CAUL Fair, affordable and open access to knowledge program: CAUL Review of Australian Repository Infrastructure

"The purpose of the CAUL project is to review the current state of the Australian institutional research repository infrastructure. The objectives were to make observations and recommendations on possible ways forward for the Australian repository infrastructure and management....

Through the completion of work packages, it was found that: • Australian repository infrastructure is diverse with numerous mature and new generation, local and cloud hosted, open source and proprietary solutions in use. • While all institutions endeavour to make published research outputs openly available, only half make research data available. • Non-research related collections – archival library collections, images and multimedia, and course materials are also in scope for a minority of institutions. • Less than half of Australia’s universities have an Open Access Policy or statement for research outputs. • Two-thirds of institutional repositories support grant funder policies, only one-third monitor compliance, and one-quarter use Research Activity Identifiers. • Only one-third of institutions have a preservation strategy for their repository collection. • While [lack of] harvesting by Google and Google Scholar was initially thought to be an issue, this is not the case with 90% of institutions reporting satisfactory harvesting. Further inspection demonstrated that the remainder of institutional repository collections are actually being harvested either directly, or via Trove which is harvested by Google, but not Google Scholar. • Interoperability is of primary concern to the current and next generation repository systems. • The Australasian repository landscape would benefit from greater connection with international repository developments such as via Coalition of Open Access repositories (COAR) membership. • The Australasian repository landscape would benefit from increased coordination and support via a technical advisory group which could lead initiatives on training, interoperability, metadata, standards and system requirements. This group would build upon the excellent work undertaken by the CAIRSS group / CAUL Repositories Community Group. • User stories based around Findable, Accessible, Interoperable and Reusable can be defined, which help to direct resources and attention for repository development. • Existing Australian repository infrastructure can be improved via the work of a technical advisory group, however the diversity of systems and investment in the landscape, means that this is complex and that benefits would not be evenly spread across institutions. Many institutions report a reluctance to further invest in their legacy repository system, having a preference for a next generation system. • A series of ideal state statements could be identified to provide direction to the technical advisory group and future project work. Again, the diversity of institutional policies and infrastructure makes it difficult to apply ideal state statements across the landscape. • A range of attractive, current and next generation repository tools are available, these being open source, proprietary or commercially hosted and supported open source. This suggests an opt-in consortium procurement and implementation process may be a suitable avenue for updating institutional repository systems to bring them uniformly up to minimum system requirements and standards. It was found only a minority of institutions, perhaps a hand full, would partake in a shared infrastructure procurement process, questioning the financial viability and sustainability of such a project. This was confirmed by the consultation process. Shared infrastructure would likely be on a cloud provider platform. • The success of international repository networks and collections may point the way for the development of same in Australasia. While a “Research Australia” collection was initially thought to be of interest to the sector, this was not found to be the case, with little interest in and support for developing such a collection, especially considering likely project costs to the sector. This was again confirmed by the consultation process. The NLA Trove system is the closest thing Australia has to such a collection and infrastructure. Working with the NLA to separate the national research collection from the library collections and developing a Research tile in Trove, and ensuring indexing via Google Scholar, offers the least work-intensive and most cost-effective solution to achieving this, while also being independent of any shared repository infrastructure project. • The Project Initiation Document (PID) scope intentionally excluded the investigation of Current Research Information Systems (CRIS) / Research Management System(s) and their interoperability and integration with research repositories. Feedback gained from the consultation processes indicates some community members regards this to weaken the value of this report...."