Key Design Patterns

Overview of Object Properties

Overview of DUL Patterns

DUL is not just a taxonomy but a library of reusable patterns. Each pattern solves a recurring modeling problem.

Major Patterns:

1. Descriptions & Situations (D&S)
2. Quality-Region
3. Participation
4. Information Realization
5. Mereological Patterns
6. Classification
7. Place
8. Sequence

The Descriptions & Situations (D&S) Pattern

Problem: How to model context-dependent relations and time-varying properties?

Solution: Separate intensional (conceptual) and extensional (actual) aspects:

When to use:

• Role assignment: Agent plays Role in Situation
• Task execution: Action executes Task in Situation
• Time-indexed relations: Relations hold within temporal Situations
• Multi-perspective modeling: Same entities, different Situations/Descriptions

Example Pattern Instance:

Recipe (Plan, a Description)
  --defines--> 'Ingredient' (Role)
  --defines--> 'Mixing' (Task)

CookingSession (PlanExecution, a Situation)
  --satisfies--> Recipe
  --includesAgent--> Chef₁ (plays 'Cook' Role)
  --includesAction--> Mixing₁ (executes 'Mixing' Task)
  --includesObject--> Flour₁ (plays 'Ingredient' Role)

The Quality-Region Pattern

Problem: How to model attributes with:

• Multiple measurement systems
• Temporal change
• Observational relativity

Solution: Reify attributes as Qualities, link to abstract Regions:

When to use:

• Scientific observation: Same phenomenon, different measurements
• Individual aspects matter: Not just "height" but "this building's height"
• Parameterized constraints: Concepts constrain regions through parameters

Example Pattern Instance:

DmitrisSkin (Quality)
  --isQualityOf--> Dmitri (Person)
  --hasRegion--> Yellow₁ (SocialObjectAttribute / Color value)

MeasurementSituation
  --includesObject--> Dmitri
  --includesQuality--> DmitrisSkin
  --satisfies--> ColorimetryDescription
    --defines--> RGBParameter

The Participation Pattern

Problem: How to link events and objects?

Solution: Symmetric participation relation:

Extensions:

• Agent participation: Action --hasParticipant--> Agent (agent-specific)
• Co-participation: Object --coparticipatesWith--> Object (derived from shared event)

When to use:

• Event modeling: Who/what is involved?
• Provenance: What events affected this object?
• Social network analysis: Who co-participated in events?

Example Pattern Instance:

TennisMatch₁ (Event)
  --hasParticipant--> Vitas (Agent)
  --hasParticipant--> Jimmy (Agent)
  --hasParticipant--> TennisBall₁ (Object)

Vitas --coparticipatesWith--> Jimmy (inferred)

The Information Realization Pattern

Problem: Distinguish abstract information from concrete manifestations.

Solution: Two-level information model:

When to use:

• Cultural heritage: Work vs. manifestations (FRBR-compatible)
• Legal documents: Legal content vs. physical copies
• Knowledge representation: Concepts vs. terms expressing them

Example Pattern Instance:

Constitution_ItalianRepublic (InformationObject)
  --expresses--> ItalianLegalSystem (Description/Norm)

PhysicalCopy₁ (InformationRealization / PhysicalObject)
  --realizes--> Constitution_ItalianRepublic

OralRecitation₁ (InformationRealization / Event)
  --realizes--> Constitution_ItalianRepublic

Mereological Patterns

Problem: Model part-whole relations with different characteristics.

Solution: Multiple parthood properties:

Pattern A: Reflexive Part-Whole

Use when: General decomposition, entity is part of itself (mereologically valid)

Pattern B: Proper Part-Whole

Use when: Strict decomposition, part ≠ whole

Pattern C: Component-System

Use when: Designed artifacts with direct structural components

Pattern D: Constitution

Use when: Different ontological strata (social/physical, organism/molecular)

Example Pattern Instance:

Car₁ (DesignedArtifact)
  --hasComponent--> Engine₁ (direct component)
  --hasComponent--> Wheel₁ (direct component)

Engine₁
  --hasProperPart--> Piston₁ (engine part)

Car₁ --hasPart--> Piston₁ (inferred via transitivity of hasPart, if rules applied)

Person₁ (BiologicalObject)
  --hasConstituent--> Molecule₁ (cross-layer: organism → molecular)

The Classification Pattern

Problem: Context and time-dependent classification.

Solution: Classification as a special Situation:

When to use:

• Role assignment: Agents play different roles at different times
• Taxonomic classification: Species membership, artifact categories
• Status tracking: Legal status, health status, employment status over time

Example Pattern Instance:

Classification₁
  --satisfies--> TrafficLaw (Norm defining vehicle types)
  --isSettingFor--> Vehicle₁ (Entity)
  --isSettingFor--> 'Truck' (Concept)
  --isSettingFor--> Interval₂₀₂₄ (when registered as truck)

Classification₂
  --satisfies--> HistoricalArchive (Description)
  --isSettingFor--> Vehicle₁ (same entity)
  --isSettingFor--> 'ClassicCar' (different Concept)
  --isSettingFor--> Interval₂₀₃₀ (after retirement from service)

The Place Pattern

Problem: Locations are both physical and social constructs.

Solution: Distinguish PhysicalPlace and Place:

When to use:

• Physical location: Coordinates, geometric regions
• Social/administrative places: Countries, neighborhoods, institutions
• Hybrid: "Paris" is both a geographic region and a social construct

Example Pattern Instance:

Paris_Geographic (PhysicalPlace)
  --hasSpaceRegion--> Region48.8566N_2.3522E

Paris_City (Place, SocialObject)
  --isExpressedBy--> ParisCharter (InformationObject / legal document)
  --describes--> Paris_Geographic
  --isMemberOf--> FrenchCities (Collection)

The Sequence Pattern

Problem: Model temporal or logical ordering.

Solution: Precedence relations:

When to use:

• Workflow: Task sequences
• Narrative: Event chronology
• Procedure: Step ordering

Example Pattern Instance:

Workflow₁ (Description)
  --defines--> Task₁, Task₂, Task₃

Task₁ --directlyPrecedes--> Task₂
Task₂ --directlyPrecedes--> Task₃

Task₁ --precedes--> Task₃ (inferred via transitivity)

WorkflowExecution₁ (Situation)
  --satisfies--> Workflow₁
  --includesAction--> Action₁ (executes Task₁)
  --includesAction--> Action₂ (executes Task₂)
  --includesAction--> Action₃ (executes Task₃)

Pattern Selection Guide

Pattern Composition

Real-world modeling typically combines multiple patterns:

Example: Project Management Ontology

Uses:

1. D&S: Project (Plan) defines Roles/Tasks, ProjectExecution (Situation) satisfies Project
2. Participation: Actions have Agent participants
3. Classification: Agents classified by Roles (ProjectManager, Developer)
4. Sequence: Tasks ordered by precedence
5. Quality-Region: Project attributes (budget, timeline) as Qualities with Amount/TimeInterval Regions
6. Mereological: Project has sub-projects as proper parts

Pattern Integration Strategy:

1. Identify core entities (agents, resources, activities)
2. Apply D&S for role/task structure
3. Apply Participation for event-object links
4. Apply Classification for role assignment
5. Apply Sequence for workflow ordering
6. Apply Quality-Region for measurements/KPIs
7. Apply Mereology for hierarchical decomposition