Preface: I want to acknowledge that the solutions discussed here are very reliant on the monetary and staffing resources that UM Library is lucky enough to have. While these specific solutions cannot be followed by everyone, I hope it is still helpful for us to share our lessons learned.
Currently frozen: Back in January, I posted a story about recent challenges in the ongoing management of our dark repository, Dark Blue. These challenges, mostly caused by well intended yet untenable over-commitments by myself and my colleagues, led to the pause of further ingest into the repository, development of new workflows, and freezing of programs such as audio/moving image digitization and digital archaeology work. Well, despite it being summer here in Michigan, things are still quite cold. No new content has been ingested and existing workflows are non-functioning.
As I am sure you can imagine, sustaining a pause like this has been challenging to say the least. None of us want to be in the position of stopping preservation and access services. Luckily, the majority of those working on the affected programs understand the risks of continuing to do things incorrectly, and support maintaining the pause until we have workable ingest workflows based on good preservation practice. My endless appreciation for their support.
Getting unfrozen: In that previous post, I mentioned that we are moving forward with implementing some new solutions that we think might solve our problems and allow Dark Blue to start taking content again. We recently signed a contract with Artefactual to provide support for a local installation of the Archivematica tool. Archivematica provides ingest pipelines that can be tailored to different types of content while providing fundamental preservation functionality such as fixity checking and AIP packaging.
While we are still in the final testing phase of new workflow development and have yet to fully complete kicking Archivematica’s tires, we are hopeful this will be a good fix for our major issue of packaging and validation. Before, we were relying way too heavily on our colleagues to develop and maintain custom code for different content types. Implementing Archivematica shifts the work of customizing pipelines and developing rules to the Digital Preservation Unit, which also allows us to directly oversee the preservation aspects of this important piece of the life-cycle. Our Library IT division is supporting the contract with Artefactual, especially by maintaining the VM and storage. This reshuffles work in a way that is just a better fit for both the Digital Preservation Unit and LIT.
People, people, people: The first few steps in this whole adventure were to be honest with ourselves, recognize that we had made some mistakes, and that we desperately needed a reset. The next was to acknowledge that a big part of the failure was an improper balance of the work associated with delivering a preservation-focused dark repository service. Even after we had done all of that, the solution was not found in a specific piece of software or the funding of a contract, but in adding staff capacity. Rebuilding each workflow from scratch and learning and testing (and testing, and testing…) a new tool has taken a tremendous amount of people power and people hours. We would still be at a capacity impasse if not for the addition of a staff member to the Digital Preservation Unit last year, which allowed our unit to take on the bulk of the workload to launch and maintain this new service approach to Dark Blue.
For me, this just reinforces the fact that digital preservation is a human endeavor. Overcommitting finite person hours got us into this situation, and the only way out was to add skilled hours in a manner that made sense. Had we not been in a position to do so, this post would be about the difficult decision to shut down several important preservation and collecting services. Proper staffing of digital preservation work remains one of our library's, and I believe our community's, primary challenges.
More to come on our workflow development, implementation of Archivematica, and more punny uses of the word “dark”.