Overview
SIST ISO 24610-1:2013 / ISO 24610-1:2006 specifies a standardized format for representing, storing and exchanging feature structures used in language resource management and natural language processing (NLP). Part 1 defines the concepts, informal semantics and an XML-based vocabulary for untyped feature structures, feature values and feature libraries. It also provides a computer-readable way to describe constraints on features and values so implementations can check conformance against a reference specification.
Keywords: feature structures, language resource management, XML representation, ISO 24610-1, linguistic annotation, natural language processing.
Key Topics and Technical Requirements
- General characteristics and notations: formalizes graph, matrix and XML-based notations for feature structures.
- Structure sharing (re-entrancy): representation of shared substructures to model linguistic phenomena efficiently.
- Collections as feature values: support for lists, sets and multisets as complex feature values.
- Typed feature structures (overview): discussion of types, type notation and type-related constraints (with full typed validation addressed in Part 2).
- Relations and operations: formal treatment of subsumption, compatibility, unification (including shared structures) and operations on feature values such as concatenation, union, alternation and negation.
- XML Representation: a standard XML vocabulary and organization for elementary and complex feature structures, feature/value libraries, default values, and linking analysis to text.
- Formal and implementation materials (Annexes): formal definitions, RELAX NG module, illustrative examples, type inheritance guidance and denotational semantics to aid implementation and validation.
- Normative references: integrates existing standards such as ISO 8879 (SGML) and ISO 19757‑2 (RELAX NG) for schema and validation.
Practical Applications and Users
- NLP engineers and computational linguists: interchange of parsed or annotated linguistic structures across tools and pipelines.
- Corpus and annotation projects: storing multi-layer annotations (morphosyntax, semantics, phonology) with shared substructures and complex values.
- Tool developers and library authors: implementing parsers, unifiers or feature-based grammars consistent with a standard XML interchange format.
- Language resource managers and standards teams: ensuring long-term interoperability of language resources and validation of feature-structure conformance.
- Linguists and lexicographers: encoding feature-based descriptions for research and publication.
Related Standards
- ISO 8879 - SGML (reference for markup principles)
- ISO 19757-2 - RELAX NG (used for schema/validation)
- TEI (Text Encoding Initiative) - collaborative origin of feature-structure chapters
SIST ISO 24610-1 is a practical, interoperable foundation for representing feature structures in linguistic data exchange, validation and tool integration.