Research data publication: Moving beyond the metaphor

Metaphors are a quick and easy way of grasping (often complicated) concepts and ideas, but like any useful tools, they should be used carefully. There are as many arguments about how datasets are like cakes1 as there are about how datasets aren’t like cakes.2 It can be easy to categorise a dataset as being a special class of academic paper. Positively, this means that the tools and services for scholarly publication can be utilised to transmit and verify datasets, improving visibility, reproducibility, and attribution for the dataset creators. Negatively, if a dataset doesn’t fit within the criteria to meet the “academic publication” mould (e.g. because it is being continually versioned and updated, or it is still being collected and will be for decades) it might be considered to be of less value to the community. It is often said that “all models are wrong, but some are useful” (Box, 1979). Hence we need to determine the usefulness and limits of models and metaphors, especially when trying to develop new processes and systems. This paper further develops the metaphors for data outlined in Parsons and Fox (2013), and gives real world examples of the metaphors from scientific data stored in the Centre for Environmental Data Analysis (CEDA) – a discipline-specific environmental data repository, and the processes that created the datasets. © 2019 The Author(s).