Select/define annotations * (p.
)
Integrate annotations into the data pipeline * (p.
)
Always produce some kind of orthographic transcription:
Define/select the orthographic transcription *(pp.
,
)
Set up the transcription rules/method *(p.
)
Define the delivery format of the transcript *(p.
)
Choose/program the tools for transcription *(pp.
,
)
Train the group of transcribers *
Set up check procedures for the transcription *(p.
)
Test for inter-transcriber agreement ***(p.
)
For each other annotation type, tagging (p. ) or
segmentation (p.
):
Define the annotation contents and rules *
Define the delivery format of the annotation *
Choose/program/test the tools for annotation *
Train the labelers *
Set up check procedures *
Test for inter-labeler agreement *** (p.
)