This is the third and final blog post in a three-part series on understanding the perspective of various stakeholders across the scholarly communication landscape regarding the adoption of persistent identifiers (PIDs) and machine-readable data management plans (DMPs) prior to the 2019 conference on “Implementing Effective Data Practices.”
In an effort to make these perspectives more widespread, we are sharing excerpts from these interviews and discussing them in the context of the final conference report and recommendations for machine-readable DMPs and persistent identifiers (PIDs).
The first blog post in this series provided a vision for what an ecosystem of connected research outputs would bring to researchers, funders, and institutions. The second blog post outlined a variety of perspectives on what the conference hoped to address and some expected outcomes.
This third blog post highlights a recurring theme we heard from stakeholders about wanting more than just high-level recommendations for implementing effective data practices. Conference attendees wanted specifics about what it would take to bring the vision described in our first blog post into existence. The interview clips below include ruminations on what would be necessary at the infrastructure level and at the item level to make DMPs machine-actionable, ongoing questions about the granularity at which PIDs are assigned, and the need for broad standardization across DMP requirements from funding agencies.
Getting down into the Details
Many of the interviewees wanted to get into the details on what a machine-readable data management plan would contain—or which parts would be machine-readable. Cliff Lynch, executive director, Coalition of Networked Information, discussed the need for a clearer understanding of these specifics.
Benjamin Pierson, then senior program officer, now deputy director for enterprise data, Bill and Melinda Gates Foundation, also expressed an interest in getting down to brass tacks, actionability, and specificity for machine-readable DMPs and PIDs.
Articulating these details is not without challenges though. As Kerstin Lehnert, Doherty Senior Research Scientist, Columbia University, states, the granularity of assigning PIDs often depends on the audience and use of the data set.
Additionally, as Lisa Johnston, research data management/curation lead, University of Minnesota, indicates, the content of DMPs needs to standardize in order to ensure they provide the needed information to realize the vision.
All of the pre-conference interviews are available on the ARL YouTube channel.
Natalie Meyers is interim head of the Navari Family Center for Digital Scholarship and e-research librarian for University of Notre Dame, Judy Ruttenberg is senior director of scholarship and policy for ARL, and Cynthia Hudson-Vitale is head of Research Informatics and Publishing for Penn State University Libraries.