Chapter 4 Repair Protocol
Use this only after retrieval/testing reveals a weakness.
If you confuse Data Architecture with Data Modeling & Design
- Redraw the enterprise-to-project abstraction stack.
- Label architecture outputs versus modeling/design outputs.
- Classify three prompts:
- target enterprise Customer blueprint,
- normalized Customer/Address logical model,
- Oracle physical keys/indexes.
- State the switch clue: enterprise blueprint/alignment vs model/design execution.
If you confuse EDM and Data Flow Design
For three examples, answer two questions: - What does the data mean / how do concepts relate? → EDM. - Where does it originate/move/store/transform / who uses it? → Flow.
Then redraw one EDM and one flow for the same Customer domain.
If you miss state or lifecycle questions
Rebuild: - current, - target, - transition, - roadmap.
Then write all eight lifecycle status labels and repeat the four-way cue:
Preferred grow | Emerging test | Containment limit | Retirement leave.
If you miss project questions
Rebuild: Scope → Requirements → Design → Implement → Feed Back
Under Implement write: Buy | Reuse | Build.
Then answer: - vendor package? → Buy, - existing governed service? → Reuse, - new structure? → Build.
If you miss framework questions
Write from memory: What | How | Where | Who | When | Why
Then say aloud: “Zachman = ontology/classification, not a project method or roadmap.”
If you miss governance/metric questions
Separate: - who sets/oversees requirements/standards/exceptions, - who creates/maintains architecture, - who provides business meaning/accountability, - who implements the project solution.
Then classify metrics as: COMPLY → CHANGE → VALUE.
If you make a confident miss
Flag it for spaced retesting. High confidence + wrong answer is a durable-risk signal because the wrong discriminator is already reinforced.
Completion test
Chapter 4 is not repaired until you can explain the changed fact that switches the answer—not merely repeat the label.