Appendix

The infrastructure manifests included for the chapters in this volume are an experiment to explore what would be required for those working in the digital humanities (DH) or related areas to identify the principal infrastructures used for their professional work. (See what follows for a list of manifest files and their template file.) Each manifest represents the infrastructure used by one or more of the authors of each chapter in researching, preparing materials for, and producing their contributions to this book. However, two of the manifests—for chapters 2 and 8—represent additionally or instead the infrastructures of the platforms, services, or tools that they discuss as topics, thus showing how the manifest concept can be used generally to document infrastructure.

In the etymological sense discussed by Matthew Hockenberry at the opening of his “Manifesting Connection” (chapter 4), “manifests” are about making “something ‘evident’ or ‘palpable.’” But unlike the Early Modern shipping manifests that Hockenberry begins by citing (e.g., those defined in the Merchant’s Magazine in 1697 as “Transcript[s] of a Master of a Ship’s Cargo, showing what is due to him for Freight from each person to whom the Goods in his Ship belong”), the infrastructure manifests in this book are intended to make evident both the payload of scholarship (as it were) and the nature and cost of everything that from beginning to end does the loading, conveying, unloading, and anything else—that is, the infrastructures for extracting, shaping, mixing, connecting, supporting, framing, storing, transporting, labeling, and otherwise making, structuring, and communicating DH scholarship. After all, the premise of Critical Infrastructure Studies and Digital Humanities is that, however abstract scholarship may appear when it finally arrives on the receiving dock, it is always also a matter of actual materials, artifacts, organizations, and embodied intellectual and physical labor—everything and everyone needed, for instance, to create the physical copy of the present book and the online version of the same book.

In its present version (1.0), the template that we created for the infrastructure manifests in Critical Infrastructure Studies and Digital Humanities is limited in form and purpose. We designed it in consultation with authors in our volume just to start learning what such manifests should be. The template, which we asked at least one creator of each chapter to fill out as a spreadsheet, is prestructured with suggestions for some kinds of infrastructure to consider. There are rows for infrastructures under the following labels: “land,” “materials,” “energy,” “transportation,” “architectural,” “civic, community, national, or regional,” “institutional,” “labor,” “research-content,” “tools,” “networked platforms,” and “high-performance computing.” But we left the door open for individual authors to add other categories and to use free-form text in naming and describing infrastructure. We also sought primarily qualitative information and only some approximate quantitative data. Columns in the template ask for a description of the infrastructure that is the subject of each row, and they also inquire about attributes of that infrastructure by means of true/false check boxes or of simple quantification measures on a Likert scale. (True/false attributes include “control point for access to or use,” “open source, open access, or public resource/utility,” and “proprietary.” Attributes measured on a 1 to 10 Likert scale include “how heavily used?,” “reliability,” and “satisfaction for user.”) Other columns allow authors to register “ethical concerns” and add “comments.”

In regard to their format, our infrastructure manifests are presently implemented as spreadsheets (originally Google Sheets) that for publication were transformed into downloadable, self-contained HTML files (and, for the blank template, also a Microsoft Excel file). In principle, however, manifests could—and should—be implemented in alternative input and output formats. After all, large spreadsheets are difficult to navigate, view synoptically, and reproduce in different page or screen sizes. Ideally, infrastructure manifests in the future will appear in multiple forms supported by a variety of portability, presentational, and archival data formats—including minimalist ones such as comma-separated values (CSV) or JavaScript Object Notation (JSON) files of the sort commonly used to transport lightly structured data between systems and formats. In the future, too, the information in infrastructure manifests could be implemented in markup languages and ontologies for linked data.

Loosely flexible in their categories and vocabularies, and provisional in their format, these version 1.0 infrastructure manifests essentially represent just a preliminary scoping exercise. They probe the authors of chapters—including the editors themselves as the authors of the introduction and this appendix—about what they feel is within scope to declare as their principal infrastructures and how best to describe such infrastructures.

Why should there be infrastructure manifests for DH? In the spirit of Susan Leigh Star’s well-known dictum that infrastructure is normally “invisible” (382), our general goal is to make visible the infrastructure of knowledge production in a field where such infrastructure, as we argue in our introduction, is a constitutive “object of study.” More specifically, we envision five discrete purposes for infrastructure manifests in DH, only the first of which is intended to be adequately addressed by the present version 1.0 manifests. We hope that the other four purposes will be addressed by others in the DH community in future iterations of manifests incorporating lessons from our present experiment. The five goals of the infrastructure manifests that we set as an agenda for DH are as follows:

1. To make visible the approximate scope, variety, and proportions of the kinds of infrastructure involved in knowledge work
2. To quantify some kinds of infrastructure
3. To reveal typical platform dependencies (and other dependencies), areas of scarcity, and areas subject to single (or dominant) control points of provision or regulation
4. To make visible geopolitical and sociocultural differences in infrastructure
5. To assess infrastructure comparatively on criteria of ethical, sociopolitical, and economic values, costs, and harm

We hope that publishing our infrastructural manifests will prompt others in the DH community to accompany their works with similar manifests; that our version 1.0 template will be evolved or “forked” for other research communities or purposes; and that DH might thus eventually move toward a shared understanding of, or productive debates about, the most important categories, attributes, quantities, and values (intellectual, sociopolitical, economic, environmental, cultural, ethical, and otherwise) of infrastructure to declare. Supplemented by metadata, quantitative data, and markup, future versions of manifests might also be machine-readable to allow for computationally assisted aggregation, analysis, and visualization. If DH is known in part for “distant reading,” then the infrastructures that enable such methods should themselves be available for distant reading.

Being able to machine-process manifests would help overcome one limitation of the present infrastructure manifests that we acknowledge. Most of our manifests represent the infrastructure just of individual authors. In the case of chapters with plural coauthors, only some manifests (for chapters 5, 8, and 12) attempt to represent the collective infrastructure of those coauthors. This is because in practice, it can be difficult to combine authors’ infrastructures in a single manifest without resulting in a hard-to-understand miscellany. Such combinations would be more meaningful if we could aggregate hundreds or thousands of authors to find statistically meaningful infrastructural patterns.

It is also important to note that the norm for the content of our infrastructure manifests is professional. Manifests are intended to declare the kinds of infrastructure that the DH scholarly community (and related scholarly and industry research and development communities) thinks can and should be made transparent under citable open-science, accessibility, accountability, provenance, reproducibility, collaboration, and other principles or protocols. Examples of such evolving norms include the following:

• Collaborators’ Bill of Rights (Tanya E. Clement, Douglas Reside, Brian Croxall, et al.)
• Datasheets for Datasets (Gebru, Morgenstern, Vecchione, et al.)
• Data Provenance Standards (Data & Trust Alliance)
• Data Transparency Standard (IAB Tech Lab)
• FAIR—Findable, Accessible, Interoperable, Reusable (ALLEA)
• Fair Cite
• FAT—Fairness, Accountability, and Transparency in Machine Learning (FAT ML)
• Open Science Framework (Center for Open Science)

Excluded under the norms of the scholarly profession (and many other professions), therefore, is any expectation that special personal, medical, familial, or other private circumstances bearing on a researcher’s needs for infrastructure should be declared, although authors are free to include such information if they wish. Privacy norms also bear on what should, or by regulation can, be made public about scholars’ infrastructures and that of their collaborators or students. As Star notably observed, infrastructure is socially “relational” (380). Infrastructure manifests are not a one-size-fits-all lading list of goods. They are in part also social contracts that negotiate relationally between the goals of professionally shared knowledge and the needs, rights, and protections of individuals and organizations at different levels of advantage in the profession and society at large.