/ Digits: Two Reports on New Units of Scholarly Publication


The Digits team (Matt Burton, Matthew J. Lavin, Jessica Otis, and Scott B. Weingart) convened around the question of how we might share, preserve, and legitimize scholarship freed from the affordances of print. For the A.W. Mellon-funded Digits Planning Grant (2016-2018), the PIs had three goals:

  1. Investigate the use of software containers for research in the sciences, social sciences, and humanities.
  2. Assess the infrastructural needs of digital humanists around publishing and preserving web-centric scholarship.
  3. Gather a team of experts to guide the above activities and plan how they might inform a beneficial intervention into the scholarly ecosystem.

Through our investigation into the scholarly uses of containers, we discovered that the technical infrastructure needed to connect containers with digital publications is underdeveloped. We see potential for container technologies to facilitate existing digital scholarly publications and afford new forms of computational scholarship, but this process would first require a series of infrastructural bridges. The digital scholarship needs assessment we conducted, as well as our advisory board meetings, made it clear that a targeted technological intervention alone would not be enough to welcome web-first publications into the scholarly ecosystem; in-tandem cultural and institutional changes are also necessary.

The first and second of our three tasks resulted in the two reports that comprise this article. The first report, A New Unit of Publication: The potential of software containers for digital scholarship, involved an environmental and secondary source scan of activities at the intersection containerization and scholarship. The second report, New Scholarship in the Digital Age: Making, publishing, maintaining, and preserving non-traditional scholarly objects, summarizes 75 interviews of scholars, technicians, publishers, and others who work towards the publication of digital-first scholarship.

Both reports were presented to the Digits advisory board, including Laurie Allen (Penn), Lauren Brumfield (Reclaim Hosting), Dan Cohen (Northeastern), Dan Evans (CMU), Martin Paul Eve (Birkbeck College, London), Ilya Kreymer (Rhizome), Alison Langmead (Pitt), Sharon Leon (MSU), David Newbury (Getty), Andrew Odewahn (O’Reilly Media), Mary Shaw (CMU), Ammon Shepherd (UVA), Ed Summers (UMD), Whitney Trettien (Penn), Amanda Visconti (UVA), Keith Webster (CMU), and David Wilkinson (Pitt).

Several lessons became apparent over the course of the grant period.

Our container study revealed myriad use-cases of containers in academia that are research oriented. Container adoption is much wider in the sciences than in the humanities, especially for dependency management and reproducibility.

Our findings also suggest that containers are used in teaching, but we didn’t fully investigate this topic. Lastly, we found few if any efforts focusing specifically on containers as a unit of scholarly publication at the time of conducting our research. In the past several years, however, some additional examples have arisen or come to our attention, including Binder and NextJournal.

Commenting on the use of a software container as a unit of scholarly publication, the advisory board stressed the importance of starting with templates or examples when creating digital platforms, of making working prototypes before over-theorizing, and of creating a platform that fits easily within the current publication ecosystem. The board further suggested a project on the long-term costs associated with creating, hosting, and maintaining digital scholarly objects would be critically useful to efforts in this space moving forward.

The first advisory board meeting settled on four important elements that a container-based intervention would need to encompass to be successful: setting specifications, creating a production environment, facilitating a publication platform, and designing with preservation as a top priority.

Our DH infrastructure study reinforced our sense of how many factors impede digital scholarship, as well as how deeply these impediments run. The diverse ecosystem around digital publications adds friction to their existence at every stage of their lifecycle, particularly at points where a digital object is transferred from one party to another. After conducting our study, we concluded that the best way to solve the identified problems would be to target every stage of the digital scholarly workflow in concert. Although these interventions, ideally, would occur all at once, various piecemeal interventions could radically improve the ability of scholars to create, publish, preserve, and receive recognition for their digital work. Some of the interventions we discuss are technical, but at least as many are entirely social. Even where technical solutions are to be found, implementing them will require scholarly buy-in and a willingness to adapt existing scholarly practices.

The second advisory board meeting echoed the findings of the infrastructure report, and additionally offered some next steps:

  1. Constructing the technology and standards for a self-contained digital scholarly object.
  2. Developing plugins for pre-existing authorship, publication, and preservation platforms to allow for the easy transfer of complex digital objects from one stage to the next.
  3. Developing a tool that can encapsulate a given system stack and solicit metadata on the scholarly object within in order to create a digital object conforming to the agreed upon standards.
  4. Creating or fostering sample publications that use the proposed technologies to act as a lightning rod to encourage wide adoption.
  5. Working with publishers and institutions to adopt and support these standards.
  6. Teaching scholars and creators to build towards these standards.
  7. Encapsulating or easing the encapsulation of several popular platforms for digital publication, to foster a broader adoption of these standards.

The combined expertise and experience of the advisory board stressed the difficulty of such an orchestrated effort, as important as it is.

The irony is not lost on us that, at the start of this collaboration, we often spoke of orchestration, but this term had a specific and technical meaning for us related to products like Docker and Kubernetes. By the time we completed our work, we were speaking almost exclusively of orchestration as a sociotechnical concept.

We remain committed to the idea that containerization, or a similar lightweight encapsulation technology, has an important role to play in the future of scholarly publication. With the completion of the Digits grant, however, we have come to believe that containerization will only ever be one piece of a much larger puzzle. We hope the ensuing two reports will be useful in revealing that puzzle’s ultimate picture.

We would like to thank everyone at the A.W. Mellon Foundation, particularly Patricia Hswe, Michael Gossett, and Donald J. Waters, for their generous support, guidance, and feedback. We would also like to thank the 75 interview subjects and the members of our advisory board for their thoughtful insights.