Skip to article content

Preserving Interactive Research Content

Challenges, Frameworks, and Best Practices

Definitions

Here we list some of the terms and definitions in this context.

Preservation

Preservation, in this context, is defined as “the activity or process of keeping something valued alive, intact, or free from damage or Decay” (https://www.merriam-webster.com/dictionary/preservation). Preserving digital content is an active process as this content needs to be monitored, changed and migrated to ensure access remains stable. Content needs to be of enough interest, or value, to others for this process to be supported and funded. The outcomes of the preservation actions should minimize any damage and ensure that the object remains useful for the intended audience over time. The actions that might be undertaken for preservation will align with the concept of Graceful Degradation, mitigating the challenges of maintaining access while preserving interactivity when possible.

Graceful Degradation

The idea that complex systems can fail or simplify incrementally—not catastrophically Decay. It’s about designing systems that provide the richest experience possible when everything works (e.g. Interactive Research Content), while still offering meaningful, functional fallback options when conditions change.

Decay

Decay is not the same as Graceful Degradation, as GD implies decay management. Decay just happens, whereas GD is a response to decay (i.e. decay stewarding).

Interactive Research Content

An interactive research product is one that allows the research content, or the way that content is displayed, to change as a result of user input. Any change that results from user input, even a cosmetic change, counts as interactivity if it clarifies the work or deepens the user’s understanding of the work. However, in the context of the Preserving Interactive Research Content WG, we are mainly focused on user inputs that enhance memorability or understanding, and not merely readability (e.g. light/dark mode, font size). The research product must be designed with user input in mind, and make it easy for the user to customize their own experience.

In other words, an interactive research product must:

  • Enhance our understanding of a particular concept or dataset

  • Solicit and “respond” to user input, allowing the user to customize their experience

  • Be designed with user input in mind, making that input easy to provide

Is it Interactive Research Content (IRC)?

OutputIRC?Explanation
PDF (with links to downloadable code and data)NoResearch is displayed the same for all users Source code is available but not easy for users to change their experience
Downloadable binary executableNoResearch is displayed the same for all users Source code is available but not easy for users to change their experience
JATS/MetadataNoIncreases understanding, potentially drives interactivity (and accessibility) but is not interactive itself
Website with research resultsUsuallyWhen user can change display of data or ideas to enhance understanding
PDF research article with widgetsYesStatic product with interactive components
Application or dashboard with data built-inYesResearch display varies based on user input
Excel spreadsheetYes, typicallyUser responds immediately to changes in the raw data (or formulae) by updating any dependent content, such as other cells, or charts. Affordances available to end-users, not just the spreadsheet developer.
Multiverse analysesYes, ideallyUser considers multiple analysis pathways including selection of variables and statistical models. Enabled by IRC.
Notebook with research code and results - “Explorable Explanations”SometimesA notebook with widgets, zoomable images, or empty input fields is IRC. If designed to be run on the command line, or perhaps published as a static PDF or HTML, is not IRC because it is not designed to solicit and react to user input. A program that just happens to be easier to edit simply by virtue of being in a notebook format is not IRC.
Package or librarySometimesIf the package is designed to enhance user understanding of a particular research product by responding to user input, then it is IRC. A package that “merely” serves as an interface for research data, and is not designed to enhance our understanding of any particular study or studies, is not IRC. This is a research tool, and while research tools are extremely useful research products, they are not research content per se.
Dockerized programSometimesA program designed to solicit and respond to user input for the purpose of enhancing user understanding is IRC. A program that was designed to run in batched or non-interactive mode does not become interactive simply because it was containerized.
Preserving Interactive Research Content
Expanding the Idea: Perspectives from the Ecosystem