ELANtype

ELANtype is a software framework and data model designed for creating and managing typed, time-aligned annotations on time-based media such as audio and video. Building on the ELAN annotation environment, ELANtype introduces a formal type system to describe annotation schemes, allowing researchers to define reusable, interoperable annotation types and constraints.

Core concepts include primitive types (TimePoint, TimeInterval, Annotation, Tier) and a hierarchy of composite types for linguistic and non-linguistic annotations. Annotations reference time intervals and can carry features as key-value pairs. Types can be extended through inheritance, enabling domain-specific schemes to share a common core. The framework emphasizes explicit schema definitions, validation, and documentation, aiding consistency across projects.

Data encoding and interchange: ELANtype provides serialization in XML and JSON, with schemas to validate conformance. It interoperates with TEI/XML ecosystems and can export to standard formats used in corpus linguistics. It can be used as a stand-alone library or as a plugin to existing ELAN installations.

Applications: It is used to annotate multimodal data, including phonetics, discourse, gesture, and alignment studies. The typed approach helps researchers compare annotation schemes and automate processing such as alignment verification and statistics across corpora.

Development and reception: Developed by a collaborative community of linguists and computer scientists; aims to improve rigorous data modeling, reproducibility, and long-term interoperability of annotations. Some practitioners caution that the learning curve for the type system and schema design can be steep.

See also: ELAN, TEI, annotation, corpus linguistics.