the RefCo Toolkit - Archive ouverte HAL
Rapport Année : 2022

the RefCo Toolkit

Résumé

The RefCo Toolkit is a toolkit for improving and evaluating the quality of oral language documentation. It provides quality criteria specifications as well as a process to be implemented by any certification entity interested in improving the quality of their language documentation corpora.

The Corpus Documentation is a spreadsheet which should be filled by the corpus submitter, that is someone who either has created a corpus or is curating it and wishes to have it evaluated and improved according to the quality criteria defined by RefCo. A previous version, unpublished, of the Corpus Documentation was written by Kilu von Prince. Even if the current version evolved in a different direction, it benefited a lot from von Prince's work.

The Reference Manual describes each item present in the Corpus Documentation. It is intended both for a corpus submitter when filling the Corpus Documentation.

The Review List contains the list of the different items that should be checked by an entity implemented the RefCo quality criteria. These items can be evaluated either manually or automatically.

Finally, the CheckList Report is the form a reviewer should fill after the evaluation of the corpus by the certification entity, in order to help the corpus submitter to fix the issues observed in the corpus, and keep track of the whole process.

For the RefCo Checker and the semi-automatic corpus evaluation it provides, see https://gitlab.rrz.uni-hamburg.de/bba1792/corpus-services/-/blob/develop/doc/README.RefCo.md. The .json and the .xsd files are formal schemes describing the RefCo Corpus Documentation.

Domaines

Linguistique
Fichier non déposé

Dates et versions

hal-04235030 , version 1 (10-10-2023)

Licence

Identifiants

Citer

Jocelyn Aznar, Frank Seifart. the RefCo Toolkit. ZAS. 2022. ⟨hal-04235030⟩
33 Consultations
0 Téléchargements

Altmetric

Partager

More