Integrated collection, storage and archiving of research data and associated research project information via Data Management Plans

Presentation at CERN conference PV2023: Adding value (to) and preserving Scientific & Technical data Using DMPs – Data Management Plans as a central node or “hub” – in accordance with the RDA DMP Common Standard model (https://github.com/RDA-DMP-Common/RDA-DMP-CommonStandard) would allow for integration of research information of different types. One way of doing this would be to collect directly from the source as much as possible of available research project information, including research data and code, packaging and storing everything together as a (pre-)SIP for (later) transfer to a proper OAIS digital archive, where the SIP can be transformed to an AIP, an Archival Information Package, while also generating descriptions in relevant local records management systems. For much of this research information, retrieval needs to be iterated on multiple occasions. An example could be the DMPs created by means of the machine-actionable DMP (maDMP) template in use at Stockholm University (SU- EOSC Nordic project and maDMP template described in Dataverse: https://doi.org/10.7910/DVN/MGZBAL). These DMPs are now being retrieved on two separate occassions, one for the initial version and later the final version at the end of the research project. The assumption is that the research information, data and code described in a DMP might change, be updated and completed over time. We believe the most efficient workflow is to collect research information and datasets that are described in a DMP directly from the source on a suitable occasion, e.g. when DMPs are archived anyway. In this way we also believe research information and datasets could more easily be retrieved and monitored as compliant with file formats and metadata standards that enable further integration, processing and transformation to archival formats in accordance with the OAIS standard (SIP/AIP/DIP). There are multiple potential benefits from keeping relevant associated research information and datasets stored in the same “packages”: 1) avoid duplication of effort when retrieving, especially when retrieval needs to be iterated; at each retrieval event the corresponding package (pre-SIP) will offer an overview of what kind of information has already been collected/retrieved, 2) it will also give an overview of available file formats and metadata standards, 3) for future dissemination (DIPs) of any or all research information or datasets in a project it will be much easier to search and find the desired information ”in the same place”; particularly if you also keep copies e.g. of PDFs in plain text formats (such as txt, csv, md, json, xml), thereby enabling more efficient full text search in larger data archives or file libraries e.g. by means of BASH grep. 4) enhanced (meta)data quality control of separate files, by means of comparison with associated information, thereby potentially promoting preservation efforts and adding value to scientific data