Visual Maps 1–6
Map 1 — Shared-Context Stack
Redraw:
TRANSACTION EVENTS → involve → MASTER ENTITIES → classified/contextualized by → REFERENCE DOMAINS
METADATA → describes meaning/source/stewardship/lineage across all layers.
Interpretation: Events depend on persistent entities. Reference values give consistent classification/context. Metadata describes the data-management facts rather than becoming another business-data domain.
Answer check: Order = Transaction; Customer = Master; Order Status = Reference; steward/update schedule = Metadata.
Source: pp. 332–339.
Map 2 — Enterprise Trust Chain
Redraw:
SOURCE SYSTEMS → assess quality/semantics → STANDARDIZE/ENRICH → RESOLVE IDENTITY → TRUSTED SOURCE → SHARE TO CONSUMERS → STEWARD FEEDBACK ↺
Interpretation: Trust is produced through governed processing and feedback. It is not created by declaring a repository “golden.”
Answer check: Preparation and identity resolution are distinct. Steward feedback improves sources/rules rather than merely cleaning forever.
Source: pp. 340–347, 351–353.
Map 3 — Authority Concepts
Redraw four labels with deciding questions:
- SYSTEM OF RECORD: Where is it created/maintained?
- SYSTEM OF REFERENCE: Where should consumers obtain it?
- TRUSTED SOURCE: What governed source/view should users rely on?
- GOLDEN RECORD: What reconciled record represents this entity instance?
Interpretation: “Authoritative” can refer to different activities. The concepts can coexist in one architecture.
Answer check: Golden Record does not guarantee perfect truth.
Source: pp. 339–340, 345–351.
Map 4 — Reference Data Structure Ladder
Redraw:
LIST → CROSS-REFERENCE → TAXONOMY → ONTOLOGY
code/description → translate representations → hierarchy → richer semantic relationships
Interpretation: The key is the relationship required, not merely “complexity.”
Example: status list → ISO code crosswalk → product hierarchy → formal domain relationship model.
Source: pp. 334–338.
Map 5 — MDM Processing Spine
Redraw:
MODEL → ACQUIRE → VALIDATE / STANDARDIZE / ENRICH → ENTITY RESOLUTION + ID MANAGEMENT → SHARE + STEWARD
Interpretation: Define common semantics and prepare evidence before identity decisions.
Answer check: Matching before adequate modeling/source assessment/standardization is a common sequencing error.
Source: pp. 341–347.
Map 6 — Matching Decision
Redraw:
CANDIDATE PAIR → similarity evidence → SAME ENTITY?
If system says SAME but truth is DIFFERENT → FALSE POSITIVE.
If system says DIFFERENT but truth is SAME → FALSE NEGATIVE.
Add: MATCH HISTORY + REVERSIBILITY ↺
Interpretation: Matching is a governed decision under uncertainty. Error costs drive thresholds and stewardship.
Answer check: false positive over-combines; false negative under-matches.
Source: pp. 343–346.