Overview
ISO 12620-2:2022 provides criteria and governance practices for managing data category repositories (DCRs) used in the creation and maintenance of language resources. The standard defines roles, responsibilities and procedures for establishing and operating a DCR within a community of practice (CoP), including how to add new data category specifications, run continuous quality assurance, and support interoperability of terminological, lexicographical and annotated corpora.
Key topics and requirements
- Scope and purpose: Management of data categories for language resources to improve interoperability and reuse.
- Governance roles: Clear definitions of CoPs, profile experts, profile managers, profile management groups, and a DCR management board responsible for oversight and publicity.
- DCR functional requirements:
- Electronic availability (online or limited environment)
- Central collection of data category specifications
- Mechanisms to prevent duplicate specifications
- Submission and feedback processes for new or revised data category specifications
- Search filters (by date, creator, fields) and export capabilities (e.g., CSV, XML)
- User access controls for write permissions
- Ability to define data category selections (DC selections) and export them
- Subsetting based on an ontology of data category concepts
- Profiles and selections:
- Support for profiles (subsets of DCR entries managed by subject-matter experts)
- DC selections to define application-specific sets of data categories in combination with a data model
- Quality and workflow:
- Procedures for specification lifecycle, governance, and continuous quality assurance (workflow for specification and maintenance is defined in the standard)
- Public information: Guidance for transparency and interaction with stakeholders
Applications and practical value
ISO 12620-2 is directly applicable to:
- Terminology managers and lexicographers building consistent field names and value sets
- NLP engineers and corpus developers who need interoperable metadata and annotation schemas
- Software developers and data architects implementing exchange formats or terminology databases
- Research communities and CoPs seeking a governed, shared source of vetted data categories
Practical benefits include improved interoperability, faster integration of heterogeneous language resources, reduced duplication of schema elements, and clearer governance for evolving data categories.
Related standards
- ISO 12620-1:2022 - Data category specifications (structure and rationale)
- ISO 30042 - TBX and data model considerations (referenced for module/dialect composition)
- ISO 16642 - Metamodel referenced for terminology exchange
- DatCatInfo is cited as an example DCR for language resource descriptions.
Keywords: ISO 12620-2:2022, data category repository, DCR, data category specifications, terminology resources, profiles, data category selection, governance, interoperability.