Instructor Guide and Review Status
This is the next authored core-module upgrade after Probability & Statistics. Continuing authoring does not waive the probability module's pending expert review or learner pilots. This module is also not independently reviewed or learner-calibrated; the semester quality matrix remains unchanged.
Technical review priorities
Review the Master Theorem's stated version, n/log(n) recurrence, potential initial state, closed versus half-open scheduling semantics, DP reconstruction, signed pruning, and unit-cost versus bit-cost claims. Solve assessment items without the key first; compare derivations and rubric interpretation afterward. Some existing concept prose remains, so executable checks must not be described as verifying every explanatory sentence.
The revision removes a claimed two-sample recurrence classifier, fixes the dynamic-array potential and binary-counter variable mismatch, corrects lexicographic ordering, qualifies Fibonacci bit costs, rejects universal greedy-versus-DP speed rankings, and replaces misleading generated-section reading links with topic assignments into book indexes.
Maintainer checks
python labs/algorithm-analysis/check.py --module reference
python scripts/package-algorithm-lab.py --check
npm run validate:content
npm run check:links
Rebuild the download after any package edit with python scripts/package-algorithm-lab.py. Only seven named teaching files are included; generated reports and learner submissions are excluded. Extract into a fresh directory and verify both reference success and unfinished-starter failure. Run the deterministic report twice and compare files. Substitute a weighted greedy or strict endpoint comparison to confirm the checker catches those mistakes.
The source reading audit inspected local Skiena chunks on job selection, contiguous arrays, and Fibonacci caching plus the local contents index. MIT's dynamic-array lecture and Kevin Wayne's weighted-scheduling material supply primary external reinforcement. New problems, fixtures, and implementations are original; no full-book review is claimed.
Pilot and assessment calibration
Ask a prerequisite-ready learner to follow the diagnostic, one proof lesson, a lab, and the assessment without undocumented help. Record actual effort, ambiguous terms, first errors, hints, and abandoned steps. Have another reviewer score sample answers to identify rubric disagreements. Recheck changed assumptions after 3–7 days.
Pilot records should identify page, task, elapsed focused time, help, error, correction, competency score, and delayed result. A small pilot reveals specific issues; it does not establish general effectiveness or a population rank. Current workload and pass thresholds remain author estimates.