Lesson 10 — Exam Reconstruction and Decision Rules
Chapter 12 becomes easier when you reconstruct the decision path, not a dictionary of isolated terms.
Final order of thought
BUSINESS NEED → METADATA STRATEGY → CONSUMER/CONTENT REQUIREMENTS → ARCHITECTURE → METAMODEL/STANDARDS → HARVEST/INTEGRATION → QUALITY/SECURITY/GOVERNANCE → DELIVERY/USE → LINEAGE/IMPACT → METRICS/IMPROVEMENT
Twenty high-value decision rules
- Data vs Metadata — classify by contextual role.
- Business vs Technical vs Operational — meaning/governance vs implementation/movement vs runtime evidence.
- Descriptive vs Structural vs Administrative — identify/retrieve vs parts/relationships vs lifecycle/manage.
- Glossary vs Dictionary vs Catalog — meaning vs structural detail vs discovery/location.
- CMDB vs Metadata repository — IT configuration scope vs broader integrated enterprise Metadata.
- Source vs enterprise repository — local producer/store vs integrated/governed environment.
- Centralized vs Distributed — persistent central copy vs live retrieval/no persistent store.
- Distributed vs Hybrid — zero central persistence vs selected central persistence.
- Hybrid vs Bi-Directional — mixed persistence/retrieval vs controlled write-back.
- Strategy vs Requirements — direction/future state vs specific consumer needs.
- Requirements vs Architecture — what is needed vs how it is organized/sourced/stored/delivered.
- Architecture vs Metamodel — environment pattern vs model of Metadata objects.
- Data Model vs Metamodel — model of business data vs model of Metadata structures.
- Standards vs Manage Stores — conventions/compliance vs repository operations.
- Integrate vs Deliver — reconcile/assemble vs make accessible/useful.
- Harvest vs manual enrichment — automated extraction vs human-added enterprise context.
- As Designed vs As Implemented — intended mapping vs actual code/path.
- Lineage vs Impact — trace provenance vs predict consequences.
- Coverage vs Quality vs Usage — exists vs trustworthy/current vs used.
- Ingest capture vs later cataloging — preserve context now vs reconstruct later.
Changed-fact practice
Architecture
Persistent harvested enterprise copy → Centralized.
Change: no copy, query sources live → Distributed.
Metadata type
Approved Customer definition → Business.
Change: CUSTOMER_ID VARCHAR(20) → Technical.
Lineage
Mapping document describes A → B → As Designed.
Change: current code runs A → C → B → As Implemented.
Dependency question
“Where does sensitive field X flow?” → Lineage.
Change: “What breaks if X is renamed?” → Impact Analysis.
Metrics
Only half of strategic scope documented → completeness.
Change: almost all documented but nobody uses it → usage/adoption.
Sixty-second reconstruction
You should be able to state, without notes: - three primary Metadata types; - Glossary/Dictionary/Catalog; - all four architectures; - five core activities; - metamodel definition; - integration vs delivery; - designed vs implemented lineage; - lineage vs impact; - why Metadata can be sensitive; - seven metric families.
If you can name a term but cannot say what changed fact would make another answer better, return to its Battle Card before the Question Bank.
Source anchor: pp. 395–423.