inter-views curation of interview data 1 feb. – 1 nov. 2010 clst, nijmegen,, // henk van den...
TRANSCRIPT
“INTER-VIEWs”Curation of Interview Data
1 feb. – 1 nov. 2010
CLST, Nijmegen,, http://www.ru.nl/CLST
Henk van den Heuvel
Centre for Language and Speech Technology (CLST)
Radboud Universiteit Nijmegen
21 sep. 2010
Overview
1. Metadataa. Sources
b. Point of departure
c. Procedure
2. Experiences so far & questions
INTER-VIEWsCLST, Nijmegen, http://www.ru.nl/CLST
Metadata (excel sheet)
a. Sources- IPNV, VT-VP, VI, DANS, DC, CMDI
b. Points of departure- SpeechCorpusProfile (full corpus)- SpeechCorpusProfile_Autonomata profile (per interview)
c. Procedure- Make Interviews profile in CMDI Comp. Registry editor
- SpeechCorpusProfile_interviews- SpeechCorpusProfile_interview
- Report any new categories to ISOcat(.org)- Make metadata schema from profiles- Fill schema for individual interviews using Arbil
INTER-VIEWsCLST, Nijmegen, http://www.ru.nl/CLST
2. Experiences & questionsA. Some elements have a fixed value for all interviews. Can we already fix this value in the
profile?
B. When entering the meta data values: can you leave elements in a component empty? Even if you have specified that the element should occur at least once.
C. In our workspace components are not hierarchically ordered, but they are all in line under each other. However in the public space we see a hierarchy in the registered examples. How come?
D. Elements in ISOcat often are just names to which you can add a string as value. This gives a lot of freedom and possibilities to divert from the original meaning. Should you introduce a new category as soon as you think it differs from the existing element?
E. Can Arbil import metadata values from import files and put these into metadata file for individual interviews
INTER-VIEWsCLST, Nijmegen, http://www.ru.nl/CLST