Practical Challenges and Considerations in Designing Digital Object Identifiers for Data
Clark, A., Ahern, T., Trabant, C., Newman, R., Benson, R.
While the basic idea of Digital Object Identifiers (DOIs) is fairly simple, a robust implementation must consider a number of factors that are not necessarily obvious at first glance. This is especially true when attempting to apply the DOI framework — which was designed for use with published works — to datasets for which concepts like identity, provenance, and lifecycle may be much more complex or ambiguous.
The IRIS Data Management Center (DMC) recently undertook an effort to build DOIs for various parts of its data holdings, and we share some of the particular challenges and solutions that emerged from this process, including:
- The often conflicting priorities of the various interested parties (the data creator, the data curator, the researcher citing the data, and the reader following the citation back to its source),
- The need to define a proper data granularity, and the options for representing data subsets, variations, and collections,
- The advantages and disadvantages of DOI opacity,
- The challenges imposed by the DOI framework’s fundamentally unidirectional nature,
- Building a scalable and maintainable system for serving DOI target URLs.