LAB18 — Complete one transfer task

Start after the core capstone. Choose a bounded task in your own appropriate project, or use the CP09 README clarification from Chapter 18. No new framework, service, or hardware is required. Preserve a starting copy and recovery path before editing.

Write a brief with the goal, starting source, applicable context, excluded changes, and observable acceptance condition. For the supplied maintenance example, explain the difference between malformed fixture input and a valid synthetic run that intentionally enters FAULT. Keep parser behavior unchanged and preserve synthetic labeling.

Complete the task using your chosen workflow. Record preparation, editing/generation, review/repair, verification, and waiting separately. Use actual units and a measurement method. If time was not measured, label the comparison qualitative; do not reconstruct plausible measurements afterward.

For the CP09 documentation task, inspect docs/formats.md and run these commands from the exercise root:

.\scripts\replay.cmd fixtures/invalid.csv
.\scripts\replay.cmd fixtures/normal.csv

Preserve the actual outcomes. The first fixture is intentionally malformed and should be diagnosed with exit 2. The normal fixture contains an intentional nonfinite sample; that policy FAULT can occur during successful replay with exit 0. Do not “fix” those teaching inputs or equate either result with a physical sensor observation.

Inspect the completed diff and connect the result to the acceptance criterion. If comparing two approaches, give each the same preserved starting state and criteria, then retain both artifacts. Count review, correction, and verification for both. A single attempt supports a limited observation, not a permanent ranking of tools.

Write a maintenance plan identifying the current source/tool version, realistic review trigger, relevant continued-behavior check, and last actual result for at least three task-relevant items. A changed skill description, compiler, or hardware target calls for different checks. Select them according to the affected behavior.

Finish with a handoff naming the completed bounded work, actual receipts, remaining limits, and next action. Submit the brief, changed artifact, diff, results, effort record, and maintenance plan. Completion depends on an observable result and explainable evidence. Do not invent timing data, human feedback, or physical validation. Figures: SS18-01–03. The chapter provides the comparison rubric and optional future directions.