Instructor Guide and Review Status
Status: authored reference candidate with executable verification recorded in the repository authoring review. Independent mathematical/code review, grading agreement and learner pilots remain pending. Do not equate a static site build with teaching effectiveness or a professional-percentile outcome.
Subject review
Review all 17 lessons and the contracts page. Focus on state sufficiency versus overlap; counting versus optimization; exact-use/unreachable bases; cache lifetime; bit complexity; strict LIS and witness reconstruction; Unicode/edit semantics; capacity update versions; interval boundaries; tree validation/rerooting cost; tour closure; digit padding/zero; dominance; greedy proof scope; dependency DAG arrow direction and multi-child combination; and the budgeted scheduling extension.
Review implementation and oracle separately. Recompute witness constraints and numeric fixtures. Check that the tests accept tied optima and reject an illegal witness even with a correct scalar. Finite oracles are bounded to tiny inputs; source review must justify complexity and general correctness.
Executable and scoring review
Extract the seven-file download and run reference checks/report generation. Confirm the untouched starter fails. Introduce a strict/nondecreasing LIS error, reuse a 0/1 item, omit a tour's return cost or alter digit tightness and confirm rejection; restore the reference afterward. Record environment and exact scope.
Independently score two sample submissions using all eight competencies, then record disagreements and revise ambiguous criteria. Pay particular attention to a correct value with an invalid witness, conceptual versus arithmetic errors, and claimed linear storage that omits reconstruction. A mean score must not hide a missing core competency.
Learner pilot and delayed transfer
Recruit learners who satisfy the earlier-skill diagnostic, including someone without prior DP fluency. Record hours by block, setup failures, state-design difficulties, assistance, rejected contracts and repairs. The 90–125 focused-hour range is an author estimate, not observed effort. After 7–14 days use a fresh constraint change and score transfer without the original solution key.
| Evidence | Record |
|---|---|
| Author revision | 2026-09-18; repository authoring review records findings and checks |
| Independent subject/code review | Pending: reviewer, scope, issues, resolutions |
| Cross-platform execution | Pending beyond the recorded author environment |
| Grading agreement | Pending: submissions, independent ratings, disagreements |
| Learner pilot | Pending: prerequisites, hours, errors, assistance and revisions |
| Delayed transfer | Pending: fresh task, date, affected competency ratings |
Use observed evidence to revise workload, explanations and rubric. Further core authoring can continue while human review is arranged; do not advance the semester quality matrix without its required evidence.