ISO 24615-2:2018 PDF

Overview

ISO 24615-2:2018 - Language resource management - Syntactic annotation framework (SynAF) - Part 2: XML serialization (Tiger vocabulary / ISOTiger) defines an XML-conformant serialization of the ISO 24615‑1 SynAF meta‑model. The standard specifies a language- and theory‑independent XML format (based on TigerXML) to promote interoperability between syntactic annotations, treebanks, syntactic parsers and language processing components.

Key topics and technical requirements

• Meta-model instantiation: Concrete XML element and attribute names mirror the ISO 24615‑1 meta‑model (graph, nodes, edges, annotations).
• Graph-based representation: Syntactic annotations are modeled as graphs G = (V, E, A) with nodes (terminals and non‑terminals), directed edges and feature-value annotations.
• XML vocabulary (ISOTiger): Core elements and conventions include:
  • (segment) containing one or more elements
  • with elements (terminal nodes)
  • with elements (non-terminal nodes)
  • elements as children of or with @target to reference nodes
  • Attribute conventions such as @word for token text
• Primary data handling: Two allowed representations
  • Sequential representation - primary data included inside elements (single XML file)
  • Standoff representation - primary data referenced externally (e.g., MAF files), enabling modular architectures
• Annotation tagsets and features: Tagsets expressed via Feature and FeatureValue constructs; categories may reference standardized data categories (DatCatInfo) via URIs in compliance with ISO 12620.
• Namespaces and interoperability: XML elements use the CLARIN SynAF namespace (http://www.clarin.eu/standards/ns/synaf), and the format is coordinated with related ISO frameworks.

Applications and users

ISO 24615-2:2018 is practical for:

• Corpus linguists and treebank curators building or exchanging syntactically annotated corpora
• NLP engineers and tool developers implementing parsers, converters or visualization tools that must interoperate with standard treebanks
• Digital humanities projects that need standardized syntactic markup for multilingual resources
• Research groups comparing annotation schemes or combining annotations from different sources

Benefits include improved reuse of syntactic annotations, easier tool integration, and language/theory neutrality for syntactic data exchange.

Related standards

• ISO 24615‑1 (SynAF meta‑model)
• ISO 24612 (Linguistic Annotation Framework, LAF)
• ISO 24611 (MAF - morpho‑syntactic annotation)
• ISO 24610 (Feature structures)
• ISO 12620 (Data category specifications / DatCatInfo)

Keywords: ISO 24615-2:2018, SynAF, ISOTiger, TigerXML, XML serialization, syntactic annotation, treebank, language resource management, standoff representation, sequential representation.