Dispatch from the Digital Preservation Coalition’s Unconference 2023
I was fortunate to attend the Digital Preservation Coalition’s Unconference in London last week. Held annually, the Unconference is a member-led event and an opportunity for operational and experienced practitioners within the digital preservation community, who are also based at DPC member institutions, to seek or contribute feedback to relevant topics within the discipline.
I left feeling there was much more to discuss and wanted to use this post to briefly summarise key points from some of the sessions I attended.
Somaya Langley from the Science Museum Group led two sessions: one on digital preservation storage and another on cybersecurity.
Along with data validation and security that come with different types of storage for preservation, the following points also guided the first session:
- The 3–2–1 backup rule (3 copies, 2 different media, 1 offsite copy)
- Legislation and mandates around data localisation and residency
There was a range of preservation storage solutions used by institutions represented in the audience, from cloud to tape — and even tape that’s stored inside a mountain! From listening to points and questions from audience members, it sounded like most institutions are taking a hybrid approach to storing digital collection materials.
Takeaways for deciding whether to use any kind of storage is staff familiarity with the solution as well as time and resource to spend on it. At CUL, we’re taking a cloud native approach to building services and storing data and have gradually learned about cloud suppliers and their services along the way. Some members of the Digital Preservation Programme team have become certified in cloud solutions, and we also learned about using the cloud for digital preservation from staff at Wellcome Collection.
At the cybersecurity session, a few audience members expressed how challenging it can be to build inroads with colleagues in IT. This could be for several reasons, including IT staff familiarity with digital preservation work and the need to download software from the web to test it out; communication only happening via a helpdesk, making it hard to form professional relationships; and general concern that colleagues outside IT might compromise cybersecurity protocols.
Cyber Essentials was mentioned by a few audience members, including how its requirements could work against staff working in digital preservation at an institution that is certified with this standard.
During both sessions, trust and risk were recurring points from audience members, with the question of how might digital preservation staff work with IT colleagues to mitigate any risks associated with their work.
Various pain points were raised during the session on appraisal, which was led by Ellie O’Leary from the DPC.
Whether staff are familiar and skilled with appraising digital materials depends on the institution. The sheer volume of material to appraise was mentioned as one challenge.
The complexity of any volume of material could depend on a number of factors, including:
- Digital materials not being created with preservation and access requirements in mind.
- Lack of records management practices that could better prepare records for later appraisal if they need to be preserved long term.
- Identifying sensitive data, which could be a time-intensive process for which no automated tools exist.
- Hybrid acquisitions where materials are appraised by different teams according to type and format.
- Getting the correct and expected digital materials from the donor or depositor.
- Extracting digital materials from handheld storage carriers so they can be appraised.
At CUL, a Transfer Service was set up to transfer data on handheld storage carriers. Transferring materials from optical media has presented a range of challenges that the Digital Preservation team are currently working through, with the aim of sharing more about these at a future point.
Bryony Hooper from the University of Sheffield Library led a session on Key Performance Indicators and measuring digital preservation impact. If you subscribe to the DPC-Discussion listserv, you might have caught Bryony’s email on this subject, where she asked for feedback on how other institutions addressed this topic.
One takeaway was deciding where quantitative metrics or qualitative metrics are needed when communicating the impact of digital preservation work to stakeholders. Quantitative metrics are great for creating a baseline and measuring improvement over time, whilst qualitative metrics can help demonstrate the success of business change and the level of user satisfaction achieved by a digital preservation project or programme.
Francesca Mackenzie at The National Archives made a good point about how use of a service or system doesn’t necessarily indicate it’s a success. Both quantitative and qualitative paint a better overall picture of success.
From the discussion, it felt like the use of KPIs, as well as defining any KPIs, comes down to the audience for the data. It could be the case, for example, that there are some KPIs that apply more to the work of senior management and different KPIs for other stakeholders of the service.
Thomas Edvardsen and Johannes Karlsen asked for feedback on their thoughts around a packaging standard for digital materials to be preserved at the National Library of Norway.
One of their sessions focused on different specifications for structuring content in preservation packages and the advantages to using current standards, including DILCIS (E-Ark) Specification, and Cinema Preservation Package (CPP), BagIt, and OCFL.
Like with any standard, all the ones mentioned above have their pros and cons, and the decision to use a standard should be guided by the requirements of the institution. Such requirements might not be limited to preservation and access, but also include other areas of collection management, like search and discovery.
Their second session focused on whether metadata for digital materials should be stored with or separately to the content: “Can a separate metadata store serve as the foundation for generating digital object metadata files (METS/PREMIS) that are stored together with the content?”
Similar to the question about how to package content and metadata, it feels like this question also comes down to requirements around use of the data being preserved, as well as exit strategy from the systems and storage being used within the wider infrastructure.
