Are Research Datasets FAIR in the Long Run?

Abstract

Currently, initiatives in Germany are developing infrastructure to accept and preserve dissertation data together with the dissertation texts (on state level – bwDATA Diss¹, on federal level – eDissPlus²). In contrast to specialized data repositories, these services will accept data from all kind of research disciplines. To ensure FAIR data principles (Wilkinson et al., 2016), preservation plans are required, because ensuring accessibility, interoperability and re-usability even for a minimum ten year data redemption period can become a major challenge. Both for longevity and re-usability, file formats matter. In order to ensure access to data, the data’s encoding, i.e. their technical and structural representation in form of file formats, needs to be understood. Hence, due to a fast technical lifecycle, interoperability, re-use and in some cases even accessibility depends on the data’s format and our future ability to parse or render these.

This leads to several practical questions regarding quality assurance, potential access options and necessary future preservation steps. In this paper, we analyze datasets from public repositories and apply a file format based long-term preservation risk model to support workflows and services for non-domain specific data repositories.

1

BwDATADiss-bw Data for Dissertations:https://www.alwr-bw.de/kooperationen/bwdatadiss/

2EDissPlusDFG-Project – Electronic Dissertations Plus:https://www2.hu-berlin.de/edissplus/