For instructors¶

This course is designed for guided self-study, a one-day quick route, or a multi-session onboarding course.

Learners with no programming experience should complete the Python preflight before the first session.

Recommended formats¶

One-day quick route¶

Session	Material
09:00	overview and lesson 01
10:00	functions and Exercise 02
11:30	arrays and Exercise 03
13:30	debugging and Exercise 04
15:00	schemas, tools, and Basilisk-log lab
16:30	explain one result and review

Aim for roughly 20% exposition and 80% tracing, coding, checking, and explanation.

Six-session full route¶

thinking, functions;
arrays, control;
debugging, schemas;
plotting, numerical verification;
tools, reproducibility;
CoMPhy labs and capstone proposal.

Facilitation¶

Ask for predictions before live execution.
Let the first test fail visibly.
Request an adversarial input, not only a happy path.
Ask “what does this number mean?” more often than “what syntax did you use?”
Pair students to explain each other's code.
Do not repair every error immediately; help the student form a discriminating check.

AI-assisted work¶

Require students to retain an initial generated attempt, identify its physical and computational assumptions, and explain the corrected version in a pull request. Grade the audit, not the prompt.

Solutions¶

Worked solutions are public by design. Use them after a genuine attempt:

compare boundaries and tests before bodies;
identify one difference in scientific policy;
improve either version;
explain the choice.

Hiding solutions in an open repository is security theatre.

Assessment¶

Use the capstone rubric. A student should not pass with a polished figure but no schema, or a comprehensive test suite but no physical question.

Extending the course¶

Prefer a new exercise when:

it represents a recurring CoMPhy task;
it teaches a distinct decision;
public or synthetic data suffice;
a compact automated check exists.

Do not add an exercise merely to cover another Python feature.