The Two Planes

FERIN separates information into two distinct planes of existence:

Concept Plane

Where meaning lives

  • Concepts — abstract ideas in concept systems
  • Concept versions — specific realizations over time
  • Concept relationships — connections between concepts
  • Concept version relationships — connections between versions

Content Plane

Where data persists

  • Register item classes — data schemas
  • Register items — data instances
  • Changes — records of modifications

Elements on Both Planes

Identifiers

Both object and functional identifiers apply to entities on both planes.

Relationships

Typed connections exist as concept relationships, concept version relationships, and content relationships.

Key Relationship

Each concept version is backed by a register item (data instance) in a particular register item class (data schema). This separation enables concepts to evolve while preserving the integrity of the data that represents them.

This separation enables concepts to evolve (e.g., refining what "meter" means) while preserving the integrity of content that references them (e.g., measurements using "m").

Key Entities

Concept

An abstract idea or category that can be defined. A concept represents "what something means" rather than how it's represented.

Example: The concept of "meter" as the SI unit of length. The concept has a definition that has evolved over time (from a physical bar to a wavelength standard to the current speed-of-light definition).

Concept Version

A specific version of a concept's definition at a point in time. When a definition changes significantly, a new version is created.

Example: The 1960 definition of meter (wavelength of krypton-86) vs. the 1983 definition (distance light travels in 1/299,792,458 second).

Register Item Class

A data schema that defines the structure of register items. Each register item class specifies what fields and data types are valid for items in that class. Register item classes can evolve independently of the concepts they represent.

Example: A "Unit of Measure" register item class might define fields for name, symbol, definition, and conversion factor. This schema can be versioned separately from the concept of "meter" itself.

Register Item

A concrete piece of content in a register—an instance of a register item class. Register items have identifiers, status, and are managed through governance processes. Each concept version is backed by a register item.

Example: The symbol "m" for meter, with identifier urn:iso:std:iso:19135:meter, status "valid", added on 1960-10-14 by the BIPM, in the "Unit of Measure" register item class.

Register

A managed collection of register items. A register has an owner, follows a register specification, and implements the FERIN framework.

Example: The Register of Units of Measure (RUM), containing all SI units and commonly used non-SI units.

Identifier

A unique label that distinguishes one entity from another. Identifiers exist on both the concept plane and content plane. FERIN requires two types of identifiers for each entity: an object identifier (non-redirectable, permanent) and a functional identifier (redirectable, supports hierarchy). This dual approach enables both stable citation and semantic addressing.

Example: The concept "meter" has an object identifier uuid:550e8400-... (permanent) and a functional identifier /concepts/units/length/meter (hierarchical, semantic).

Relationship

A typed connection between entities. Relationships exist on both planes: concept relationships and concept version relationships on the concept plane, and content relationships on the content plane. Relationships enable lineage tracking, semantic navigation, and temporal reconstruction.

Example: A "supersedes" relationship from meter v2 to meter v1 indicates that v2 replaces v1. A "has-part" relationship from "Address" to "Street" indicates composition.

Change

A record of a modification to register content. Every action (add, modify, invalidate, supersede, etc.) creates a change record with complete information about what was modified, when, by whom, and why. Change records enable point-in-time reconstruction and complete audit trails.

Example: When the definition of "meter" was updated in 1983, a change record was created noting the action (add-version), the affected concept, the new definition, the timestamp, and the deciding body (CGPM).

Identifiers

Every entity in FERIN requires two types of identifiers:

Object Identifier

Non-redirectable, permanently assigned to a specific object. Provides permanent reference for citations.

Example:550e8400-e29b-41d4-a716-446655440000 (UUID)

Functional Identifier

Redirectable, reflects semantic intent. Must support hierarchical specification.

Example:/concepts/meter/v2.0 (URI)

What Needs Both Types

EntityObject IDFunctional ID
The RegisterRequired
ConceptsRequiredRequired
Concept VersionsRequiredRequired
Register Item ClassesRequiredRequired
Register ItemsRequiredRequired
ChangesRequiredRequired

See Identifier Design for implementation guidance.

Concept Granularity

The Challenge

One of the most common questions in FERIN implementation is: "When should I create a new concept vs. a new version of an existing concept?"

This isn't explicitly defined in the standard, so here's practical guidance:

Decision Framework

ScenarioActionRationale
Definition clarified but meaning unchangedNew versionThe concept's identity remains the same
Definition refined with additional precisionNew versionConcept is the same, just better understood
Definition changed to mean something differentNew conceptIdentity has fundamentally changed
Scope or domain of application changedNew conceptContext has shifted
Same term, different meaning in different contextsMultiple conceptsHomonyms require distinct concepts

The "Would It Confuse Users?" Test

A practical heuristic: if existing references to the concept would become confusing or misleading after a change, create a new concept. If references would still be accurate (just less precise), create a new version.

Example: Defining "meter"

When the definition of meter changed from a physical bar (1889) to a wavelength (1960) to the speed of light (1983), these were new versions because the concept "meter" remained fundamentally the same unit of length—the precision just improved.

However, if meter had been redefined as "1/10 of a nautical mile," that would require a new concept because the fundamental meaning changed.

Concept Systems

Concepts are organized into concept systems—structured sets of related concepts. Common patterns include:

Generalization-Specialization

Broader concepts (general) relate to narrower concepts (specific).

Unit → Length Unit → meter

Whole-Part

Concepts that are composed of other concepts.

Address → Street, City, PostalCode

Sequential

Concepts that follow each other in a sequence.

Draft → Review → Approved → Published

Concept Lifecycle

Concepts progress through a lifecycle as they are defined, refined, and potentially superseded:

1

Proposed

A new concept is submitted for consideration

2

Under Review

The concept definition is being evaluated

3

Valid

The concept is approved and available for use

4

Superseded

A newer version or replacement concept exists

5

Invalid/Retired

The concept is no longer valid for use

Related Topics