systematic-debugginglisted

# Systematic Debugging ## Overview The single discipline: find the root cause before writing fix code. Not after. Not simultaneously. Before. The process is deliberately sequential because the most common debugging error is skipping ahead to a fix before understanding the problem. The process applies to every class of defect: logic errors that produce wrong output, performance regressions that make things slow, integration failures where services disagree, flaky tests that pass "most of the time," and data corruption that only shows up downstream. The steps are identical. The evidence bar is identical. There are no shortcuts for "simple" bugs because the bugs that look simple are the ones that get patched incorrectly. ## Process Follow these seven steps in order. Do not skip to step 5 because the fix seems obvious — skipping steps is how wrong fixes get committed with confidence. The useful signal for progress is not the clock but narrowing: if you have cycled through hypotheses several times with no narrowing, you are stuck at step 3 or 4 and should consult the Decision Points section below. ### Step 1: Reproduce reliably **What to do:** Before touching any code, write down the exact steps that trigger the bug. Run them. Confirm the bug appears consistently. **How to do it well:** Create a concrete checklist with these elements: - **Input data:** the exact request body, file, or user action that triggers the bug. - **Environment state:** database contents, configur