Everyone seems to be on the scent of good research, researchers looking for collaborators, research managers developing industrial partnerships, and funders wanting to demonstrate the value of the investments made.
The key to finding research is good metadata: information about the books, articles and so on that researchers produce. Developed by UKOLN at the UK’s University of Bath, with input from Chygrove and funding from Jisc, a new Metadata Application Profile and Guidance for repositories called RIOXX will be launching in April 2013 and will make this process easier and more reliable.
Over the years various metadata schemas and models have emerged, but clarity on the best metadata strategy to adopt or how to achieve interoperability between scholarly systems has been a hard nut to crack. Rather than trying to harmonise the whole of the scholarly metadata landscape, Jisc decided to focus on a particular use case – the ability to track research outputs across systems. We then started to ask what would be the essential, minimum and common fields required to allow the tracking of research outputs.
After discussions with research funders, institutional repository managers and research managers, it was clear that being able to efficiently and consistently track research outputs was a common problem. Time and effort could be saved if this could be resolved. Funders want to track outputs associated with their grants, universities want to track outputs associated with their researchers, researchers want to have their admin burden reduced, and track their own outputs. Publishers, data archives, repositories, funders and others want to provide services based on good quality metadata.
In surveying the landscape, we found that metadata implementation has been mixed. Some repositories have customised their metadata fields according to their requirements but not necessarily followed known metadata schemas to do so. Some may have adopted known schemas and others have simply opted to accept the default ‘out of the box’ metadata from the software solution. The use of schemas and also vocabularies associated with particular fields (restricted set of keywords/classifications) has been patchy at best.
So, given the patchy adoption of previous approaches, the RIOXX Metadata Application Profile and associated software improvements really needed to be pragmatic and implementable across the board. The first step in the process was to focus on applying consistency on two key sets of metadata fields, namely project ID (grant number) and funder name. Reliably linking funding information with research outputs will benefit anyone needing to track research across scholarly systems.
An increasing number of universities are installing current research information systems (CRIS), which provide a richer (and inevitably more complex) picture of how research funding, projects, people and outputs are related to each other. CRIS often use the Common European Research Information Format (CERIF). There are a number of UK projects that are based on its adoption and implementation. For example the CERIF in Action (CiA) project is developing plug-ins for CRISs to allow import and export of CERIF compliant metadata. An example of a European CERIF implementation is the Current Research Information System In Norway (CRIStin) initiative.
With the emergence of CRISs there has been a need for effective data exchange between systems. How institutions have set up their repositories and CRISs varies. The majority have set their repository to co-exist with their CRIS. However, others have set up the repository as a backend system to hold their documents with the metadata being held via the CRIS. The repository and CRIS configurations may change over time, but the important aspect to focus on is the integration and interoperability between systems however they are set up. The European OpenAIRE repository infrastructure initiative recently announced that it will support the use of CERIF-XML for interoperability.
Given an expected slow migration to CERIF as a basis for interoperability, the RIOXX guidelines are seen as an interim, yet vital, measure. Nevertheless, interoperability work is expected to be ongoing beyond these interim guidelines to ensure that repositories can meet user needs, for example by linking to publisher versions of papers via their Digital Object Identifier (DOI).
Metadata schemas are all very well but, without a common understanding of the terms used in them, the information is not truly shared. One answer is to develop common vocabularies, and the UK Open Access Implementation Group (OAIG) is working with the Jisc Publishers and Libraries Metadata and Interoperability Group to agree some of these. The Vocabularies for Open Access (V4OA) project will work with a wide variety of stakeholders such as National Information Standards Organisation (NISO), Open Access Scholarly Publishers Association (OASPA), Society of College and National University Libraries (SCONUL) and United Kingdom Council of Research Repositories (UKCoRR) to try and achieve a consensus on key vocabularies related to open access (OA. The intention is that they will be incorporated into the RIOXX application profile and the guidelines will be updated accordingly. They will also be relevant to the implementation of CERIF.
RIOXX and V4OA are timely interventions as they support the Research Councils’ revised policy on OA, which comes into force from April 2013. The policy mandates that the funder and grant number of research papers be declared when these papers are made available, which will make it easier for everyone to monitor compliance with the policy, and will help repositories report to the Councils’ Research Outcomes Service (ROS).
Better research information management and resource discovery require robust accurate metadata. It is not glamorous, but it’s essential for improving interoperability between local, national and international services. For example, it remains bizarre that in the UK we are unable to say accurately how many research outputs were produced by universities, and who funded them. Furthermore, there is increasing demand on academic services within universities to be efficient and provide added value but, as the retail sector knows, business intelligence is only as good as the information you have. The initiatives above are starting to support universities in their need to provide services that meet growing institutional requirements for better metrics, and more efficient research and reporting, all of which should serve, rather than drive, research, teaching, and business and community engagement.
Reproduced courtesy of Research Information.