Visual Atlas — Maps 1–5
Map 1 — Requirements → Implementable Design
REQUIREMENTS
↓ add agreed business scope/meaning
CONCEPTUAL DATA MODEL (CDM)
↓ add detailed technology-independent requirements
LOGICAL DATA MODEL (LDM)
↓ add technology/platform implementation detail
PHYSICAL DATA MODEL (PDM)
Interpretation: movement to the right/down adds precision and implementation specificity. Iteration can move backward when a later design exposes a misunderstanding.
Redraw check: Can you name three details that belong in LDM but not CDM, and three that belong in PDM but not LDM?
Map 2 — Core Data-Model Components
ENTITY
= type/concept about which information is collected
└─ has ATTRIBUTES
= properties that identify / describe / measure
└─ each attribute has a DOMAIN
= complete valid value set / structure
ENTITY ↔ RELATIONSHIP ↔ ENTITY
= business association
├─ CARDINALITY: zero / one / many instances
└─ ARITY: one / two / three entity types
Remember: an entity instance is an occurrence of the entity. Domain is not a peer entity type.
Map 3 — Key Taxonomy: Build vs Job
KEY CONSTRUCTION — “What is it made of?”
SIMPLE = 1 attribute
COMPOSITE = 2+ attributes
COMPOUND = composite whose components are foreign keys
KEY FUNCTION — “What job does it perform?”
SUPER = any unique set
CANDIDATE = minimal unique set
PRIMARY = chosen candidate
ALTERNATE = unchosen candidate
BUSINESS = business-meaningful retrieval identifier
SURROGATE = system-generated; no business meaning
Critical: one key can belong to multiple classifications.
Map 4 — Five Domain Forms
DOMAIN = valid value space for an attribute
DATA TYPE → what basic kind of value? (integer, date)
FORMAT → what pattern? (AAA-9999)
LIST → which enumerated values? (Open/Closed/Cancelled)
RANGE → between which boundaries? (1..5)
RULE-BASED → what conditional rule? (HireDate < today)
The five are alternatives, not stages.
Map 5 — Six Modeling Schemes
RELATIONAL → operational rules / normalized business meaning
DIMENSIONAL → analytics / facts + dimensions + grain
OBJECT-ORIENTED→ classes + attributes + methods/operations
FACT-BASED → objects + facts + roles + controlled verbalization
TIME-BASED → historization; Data Vault / Anchor examples
NoSQL → document / key-value / column / graph physical shape
Redraw test: invent one requirement that would make each scheme the strongest recognition answer.
Source anchors: current Visual Memory Map, Chapter 5 maps 1–5.