For instructors
Teach the course
The book is built for a 15-week semester: structured programming in C, then C++ and objects, then basic data structures, with the agent's role growing from tutor to pair to team.
The instructor's guide
A separate instructor's guide supports the course: the teaching stance in full, a week-by-week course plan, assessment and rubrics for the Audit Report and oral defenses, and notes on running labs. It is available to instructors who adopt the book. Write to jerry@drjerryasmith.com from your institutional address to request a copy.
One chapter a week. The book follows a 15-week arc at roughly one chapter a week; a few weeks carry two shorter chapters (Table H.2). Chapter 9 is read in two parts, in weeks 6 and 8. Chapter 19, the capstone, runs alongside the last weeks or in the exam period.
Table H.2. A 15-week schedule.
| Week | Chapters | Phase | Assignments |
|---|---|---|---|
| 1 | 1, 2 | Tutor | AI policy signed; work log starts |
| 2 to 4 | 3, 4, 5 | Tutor | PA1 (weeks 2 to 4) |
| 5 to 6 | 6, 7, 8, 9 (part 1) | Tutor | PA2 (weeks 5 to 6) |
| 7 to 8 | 10, 11, 9 (part 2) | Pair (9 stays Tutor) | PA3 (week 7) |
| 9 to 12 | 12, 13, 14, 15 | Pair | PA4 (weeks 9 to 11); PA5 (week 12) |
| 13 to 15 | 16, 17, 18; 19 | Team | PA6 (weeks 13 to 15); final feature |
The instructor notes. Each chapter's instructor notes are removed from the student edition and published as a separate volume. Each has a lecture card (board drawing, live demo, expected misconception, question for the room); a session plan for a 75-minute class and a 110-minute lab; the threshold concepts that must stay AI-off; checkpoint answers; answers or sketches for every Check and Audit the agent exercise; assessment suggestions; a procedure for re-testing each agent trap; and a log of how each listing was tested.
The AI policy, by phase. Students sign a policy in week 1 (Chapter 2). This text is written to paste into a syllabus; adapt it to your rules.
Suggested syllabus text: use of AI and coding agents.
Purpose. You will learn C and C++ well enough to direct and check coding agents, and you will be graded on evidence of understanding that an agent cannot supply.
Tutor phase (Chapters 1 to 9). Agents may explain, quiz and generate practice. No agent-written code may be submitted.
Pair phase (Chapters 10 to 15). An agent may implement units from a specification you wrote first. You audit every line.
Team phase (Chapters 16 to 20). You may direct several agent sessions in defined roles. Nothing is merged without a reviewed diff, a sanitizer-clean run and your sign-off.
Always AI-free. Exams, memory checkpoints, the Predict step of every lab, your first debugging hypothesis, and the ownership section of any specification.
Disclosure. Attach every transcript used for graded work to its Audit Report. Undisclosed use is an integrity violation; disclosed use within these rules is not.
Ownership. A line you cannot explain in the oral defense counts as a line you did not write.
Access. Every exercise can be done with the course container and a free or course-provided tutor.
Suggested grading weights. The book recommends; you decide. Handwritten, proctored, AI-free exams carry most of the grade: 55% across two midterms and a final. Labs carry 15%, programming assignments 25% and oral defenses 5%. Assignments are graded on hidden scale tests, sanitizer-clean runs, the Audit Report and the specification, not on code that merely runs. Within an assignment, a suggested split is 40% hidden tests, 15% clean sanitizer and Valgrind runs, 35% Audit Report (complexity derivation weighted most) and 10% specification quality.
Placing Chapter 20. The optional concurrency chapter can replace Chapter 19 as the closing chapter, serve as extension reading, or run as a two-week module once Chapters 14 and 16 are done. Appendix G gives all three schedules.
Re-testing the agent traps. Every N.3 trap is a pattern agents have shown, not a guarantee that they still show it. Before each term, run each chapter's "Re-testing the traps" procedure against the agents your students will use, in fresh sessions, and record the results in your copy of Appendix B. If an agent no longer makes a mistake, keep the section: the student's job becomes proving the agent got it right, with the same instrument.
The 15-week schedule
Table G.3. A 15-week schedule (as in How to Use This Book).
| Week | Chapters | Phase | Assignments |
|---|---|---|---|
| 1 | 1, 2 | Tutor | AI policy signed; work log starts |
| 2 to 4 | 3, 4, 5 | Tutor | PA1 (weeks 2 to 4) |
| 5 to 6 | 6, 7, 8, 9 (part 1) | Tutor | PA2 (weeks 5 to 6) |
| 7 to 8 | 10, 11, 9 (part 2) | Pair (9 stays Tutor) | PA3 (week 7) |
| 9 to 12 | 12, 13, 14, 15 | Pair | PA4 (weeks 9 to 11); PA5 (week 12) |
| 13 to 15 | 16, 17, 18; 19 | Team | PA6 (weeks 13 to 15); final feature |
Chapter 9 is read in two parts, in weeks 6 and 8. Chapter 19, the capstone, runs alongside the last weeks or in the exam period.
Re-test the traps each term
Agents improve, so every trap in the book is a pattern to test for, not a guarantee. Before each term, run each chapter's traps against the agents your students will use and record seen, not seen or different error. Keep the check either way: the instrument costs seconds, and the next model may bring the trap back. The map lists all 63 traps with the instrument that catches each.