The following is a post by Olivia Dorsey, an LC Labs team Innovation Specialist working on the Computing Cultural Heritage in the Cloud (CCHC) initiative at the Library.
Hello! Things have really started to get underway with the Computing Cultural Heritage in the Cloud (CCHC) initiative, like rounding out our team and the arrival of our researchers, and we wanted to share a sneak peek at what we’ve been up to!
The CCHC initiative is supported by a grant from the Andrew W. Mellon Foundation. Our aim is to understand what would be necessary for the Library to support research and access to collections at scale via a cloud computing environment. Over the next two years, we’ll be exploring the services, policies, and technical infrastructure that would support large-scale research with the Library’s digital images, audiovisual material, web archives, textual data, and metadata. This post describes some of the foundational thinking behind our processes.
Developing Our Values
As we began to build out our team, one of the first things we did was develop project values that encapsulate how we intend to undertake our work and what we intend to achieve with CCHC. We participated in a project charter workshop where we each came up with values that we felt were important to the work of CCHC. Then we “up-voted” the values that we agreed on, discussed, and finalized them.
In the end, we came up with four core values:
Members of the CCHC Initiative team are committed to the following values:
- Inviting: We encourage others to join us on this journey of discovery and welcome thoughts, suggestions, and questions as we take on this complex work. We integrate experience and expertise from colleagues, peers, and external collaborators. We want to learn together with you!
- Transparent: We will share our processes, successes, and failures for the benefit of the Library, its users, and the cultural heritage community at-large. Our hope is that our work will lay the foundation for user-centered cloud computing solutions in other GLAM (galleries, libraries, archives, and museums) institutions in the near-future.
- Exploratory: We lead through experimentation and share what we learn with others. We recognize that we do not have all the answers and will rely upon each other, our colleagues, and the wider cultural heritage community for their collective expertise. We will build upon and extend this expertise for the benefit of the team, initiative, Library, and its users.
- Flexible: We stay nimble in the face of uncertainty and are ready to adapt to changes. We welcome new ideas that challenge our thinking and processes. We create space for iteration and reflection in our ongoing work.
So, great! We have some values. But what does that mean and what does that look like in practice?
Using Values to Design and Guide Our Work
A big part of our work with CCHC involves building understanding and collaboration with our colleagues around the Library. Their expertise informs the ways we address barriers we encounter and better achieve the goals we have in serving researchers digital content. These discussions also allow us to identify opportunities for even more shared problem-solving. We are driven by opportunities that enable the Library and other GLAMs to consider how workflows can be used to create new paths of access for users of these vast resources. And we think that we can use insights from CCHC to help colleagues achieve their goals, if we work to understand their perspectives and challenges.
We keep our meetings open to LOC staff and interns, as a way for them to take an inside look at our process. Several tools are helping us to make this work more visible inside and outside of the Library, such as project management and documentation software, internal report-outs, and future presentations and blog posts!
We are also establishing working groups within the Library to talk through cloud architecture and collection structure considerations. We meet on a monthly basis to chat through researcher needs, collection and technological limitations, and approaches for supporting computational researchers in the future.
Unexpected obstacles and even barriers are inevitable, but they are a reminder to stay flexible! Embracing this value allows us to analyze multiple options for gaining access to collections’ data. We are also documenting workflows and obstacles so that we can remove them in future practice. The purpose of this and other pieces of documentation are to record what we learn over the course of the initiative and share these findings in ways that make the most sense for our audiences.
Doing the Work
We pull from a variety of methodologies for our work, but largely follow an Agile approach. As we figure out the tasks to be done, we create a list, or “backlog” of them. Every two weeks, we start a new iteration (or sprint) of tasks that revolve around specific goals. These goals are based on some of the tasks from the backlog and our priority level for those tasks. For example, in one recent iteration, some of our goals included:
- Explore cloud infrastructure needs more deeply
- Create diagrams and user stories
- Refine understanding and initiate support for researchers
Some associated tasks for these goals may include “Developing workflow diagrams for processes we have encountered” or scheduling meetings with relevant stakeholders.
We organize our tasks into four main statuses on a sprint board. Those statuses include “To-Do,” “In-Progress,” “On Ice/On Hold,” and “Done.” As we work on each task, we make sure to update its status and note any resources or other details on that task. If a task needs to be moved to another person, we re-assign the task to that person.
A day before the iteration ends, we have a refinement meeting where we review the statuses of our current tasks and identify what tasks we need to include in the next iteration. At this point, we assign each other to the next iteration’s tasks. Recently, we began implementing mid-iteration checks to identify blockers halfway through the iteration. By doing this, we can, hopefully, remove any obstacles before the next iteration!
Finally, we have an iteration begin/close meeting where we demonstrate an outcome from the previous two weeks and undertake a retrospective on the previous iteration. We then pinpoint areas of accomplishment, blockers, and overall feelings about the flow of work to identify what we could address and improve in the coming weeks. Then we close the current iteration and start the next one.
Additionally, an important part of this work is sharing a ground-level perspective to cloud computing and machine learning, especially for those who are new to these approaches. We intend to keep the “human” in Digital Humanities by communicating how these methods actually work, identifying how to get started with them, and reasons for using certain tools. We regularly ask what audiences might benefit from learning about the work and the best ways to share with them – like you, here on the Signal blog!
In the coming months, we will be working with three expert researchers whose projects use the Library’s collections data in a cloud computing environment. They will be using machine learning, AI, and other methods to analyze data from our collections. Toward the end of the researchers’ time, we will provide a GitHub repository to hold their code, their output, and the derivative datasets that they used from our collections. We hope that his will help others understand how we have made decisions and designed and executed the work, while opening a world of possibilities for future projects.
These are exciting times and we can’t wait to share more very soon!