Capstone Review Worksheet

Page Maps

graph LR
  family["Python Programming"]
  program["Python Object-Oriented Programming"]
  section["Capstone"]
  page["Capstone Review Worksheet"]
  capstone["Capstone evidence"]

  family --> program --> section --> page
  page -.applies in.-> capstone

flowchart LR
  surface["Pick one boundary or scenario surface"] --> owner["Name the owner"]
  owner --> evidence["Name the first test or bundle"]
  evidence --> risk["State the drift risk"]
  risk --> verdict["Decide whether the design still deserves trust"]

Use this page after a module, after a rewrite, or before an extension. It is not a page for praise. It is a page for pressure. The goal is to answer whether the capstone still teaches good object-oriented design under review.

Ownership review

• Can you point to one authoritative owner for rule lifecycle?
• Can you point to one authoritative owner for alert creation?
• Can you point to one clear seam for evaluation variation?
• Can you explain why runtime.py is not the domain model?
• Can you explain why projections and read models are useful but non-authoritative?

Boundary review

Boundary	Review question	Evidence to inspect first
aggregate	Does model.py still decide the legal lifecycle transitions?	tests/test_policy_lifecycle.py and rules.txt
evaluation seam	Can a new evaluation mode land in policies.py without widening the aggregate?	tests/test_policy_evaluation.py and rate_of_change.txt
runtime	Does coordination stay outside domain truth?	tests/test_runtime.py, timeline.txt, and walkthrough.txt
projections and read models	Are the derived views downstream of events instead of upstream of authority?	history.txt, snapshot.json, and retirement scenarios
persistence and rollback	Would a sink or storage failure roll back state without relocating business rules?	tests/test_unit_of_work.py, tests/test_runtime.py, and pytest.txt
public surface	Can a learner drive the scenario without reaching into internals?	application.py, demo.py, and walkthrough.txt

Drift warnings

Mark these aggressively. One "yes" is enough to justify a deeper review.

• Does the runtime now answer a question that used to belong to the aggregate?
• Does a projection need to be consulted before the aggregate can decide?
• Does a persistence concern now hide domain policy from a reader?
• Does one feature force changes in aggregate, runtime, and projection code at the same time?
• Does the strongest proof route still pass, but the owning boundary feel less legible?

Change-safety review

• If a new rule mode arrived tomorrow, would the first edit point be obvious?
• If a new incident sink arrived tomorrow, could it stay outside the aggregate?
• If rollback semantics changed tomorrow, would repository.py be the natural pressure point?
• If a learner asked where state truth lives, would you answer with one file, not three?

Proof review

Question	Best first route
What state exists right now?	inspect
How does the scenario unfold?	capstone-walkthrough
Do tests and saved review artifacts agree?	capstone-verify-report
Does the strongest local route still hold?	capstone-confirm
Does the published learner route still deserve trust?	proof

Review prompts by course stage

Course stage	Most important question
Modules 01-03	Which semantic or lifecycle rule would fail first if ownership drifted?
Modules 04-07	Which boundary is authoritative once collaboration, projections, and runtime pressure are involved?
Modules 08-10	Which saved route proves the design to someone who did not watch the implementation happen?

Minimum honest review notes

When you finish a capstone review, you should be able to fill in all four blanks:

1. The owner is ____.
2. The strongest supporting boundary is ____.
3. The first proof route I would show another reviewer is ____.
4. The first place I would expect drift is ____.

Stop conditions

Stop and reopen the architecture or extension pages if either becomes true:

• you can describe behavior, but not ownership
• you can name a route, but not why that route matches the claim

Exit check

Leave this worksheet only when you have a review verdict that sounds like this:

The design still deserves trust because ____ owns ____, ____ remains derived, and ____ would fail first if that split drifted.