ELAN Annotation Format
version 3.0
December 2016
Schema by Alexander Klassmann 17/01/03
Adapted by Hennie Brugman, Han Sloetjes, Micha Hulsbosch
The entry id should be unique within the collection of entry elements
The Tier name/id should be unique within the collection
of Tier elements
A Tier can be associated with a parent Tier by referring to an existing Tier id.
The Linguistic Type name/id should be unique within the collection
of Linguistic Type elements
A Tier must refer to an existing Linguistic Type id.
The Controlled Vocabulary name/id should be unique within the
collection of Controlled Vocabulary elements
A Linguistic Type can be associated with a Controlled Vocabulary by
referring to an existing Controlled Vocabulary id.
The Lexicon Service name/id should be unique within the
collection of Lexicon Service elements
A Linguistic Type can be associated with a Lexicon Service by
referring to an existing Lexicon Service id.
A key and keyref pair to enforce that a previous annotation idref at least refers
to an annotation id of a reference annotation.
Two key-keyref pairs to enforce that time slot references refer to the id of a time slot.
The ID of a language identifier, can be referred to by any element that
needs a reference to a language identifier.
Reference from a value in a multilingual CV to a language identifier.
Reference from a description in a multilingual CV to a language identifier.
Reference from a tier to a language identifier, to indicate the (main) language recorded
on that tier.
Reference from an individual alignable annotation to a language identifier.
Reference from an individual reference annotation to a language identifier.
This attribute is deprecated. Use MEDIA_DESCRIPTOR elements instead.
Ignore
This is in fact a reference to the parent annotation.
A reference to an url of an external Controlled Vocabulary.
Is intended to be mutually exclusive with a sequence of CV_ENTRY_ML elements.
An entry in a multilingual controlled vocabulary, containing the values and the descriptions
in multiple languages.
A controlled vocabulary entry value with a language attribute.
This allows multilingual controlled vocabularies. It adds a language reference attribute
compared to the mono-lingual cv entry element.
A description element with a language reference attribute.
A reference to the id of an ISO Data Category (url including id).
A reference to an external (closed) Controlled Vocabulary (url).
A reference to the id of an Entry in an external Controlled Vocabulary (id).
A reference to the id of an entry in a lexicon (url, url+id or id)
A reference or hyperlink to any type document (url)
The Language element containing a reference to a language name or (if possible persistent) definition.
ISO-639-3 still seems to be the best choice for language codes and closest to persistent language ID's
seem to be the http://cdb.iso.org/lg/... identifiers also used by the iso-language-639-3 component in
the CLARIN ComponentRegistry?
The license element can be used to include license information in the eaf file itself.
A set containing referential links.
A set can contain both cross-references and grouping referential links.
Apart from an ID the set can have a meaningful, "friendly" name.
A set can have an external reference, a language and a CV reference.
A cross reference is a referential link between two existing elements (REF1 and REF2).
Each of these elements can be either an annotation or a referential link.
Optionally the direction of the link can be specified.
A referential element for grouping any number of existing elements (the REFS).
Each element can be an annotation or a referential link.
Attributes common for both cross- and group references.
Apart from an ID it is possible to associate a meaningful, "friendly"
name to the link. Furthermore a link can have an external reference, a language and a
CV entry reference and a type attribute.
An attribute that allows to specify the type of the cross- or group reference/link.
The entry id should be unique within the
collection of entry elements