Replikation von Sprachdokumentationssammlungen für eine bessere Wiederverwendbarkeit und Aufbewahrung von Daten

Poster des Text+ präsentiert bei der Herrenhäuser Konferenz vom Rat für Informations- infrastrukturen (RfII) und VolkswagenStiftung am 24. und 25. April 2023 in Hannover, Datenräume in Deutschland und Europa gestalten - Impulse der Wissenschaft: The DOBES Archive contains language documentation data for a great variety of languages from around the world that are in danger of becoming extinct. The data collections that resulted from the DOBES endangered languages documentation programme are preserved and made available via The Language Archive at the Max Planck Institute for Psycholinguistics in Nijmegen. It is evident that the data in the DOBES archive are valuable scholarly assets whose wide and long-term availability will contribute considerably to the preservation of some of the world's cultural heritage, and to research in linguistics and other fields. While the DOBES collections are of interest to a broad scholarly community, they are currently made available form a single repository and in formats that are widely used among language documentation researchers. Having the data available in an additional repository with a different focus and in a standardized pivot format that allows for easier automatic migration to a variety of other formats will make the data more easily usable to a larger community within the humanities as well as in other domains. In addition, it has the potential to connect the data to a broader range of tools and services, e.g. within the planned Text+ Research Data Infrastructure in Germany. Our project is aimed at creating an additional copy of the DOBES collections at the Leibniz-Institut für Deutsche Sprache (IDS) in Mannheim, and at converting the textual data into a pivot format: TEI for Spoken. In the process, various technical, organizational, legal and ethical aspects of the migration will be thoroughly investigated and documented by a multidisciplinary team of researchers. In order to create a functional mirror of the DOBES collections at the side of the IDS, the following aspects are being worked on: (a) conversion of the structured text formats into a pivot format (TEI for spoken language); (b) improvement and unification of documentation for each of the collections (e.g. regarding the use of annotation formats and conventions); (c) improvement of metadata; (d) clarification of the legal situation and the creation of agreements that will allow the creation of the mirror.