoid it by enforcing a measurement-first approach: instrument the codebase, gather data, and assess trade-offs before attempting any algorithm selection or performance tuning.

Decide when to optimize by applying rule-based guidelines that require concrete measurement data first. Collect metrics from code instrumentation to assess performance trade-offs, ensuring optimization actions target verified bottlenecks rather than speculative inefficiencies.

Prevent premature optimization during code reviews and architectural planning by enforcing a measurement-first mindset. Require engineers to instrument code and present performance data before approving any optimization decisions or algorithm changes.

This measurement-first approach does not rely on external tooling beyond standard measurement practices. It provides clear instructions for activation and enforces instrumentation using standard practices, requiring no specific dependencies or external frameworks.

The limitation of a measure-first workflow is that it may delay immediate performance fixes until instrumentation and data collection are complete. Engineers must balance the discipline of gathering measurement data against urgent bottleneck resolution requirements.