FR ELAN structure

ELAN structure standard used by the Freiburg Research Group in Saami Studies

name	parent	type	stereotype	obligatory	LANG_REF attribute	purpose
ref@	root	refT	time-aligned	x		Uniquely identifies annotation (utterance for a speaker); pattern: .0024?
.orth@	ref	orthT	symb.-assoc.	x	x	Standard orthography; input for FST-script
..word@	orth	wordT	symb.-subdiv.	x	x	Wordform; Extracted by FST-script; input for FST
...lemma@	word	lemmaT	symb.-subdiv.	x		Lists lemma/possible lemmas; output from FST
....pos@	lemma	posT	symb.-subdiv.	x		Lists part of speech for lemma; output from FST
.....morph@	pos	morphT	symb.-subdiv.	x		Lists values for morphological categories (Giellatekno-style); output from FST
.....gloss@	pos	glossT	symb.-assoc.			English gloss for each PoS of each lemma; extracted from sje lexical database; currently only used for sje
..ft-lang@	orth	ft-langT	symb.-assoc.			Free translation of orth; ‘lang’ is replaced with specific language (eng=english, deu=deutsch -- use ISO-codes?); can have multiple derivations
..orth-orig@	orth	orth-origT	symb.-assoc.			Orthography used in the original source

Each ref annotation must have content (which content?) and it has to be unique (usually a unique number identifier such as .001)
Each tier (except for general notes tier?) must have participant, uniquely identified after @ in tier name; standard for naming participant can be project-specific

Niko: add orig-convention to places where it is needed
Anyone: come up with some solution to resolve note-tiers, for example three types assigned to currently existing (check this) notes:
what to do with tiers for annotations with notes ⋅⋅+note-refT ⋅⋅+note-orthT ⋅⋅+note-wordT
what to do when there are more than one orth-orig? (e.g. Wickman's sje-text with phonetic (FUPA), phonemic (more-less Swedish-Saami), AND adapted north-Saami orthography)