Referencing: The reuse factor : Nature News & Comment

"Researchers are still struggling to find, manipulate and cite research outputs other than published papers. Data-management plans for research — detailing what data will be created and how, and outlining plans for data sharing and preservation — are a core requisite of all grant applications for a long list of US and UK funding agencies. These include the US National Science Foundation, the US National Institutes of Health, NASA, the UK Biotechnology and Biological Sciences Research Council, the UK Medical Research Council and the Wellcome Trust. Funders require grantees to make all their products available in a citable, sharable and discoverable manner. As a result, tools such as the California Digital Library's Data Management Plan and the UK-based Digital Curation Centre's DMPonline have materialized to help optimistic grant applicants to fill in this section. Even with the help of such tools, data-management plans are being rejected in some grant applications. This is a wake-up call. Researchers should be long past thinking that depositing their data in a file-hosting service such as Dropbox is sufficient. Yet the majority of academics still consider journal articles to be the only valid, formal record of their research — the main currency for credit. In my view, the current model of grouping non-image files together in supplemental data appended to a paper is laughable. Data are the bedrock of a paper and come in myriad forms; for instance, videos for research on cell motility or biomechanics, or code for climate or epidemiology models. Scholars are increasingly sharing these sorts of raw data through repositories such as the Dryad Digital Repository and GenBank, which allow for the citation of data sets, videos, genetic sequences and other such files that publishers often struggle to accommodate. Nature special: Impact In 2011, I set up another such company, figshare, to improve the way that the 'data behind the graph' is disseminated, exploiting visualization tools such as D3.js and Jmol1. At figshare, we work with research groups and publishers to make data reusable, reproducible and interactive. (Macmillan Science and Education, the publisher of this journal, is an investor in figshare.) But the generation of huge numbers of citable research outputs is confusing researchers who are accustomed to citing only papers. The FORCE11 Amsterdam manifesto on data-citation principles, drawn up in 2011 by a community of scholars, librarians, archivists, publishers and research funders, states: 'A data citation in a publication should resemble a bibliographic citation and be located in the publication's reference list.' A quick look at the most recent journal citations of data held in figshare shows that this recommendation is not enforced by publishers or authors: only one in five cited the data in the reference list; the rest mentioned it in methods or deposition sections ..."