rseng-lessons-learned

# Lessons learned and project memory Every project pays tuition - the bug that took three days, the approach that silently failed, the review comment made for the fourth time. The tuition is only wasted if the lesson evaporates. This skill is the capture mechanism: write the lesson down at the moment it is learned, in the repository, and then - the step most teams skip - route it into an artifact that makes repetition impossible. A lessons file nobody acts on is a diary; a lesson that became a test is institutional memory. ## Capture at the trigger moments The lesson is freshest and cheapest at the moment of resolution. Capture proactively - offer, in one line, to record it - when: - A nontrivial bug is fixed: what the symptom was, what the cause turned out to be, what evidence found it, what would have prevented it (rseng-debugging's evidence trail is the draft). - A code review surfaces a recurring pattern: the third time the same class of comment appears, it is a lesson (and a candidate lint rule or checklist line - rseng-pair-programming). - A research approach is abandoned: WHY it lost - data, method, numerics, scaling - so the next person (or the same person, next year) does not re-walk the dead end. Negative results are project knowledge (rseng-software-reuse records reuse decisions the same way). - An incident or near-miss resolves: data almost lost, wrong results almost published, credential exposed - these get the fuller postmortem below. - S...