Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ISO/IEC 11179-2-06 (CAN/CSA-ISO/IEC 11179-2-06): Metadata Registries Classification — Technical Guide
Comprehensive overview of the classification schema requirements for metadata registries, including implementation and compliance guidance
1. Scope and Purpose
The standard ISO/IEC 11179-2-06, formally adopted in Canada as CAN/CSA-ISO/IEC 11179-2-06, defines the structure and requirements for classification schemes within a metadata registry (MDR). As Part 2 of the ISO/IEC 11179 family, it focuses exclusively on the classification component — providing a standardized metamodel for organizing administered items into taxonomies, thesauri, or other classification structures.
This part applies to anyone designing, implementing, or maintaining a metadata registry that needs to categorize metadata items (e.g., data elements, value domains, conceptual domains) according to a consistent classification scheme. The scope includes:
- Definition of classification scheme and classification scheme item constructs.
- Required and optional attributes for classification scheme items.
- Relationships between classification scheme items (hierarchical and associative).
- Binding of administered items to classification scheme items.
- Conformance criteria for classification-related parts of an MDR.
By adhering to this standard, organizations ensure that their metadata classification systems are interoperable, well-documented, and support unambiguous identification and retrieval of metadata across distributed systems.
Tip: ISO/IEC 11179-2-06 does not prescribe a specific classification system (e.g., UNSPSC, NAICS). Instead, it provides the general model into which any classification scheme can be mapped, making it domain-agnostic.
2. Technical Requirements for Classification Schemas
2.1 Core Constructs
The standard introduces two primary constructs: Classification_Scheme and Classification_Scheme_Item.
- Classification_Scheme – A logical grouping of classification scheme items that share a common purpose or source. It includes attributes such as identifier, name, definition, and source document reference.
- Classification_Scheme_Item – A node within a classification scheme. Each item has a unique identifier within the scheme, a name, and a definition. Items are arranged hierarchically via parent-child relationships and may be linked to administered items (such as data elements or value domains).
2.2 Mandatory and Optional Attributes
The following table summarizes the key attributes defined for a classification scheme item. Compliance requires the mandatory attributes to be recorded for every item.
| Attribute | Mandatory | Description |
|---|---|---|
identifier | Yes | Unique identifier within the classification scheme (e.g., a numeric code or alphanumeric string). |
name | Yes | Human-readable label for the item. |
definition | Yes | Clear, unambiguous statement explaining the item’s meaning and scope within the scheme. |
parent_item | No | Reference to the immediate parent classification scheme item (inherits hierarchical position). If omitted, the item is a top-level node. |
definition_source | No | Reference to the source from which the definition is taken (e.g., an external standard). |
effective_date | No | Date when the item became valid within the classification scheme. |
change_description | No | History notes describing modifications to the item. |
2.3 Relationships and Bindings
Classification scheme items participate in two key relationships:
- Parent-child – A strict hierarchy that defines the taxonomy structure. It must be a directed acyclic graph (DAG) without cycles.
- Associative – Non-hierarchical links (e.g., “see also” or “related”) that connect items across different branches.
The binding between classification scheme items and administered items (e.g., data elements, conceptual domains) is done through the Classification_Scheme_Item_Relationship construct, which allows multiple items to be linked to a single administered item and vice versa.
Important: When implementing a classification scheme, ensure that the hierarchical relationships do not create cycles. The standard requires that the parent-child tree be a proper DAG. Use integrity constraints to enforce this.
2.4 Classification Scheme Types
ISO/IEC 11179-2-06 distinguishes two broad types of classification schemes:
- Enumerative / Heirarchical – A tree structure where items are mutually exclusive at each level (e.g., product classification).
- Faceted – Items belong to independent facets (dimensions) that can be combined. A faceted scheme can be represented by multiple classification schemes, one per facet, with items that are non-hierarchical.
The standard provides generic models that accommodate both types without requiring changes to the metamodel.
3. Implementation Highlights
3.1 Mapping Existing Taxonomies
Many organizations already use classification systems (e.g., industry codes, library classifications). To conform to ISO/IEC 11179-2-06, organizations map each existing category to a classification scheme item, assigning the mandatory identifier, name, and definition. The original code becomes the identifier, the category label becomes the name, and the category description becomes the definition.
3.2 Tool Support and Metadata Repositories
Popular MDR tools (e.g., IBM InfoSphere, Manta, or open-source solutions like CKAN) typically support classification schemas. Compliance with Part 2 means the tool’s underlying data model must capture all mandatory attributes, enforce DAG constraints, and allow binding between classification items and metadata records.
3.3 Integration with Data Governance
Classification schemes defined per Part 2 form the backbone of a data governance program. They enable:
- Consistent tagging of data assets across systems.
- Impact analysis when a classification item changes.
- Faceted search and filtering for data discovery.
Best Practice: Start with a controlled vocabulary for your classification scheme names and definitions to avoid duplication and ambiguity. Use the standard’s “definition_source” attribute to track provenance when adopting external schemes.
4. Compliance and Conformance
4.1 Conformance Levels
ISO/IEC 11179-2-06 defines two conformance levels relevant to classification:
- Level 1 (Structural Conformance) – The registry must provide the ability to record classification schemes and items with the mandatory attributes and relationships as specified.
- Level 2 (Behavioral Conformance) – In addition, the registry must support query, retrieval, and maintenance operations that preserve the integrity of the classification structure (e.g., enforcing DAG, maintaining cross-references).
Full compliance with the standard requires both levels.
4.2 Auditing for Compliance
When auditing an MDR implementation, the following aspects must be verified:
- All classification scheme items have a unique identifier, name, and definition.
- Parent-child relationships form a valid DAG.
- Associative relationships, if used, are recorded with appropriate types (e.g., “synonym”, “broader”).
- Binding records exist between classification scheme items and administered items, and these bindings are accurate.
- Changes to classification items are tracked via change descriptions and effective dates.
4.3 Canadian Adoption Notes
As a CAN/CSA adoption, CAN/CSA-ISO/IEC 11179-2-06 is identical to the international edition except for editorial changes required by the Canadian standards body. Organizations in Canada are recommended to reference the CAN/CSA version for conformance with federal and provincial metadata policies. The technical requirements are unchanged from the ISO/IEC text.
CAUTION: Incomplete classification metadata — such as missing definitions or ambiguous names — can break data interoperability and lead to incorrect classification. Always validate that each classification scheme item meets the mandatory attribute requirements before production use.
Q: What is the difference between ISO/IEC 11179-2 and ISO/IEC 11179-3?
A: Part 2 (Classification) provides the model for classification schemes and their items, whereas Part 3 (Registry metamodel and basic attributes) defines the fundamental structure of the metadata registry itself, including administered items like data elements. In practice, Part 2 is used within the context of the Part 3 metamodel; a registry compliant with Part 3 can import classification structures defined per Part 2.
A: Part 2 (Classification) provides the model for classification schemes and their items, whereas Part 3 (Registry metamodel and basic attributes) defines the fundamental structure of the metadata registry itself, including administered items like data elements. In practice, Part 2 is used within the context of the Part 3 metamodel; a registry compliant with Part 3 can import classification structures defined per Part 2.
Q: Can I use ISO/IEC 11179-2-06 to represent a faceted classification?
A: Yes. The standard accommodates faceted schemes by allowing multiple classification schemes (one per facet) that operate independently. Items within each facet need not have hierarchical parents. The standard’s relationship types can then be used to link items across facets for compound classification.
A: Yes. The standard accommodates faceted schemes by allowing multiple classification schemes (one per facet) that operate independently. Items within each facet need not have hierarchical parents. The standard’s relationship types can then be used to link items across facets for compound classification.
Q: Are there any formal conformance testing programs for this standard?
A: While ISO/IEC 11179 parts are often used as design references rather than requiring mandatory certification, some governments and large enterprises require Level 1 conformance as part of procurement. Testing typically involves reviewing the registry’s data model and export files to ensure classification attributes conform to the mandatory set. There is no central ISO certification body for MDR implementations.
A: While ISO/IEC 11179 parts are often used as design references rather than requiring mandatory certification, some governments and large enterprises require Level 1 conformance as part of procurement. Testing typically involves reviewing the registry’s data model and export files to ensure classification attributes conform to the mandatory set. There is no central ISO certification body for MDR implementations.