SBN DOI Policy and Philosophy
The SBN creates DOIs for the following reasons:
- To give appropriate credit to the creators of research data sets accepted into our archive
- To provide reproducibility and provenance for research based on products in our archive holdings
- To increase the discoverability and reusability of our data sets for current, potential, and future users
The ADS lists SBN archived data sets as "refereed", which is a designation that carries with it some obligations on our part. In light of this, when SBN is creating DOIs primarily to give appropriate credit to creators it will:
- Strongly advise data preparers to consider authorship as a key criterion for determining the design of collections and bundles;
- Assign DOIs in a way that does not artificially inflate the "refereed" publication count;
- Update DOI metadata to reflect citation and other relationships between existing data sets and data sets created subsequently;
- Notify the ADS of the newly issued and updated DOIs; and
- Never assign a DOI to any bundle or collection that does not identify all of its members by LIDVID.
Reproducibility and Provenance
When authors have selected products from the archive and are required to cite them by DOI, SBN will create a secondary collection that identifies the list of selected products, and assign a DOI to that collection. In this case the primary purpose for the DOI is reproducibility and establishing provenance of the data used in the related analysis. When SBN is creating a DOI primarily for reproducibility and provenance it will:
- Flag this DOI as "non-refereed" to the ADS;
- Cite the source data set(s) in the DOI metadata;
- Include the DOI of the work to be published in the collection DOI metadata; and
- Update the source data set(s) metadata to include a "cited by" notation for the new DOI.
Discoverability and Reusability
SBN aspires to the FAIR data principles, and therefore seeks to collect as much high-quality metadata for its DOI records as feasible. Therefore, when creating DOI metadata for any reason, that metadata will include:
- Citations of source materials (journal articles, other data sets, etc.) in the DOI metadata;
- Digital identifiers for creators (e.g., ORCID) and agencies (e.g., ROR) where available;
- Size and format information, in multiple units and using standard terminology (like MIME types) where applicable;
- An appropriate license notice; and
- Rich metadata, including subject keywords, meaningful abstracts, and the identification of contributors beyond the traditional author/editor roles.