How to make a CRIS OpenAIRE-compatible: the case of METIS: METIS2OpenAIRE case study

Blogpost by Pablo de Castro, METIS2OpenAIRE project coordinator at euroCRIS
The OpenAIRE-funded METIS2OpenAIRE project has run from early Feb to mid-May 2018. Its aim has been to make a first institutional CRIS system OpenAIRE-compatible as per the updated CERIF-XML Guidelines for CRIS Managers. The process has involved the development of a minimally sufficient validator that future CRIS systems may use as a means to confirm their own compatibility with these Guidelines very much in the way the process works for Open Access repositories and journals.

  Issues around OpenAIRE compatibility via CRISs: non-compliant institutional & national-level systemsAs of mid-May 2018 no CRIS system is listed yet on the very long list of OpenAIRE -compatible data providers. This can impact the comprehensiveness of the OpenAIRE content aggregation in areas where the Open Access repository layer is not consolidated enough or is not OpenAIRE-compliant. Such is the case for instance for the United Kingdom, a small fraction of whose institutional repositories are able to deliver their content into the EU-wide aggregation, or for countries where a strong national-level CRIS system could make available significantly higher levels of research information. Taking the first steps towards enabling this additional degree of system interoperability is seen as a key development in the future of the Open Science movement. In this sense, although the METIS2OpenAIRE project is officially over as an OpenAIRE-funded project, its effort towards enhancing system interoperability towards OpenAIRE is an ongoing one.

The handshakeThere are two sides where hands need to be stretched out for a handshake to happen in the realm of system interoperability: the CRIS side and the OpenAIRE aggregation one. This project is just dealing with the first of these, i.e. making sure that the data feed exposed by a CRIS system via an OAI-PMH endpoint meets the requirements expressed in the OpenAIRE Guidelines. An additional workline needs to be in place from the aggregation to enable the ingestion of such compatible data stream. This has only been broadly discussed thus far, but will be addressed in more detail as the feed is ready for its inclusion in the content aggregation.

Tackling the CRIS handshake from the CRIS side: CRIS feed validation against the updated CERIF-XML Guidelines for CRIS managersA process for OpenAIRE-compatible metadata exposure via an OAI-PMH endpoint has been tested and completed for the METIS institutional CRIS system at Radboud University Nijmegen in the Netherlands. This metadata provides a much more comprehensive degree of contextual information for research outputs than the average repository, including for instance information on funded projects or affiliations among others. The development of the above-mentioned minimally sufficient validator that other systems will be able to re-use is one of the key outcomes of the METIS2OpenAIRE project. This validator is simultaneously being tested by other systems taking part in the project as budget-neutral stakeholders such as the Polish Omega-PSIR. The testing of the CRIS metadata feed against the CERIF-XML Guidelines takes place in a loop where adjustments may be made to the latter in order to accommodate the specific needs of a given system. These Guidelines will at some point be ‘frozen’ before a new update takes place again once a sufficiently large number of required updates has been identified, and it’s useful to have them tested by as many different systems in different countries as possible.

The work on METISThe work on the main institutional CRIS system the project has focused on has involved the setting up of the OAI-PMH endpoint and the multiple loops of testing and data model adjustment resulting from the compatibility rules to be met. Because of the previous participation of METIS in the P-O-PF projectfor information exchange across CRIS systems in the Netherlands, the first entities to be exposed via the endpoint were ‘Persons’, ‘Orgunits’ and ‘Projects’. These were later followed by ‘Publications’, ‘Products’ and ‘Patents’. The technical team at METIS have reported a relevant degree of data model fine-tuning as a result of the project that will be helpful for other institutional systems. These in-house-built CRIS systems are seen as the ideal first candidates to test OpenAIRE compatibility as they tend to have full control over their development needs in terms of data model evolution.

The budget-neutral project stakeholdersThis said, in-house-built institutional CRISs tend to be one of a kind and may not be the best choice to ensure a widespread adoption of the CERIF-XML Guidelines for CRIS Managers. This is the main reason why the METIS2OpenAIRE project has teamed up with two external budget-neutral vendors to ensure a wider implementation of its solutions: these are Omega-PSIR in Poland and Elsevier PURE in Denmark. The latter stakeholder brings in the possibility for OpenAIRE compatibility to be adopted across the very wide range of implementations of this system worldwide and specifically in areas like the UK, thus helping to tackle the current shortcomings in the domain of system interoperability with OpenAIRE. Although these external project partners have committed to the implementation of the OpenAIRE Guidelines on their platforms, they are not subject to any requirements in terms of the official METIS2OpenAIRE project time-window, meaning that they were not expected to be OpenAIRE-compatible by mid-May. In fact, Pure is aiming to become compatible via their Oct 2018 release, while Omega-PSIR intends to present the first examples of information exchange with OpenAIRE at the euroCRIS membership meeting to be held in Warsaw next November. Both Pureand Omega-PSIRhave in the meantime submitted contributions to the CRIS2018 conference to be held June 13-16 in Umeå, Sweden.

Communication: teaming up with 4Science and the CRIS2018 conference in UmeåThe communication of the METIS2OpenAIRE project progress was seen as a key area of activity since there awareness of the gradually emerging interoperability opportunity needs to raised across the community of Open Science practitioners. This has involved a specific project webpagethat will keep getting updated beyond the official mid-May project end-date, plus dissemination of the project via social media and presentations. On top of this, the project WP3 included a proposal for a joint webinar on making CRIS systems OpenAIRE-compatible to be delivered with 4Science, who have been running their own parallel OpenAIRE-funded projectfocused on the DSpace-CRIS platform. Bearing in mind that all partners involved in both parallel projects will attend the CRIS2018 conference in Umeå in June 2018, the opportunity that the event will provide to advance the discussions on this area of activity and on the organisation of the webinar will be very significant. While there is still no specific date for this dissemination activity, it remains very much a part of the communication plan, and might benefit from a later scheduling at a point where there’s solid and sufficiently widespread progress to report.