“Telling Fedora 4 Stories” is an initiative aimed at introducing project leaders and their ideas to one another while providing details about Fedora 4 implementations for the community and beyond. The following interview, conducted by Carol Minton Morris with Geoff Harder, Associate University Librarian for Digital Initiatives, Peter Binkley, Digital Initiatives Technology Librarian, and Leah Vanderjagt, Digital Repository Services Coordinator at the University of Alberta Libraries includes personal observations that may not represent the opinions and views of the University of Alberta Libraries or of the Fedora Project.
What are your roles with Fedora at your organization or institution?
Leah Vanderjagt, Digital Repository Services Coordinator, is the service manager for the University of Alberta’s ERA: Education and Research Archive, a Hydra-Fedora 4 based repository and the cornerstone of the Libraries’ Digital Asset Management strategy.
Peter Binkley, Digital Initiatives Technology Librarian, takes a guiding hand in architecting and leading UAL’s repository development and its use of Fedora.
Geoff Harder as Associate University Librarian leads the strategic development of UAL's digital initiatives, which includes research data management, digital repositories and digital scholarship support.
Tell me a little about your organization or institution.
University of Alberta Libraries is one of Canada’s major research libraries with a print and electronic collection exceeding 5.2 million titles and over 8.7 million volumes, serving an audience of 44,000 students and 4,000 staff. UofA has an interest in “activating our collections” – ensuring that the breadth and depth of our collections are findable, well used, and made available to the broadest of audiences for many years to come. To that end we have been developing a robust Digital Asset Management System (DAMS) that better enables the management and stewardship of the many thousands of digitized and born-digital materials entrusted to our care.
We are currently operating a Fedora 4 Hydra-based digital repository housing more than 40,000 items. The current phase of development will extend the repository infrastructure to manage digitized collections from various projects that, in the past, have had their own bespoke portals. We have very rich digital collections of Canadian prairie publications and ephemera (for example, the Peel’s Prairie Provinces and Sir Samuel Steele collections) and it is timely and, we believe, a very sustainable direction for us to bring these collections into a common digital library management structure. This is very much a work in progress, but we have made significant gains towards our DAMS strategy which we embarked upon in 2013.
Why did you decide on Fedora?
We began our IR service in 2006 with a DSpace repository which we continued to use for a few years to support ETD submission workflows. At the same time, we evaluated other available frameworks that would allow for easier maintenance, functionality and extensibility across our digital collections. The Islandora framework (Drupal plus Fedora) was helpful in making the decision to adopt Fedora though in the end, we chose to go with an in-house front end developed in JSP. The ability to tap into an active community, and support for scalability built into the Fedora framework were also factors in implementing Fedora as our institutional repository. Based on our observations of these frameworks, we saw Fedora as our likely best option to promote future flexibility in building services around our digital collections, and we were also very interested in how Fedora could support our aims to develop sound digital preservation practices.
What were your requirements going in?
As previously mentioned, we needed a system that would move us ahead quickly with good digital preservation readiness. We also have a strong preference for open source technology, so integrating Fedora into our stack was an appealing prospect. Primarily, what we knew is that the needs of our users and the management demands of our collections would change, so we saw Fedora as our most potentially flexible option.
What strategic organizational or institutional goals did Fedora help you meet?
We take our role as long-term stewards of digital resources seriously. We had a goal of developing a very feature-rich, flexible institutional repository, and building our own front end on top of Fedora’s strong foundation enabled us to do that quite quickly. We have been able to respond nimbly to user requests for changes and new features, which helped us promote the repository as a useful and valuable service. Because others in the Fedora community share our goals we do not have to negotiate a lot of changes to a big code base. Because our DAMS strategy sits around Fedora we are able to provide flexible support to our user community.
What are your plans for Fedora in the future?
Moving ahead, support for research data management and data archiving is a priority. Researchers need to have access to infrastructure that allows them to easily manage and archive their data, and we feel we can help support that need. We are excited about sharing our expertise as a preservation partner with other institutions. We are now officially committing one of our developers, Piyapong Charoenwattana, to manage the fcrepo4-oaiprovider project. The Fedora community offers opportunities for growing and learning with parallel projects, for example in developing best practices for deploying and backing up the Fedora datastore. We plan to take advantage of that community alignment of purpose.
We have a lot of eggs in different baskets. With Fedora we will preserve the diversity of content we have and add to our assets. In 2017 we plan to implement Fedora 4.8 including the PCDM (Portland Common Data Model) in the Hydra framework.
We are very interested in leveraging the use of Linked Open Data (LOD) out of our repository, working with some of Canada’s top research libraries who are part of the Canadian Linked Data Initiative. Currently we are participating in a new project underway to demonstrate the power of LOD to surface unexpected connections and to bring forth the uniqueness and richness of ETD scholarship in Canada.
We are also investigating closer integration between Fedora and our digital preservation platform (built on Openstack Swift), by means of a service to pull new or changed objects out of Fedora, compose them into well-defined Archival Information Packages, and pass them to our Swift cluster, which provides redundant, distributed, self-healing storage.
What is at the top of your Fedora “wish list”?
Performance. As we get deeper and deeper into RDF structures we are finding that it takes a long time to retrieve nested objects comprising large numbers of triples. Extracting the RDF graph representing a monograph of 1000 pages, for example, is very time consuming. We found the recent discussions at HydraConnect on Fedora performance in the Hydra stack to be very helpful.
As we’ve mentioned Linked Open Data services are also of interest. We are experimenting with synchronizing triples from Fedora into an external triple store for metadata evaluation and testing using SPARQL.
What advice would you give to other organizations that are planning a Fedora implementation?
Rely on the Fedora community–including Hydra and Islandora. There is a wealth of help available and a lot of great people willing to share their time and expertise. It can take a while to figure out how to navigate these communities – where to go to communicate and ask questions – but it is worth the effort to make connections.
Others may find, as we have, that active institutional commitment to development communities such as Fedora is a distinct advantage in recruiting excellent technical staff. The opportunity to participate in community-based open source projects can be a real draw.