Alastair Dunning (TU Delft Library / 4TU.Centre for Research Data)
Marta Teperek (TU Delft Library)

1 November 2017

On 26 October 2017, the European Open Science Cloud published the EOSC Declaration, which is available to all scientific stakeholders, for their endorsement and commitments to the realisation of the EOSC by 2020. Below are our thoughts and comments on the EOSC declaration and the EOSC Strategic Implementation Plan 2017-2020.

European vs Global?

Research nowadays is more and more cross-border; in-line the idea of transparency, sharing and interoperability championed by EOSC. What is unclear in the current EOSC proposal is the relationship between “European” vs “global” accessibility of EOSC. The EOSC Strategic Implementation Plan 2017-2020 suggests that EOSC will provide access to all European research data and that free access to the platform will be offered to all European researchers. The very last point of the EOSC declaration states that EOSC will be “open to the world” and “reaching out over time to relevant global research partners”.

There are already too many tools and resources burdened with institutional and national constraints, which limit collaboration and exchange. Therefore, given that research happens internationally, it is important that the openness to the world and coordination with global research partners is planned as one of the top strategic priorities of EOSC, stated from the outset in the Declaration.

Efforts required for development of standards, community consultation and service integration should not be underestimated

The EOSC Strategic Implementation Plan 2017-2020 notes that in the Federated model for EOSC development, common service standards will need to be established for all federated resources. It is also stated that the costs of Federated model for EOSC were only “marginally higher than baseline”.

However, functional interoperability will depend on agreeing on common interoperability standards. The EOSC Declaration stresses on several occasions the importance of a disciplinary approach to FAIR principles and standards development. Inevitably, this will lead to tensions between granular, community-driven approach to standards (relevant for making research outputs FAIR), and the need for deciding on minimal requirements which could make the whole service interoperable.

Overall, the EOSC Declaration tend to overstate the simplicity of technical implementation and underplays the necessary community development and engagement efforts. The document rightfully states that “research data must be both syntactically and semantically understandable, allowing meaningful data exchange and reuse among scientific disciplines and countries”. However, getting consensus across the various communities and across all scientific stakeholders will require considerable work (which means both time and financial investment). In particular, local subject-specific communities will need to be engaged, consulted and they will most likely require dedicated support to achieve interoperability and integration with EOSC.

This should also be taken into account in financial planning for EOSC development. The EOSC Strategic Implementation Plan 2017-2020 states that no/little fresh money is needed in implementation until 2020. However, for the project to be successful, it is important that work on defining common standards and assisting the various scientific stakeholders in integration efforts starts as early as possible, and this will require both time and financial investment.

In addition, it might be desirable to think about the inclusion of disciplinary stakeholders within the different EOSC Working Groups to ensure that their views are taken into account from the outset.

Research outputs other than datasets need to be recognised

The current EOSC Declaration, as well as the EOSC Strategic Implementation Plan 2017-2020, are mainly focused on research data as an output. However, software, supporting methods and protocols are equally important for effective reuse of research data. In fact, most research projects have now a computational element, therefore a more holistic approach to all research outputs other than traditional publications is needed.

In the current EOSC Declaration the idea that EOSC should support the whole research lifecycle and that “software sustainability should be treated on an equal footing as data stewardship” is only mentioned in the context of Service development. It is key that the Declaration emphasizes all research outputs from the outset to ensure that other objects crucial for research reusability and interoperability, such as code and protocols, do not become second-class category objects.

The need to reward open practices

We welcome the Commission’s view that researchers’ commitments to open practices need to be appropriately rewarded: both at the recruitment stage and during career progression. However, and as already expressed in our comments on the EC’s report on “Evaluation of Research Careers fully acknowledging Open Science Practices”, we believe that examples of immediate adoption are needed.

Researchers who are currently trying to implement Open Science practices need a new reward system now. A pilot grant scheme with the requirement to demonstrate a commitment to Open Science practices as one of the eligibility criteria would be a good starting point. Not only would such approach provide immediate recognition and reward for researchers already practising Open Science, but would also contribute necessary information and feedback for possible further implementation of Open Science practices in future funding schemes (such as FP9).

In addition, there are currently no actions committed for the “[Rewards and incentives]” capacity in the Action list and it would be helpful if some declarations and commitments in this area were solicited.

Data management plan requirements

We welcome the suggestion that data management plans (DMPs) should become an integral part of every research project and we welcome the wish to align funders’ and institutional requirements for DMPs in the EOSC Declaration. However, the EOSC Declaration also mentions that “researchers’ host institutions have a responsibility to oversee and complete the DMPs and hand them over to data repositories”.

We believe that making institutions responsible for overseeing and completion of DMPs removes the responsibility for good data stewardship from researchers. In addition, it poses a risk that DMPs will be perceived by the research community as yet another administrative burden. Making researchers responsible for data management plans provides a good opportunity for them to embrace the idea of good data management practice as an integral part of their research. Institutions could provide advice, expertise and training for researchers, but they should not be the ones responsible for overseeing and completing the DMPs.

In addition, we are unsure about the statement that institutions should be handing over DMPs to data repositories. Which data repositories are meant? Multiple data repositories? Would it not create an unnecessary administrative burden? Perhaps instead there could be a central, European level registry of data management plans, which would also facilitate the alignment of information collected in data management plans and allow more efficient reuse of information collected in data management plans (for example, about the needs for specific IT infrastructure, training support etc.).