Reference Module Standard
Use the Probability & Statistics module as a candidate pattern for future upgrades. It is author-reviewed and executable, but still awaits independent subject review and learner pilots. More content alone is not evidence of teaching quality or an elite percentile outcome.
Authoring checklist
| Requirement | Reviewable artifact | Failure that blocks completion |
|---|---|---|
| Prerequisites | Diagnostic of previously taught skills and repair paths | Entry test assumes the module has already been learned |
| Lesson correctness | Definitions, assumptions, derived examples, counterexamples | Formula is applied to a different experiment |
| Practice progression | Guided example, independent attempt, changed assumption | All exercises can be completed by copying a template |
| Sources | Exact source or honest book-level topic assignment | Link label claims a section the destination does not identify |
| Runnable work | Starter, fixture, reference, contracts, independent checks | Setup requires undocumented tools or samples do not run |
| Solutions | Separate derivations and error explanations | Answers are absent or numbers lack reasoning |
| Assessment | Competency map, rubric, critical-error gates, recheck | Aggregate score hides a missing foundational competency |
| Usability | Workload estimate and observed pilot record | Ambiguous instructions or hidden prerequisites block progress |
| Maintenance | Content checks, built routes, reproducible downloadable package | Website and source package diverge |
Rollout
First obtain independent mathematical review and small learner pilots for the probability candidate. Record actual workload, errors, reviewer disagreement, and delayed transfer. Revise the pattern from that evidence before copying it across the core. Then choose the next module by prerequisite importance and demonstrated gaps, rather than expanding every reading list at once.
Keep authored, executed, independently reviewed, and learner-piloted statuses distinct. The semester quality matrix should only advance when its own required evidence exists; one improved module does not validate a whole semester.
Current authored candidates
| Module | Concrete package | Human validation |
|---|---|---|
| Proof Techniques & Discrete Structures | Extended lessons, proof/model contracts, separate practice and assessment keys, finite-model labs and workflow quotient project | Independent review and learner pilots pending |
| Combinatorics & Graph Theory | Extended counting and graph lessons, separate solution keys, finite enumeration labs, matching certificates and a workshop planning project | Independent review and learner pilots pending |
| Probability & Statistics | Corrected lessons, practice/assessment keys, offline probability labs | Independent review and learner pilots pending |
| Linear Algebra for CS | Revised exact/numerical lessons, separate solution keys, rational certificates, QR/Cholesky/spectral labs, numerical experiments and calibration project | Independent review and learner pilots pending |
| Problem-Solving Strategies | Corrected and extended strategy lessons, 20 practice tasks, ten gates, separate keys, four offline labs and a changed-model project | Independent review and learner pilots pending |
| Algorithm Analysis & Design | Corrected analysis/proof lessons, scheduling/search labs, changed-assumption assessment | Independent review and learner pilots pending |
| Sorting, Searching & Fundamental Structures | Revised concept lessons, stable sorting/search/selection/map/heap/stream labs, stateful checks and resource accounting | Independent review and learner pilots pending |
| Graph Algorithms & Network Analysis | Revised graph proofs, eight implementations with output witnesses, exhaustive small-graph oracles, flow/cut certificates and dependency-planner transfer | Independent review and learner pilots pending |
| Dynamic Programming & Optimization | Revised state/proof/resource lessons, eight implementations with witnesses, exhaustive tiny-instance oracles, digit-bound contracts and budgeted scheduling transfer | Independent review and learner pilots pending |
| Advanced Data Structures & Amortized Analysis | Revised invariants and cost models, eight implementations, stateful model/ownership checks, range-update composition and historical-index transfer | Independent review and learner pilots pending |
| OOD Foundations & Code Smells | 19 lessons, 20 practice tasks, ten gates, separate keys, four offline contract labs, three design experiments and a compatibility project | Independent review and learner pilots pending |
| Refactoring Techniques | 18 lessons, 20 practice tasks, ten gates, separate keys, five offline refactoring labs, three experiments and a compatibility migration project | Independent review and learner pilots pending |
| Behavioral Patterns | 23 lessons covering all eleven behavioral patterns and runtime comparisons, 20 core tasks, C01–C10 plus BD1–BD6, seven core labs and an EventBus/Editor kit, three experiments and a workflow project | Independent review and learner pilots pending |
| Structural & Creational Patterns | Prior-skill diagnostic, 20 core tasks, five offline construction/copy/tree/lifetime groups, three experiments, C01–C06 plus SC1–SC6, separate practice/quiz/assessment calibration | Independent review and learner pilots pending |
| Applied Design & Code Review | Prior-skill diagnostic, 20 review tasks using the existing regression kit, three experiments, AR1–AR6 and retained HCI/accessibility gates, separate practice/quiz calibration | Independent review and learner pilots pending |
Further authoring can proceed while those reviews are pending. That does not establish that the pattern is learner-validated or remove the need to revise the candidates after pilot evidence.
Semester 2 now also has an integration project and cumulative assessment, with worked exam solutions, six competency gates, a versioned scheduling kit, independent tiny-instance checks and an explicit workload reconciliation. This integration is an authored candidate; independent review and learner pilots remain pending.
Semester 1 also has a revised integration project and assessment, with a separate complete exam key, per-competency progression rules, a graph investigation requiring interactive geometric evidence, and an explicit 520–750 hour workload estimate. This is an authored candidate awaiting independent review and learner pilots; the estimate is not an observed completion time.
Later-semester parity repair
Semesters 4–10 retain their existing lessons, practice, labs and assessments. Each of their 35 modules now adds a prior-skill check with repair calibration, an independent case and changed assumption, an experiment tied to its existing lab, and a separate quiz key. Workload pages explain artifact reuse and keep real-system, independent review and delayed-transfer gates distinct. This is a bounded teaching-consistency repair, not a claim that every lesson has received independent subject review or learner validation.