OpenAIRE feedback to the EOSC Strategic Research and Innovation Agenda
31 August 2020
Executive Summary: OpenAIRE would like to see changes in the following areas of the SRIA: recognition and support for Open Access to publications as an active and guiding principle of EOSC; a recognition of the interoperability framework as a key pillar in EOSC; a more comprehensive recognition of the legal and ethical aspects around access to research data and finally the need for EOSC to be an enabler and coordinator of community-driven activities around Open Science, rather than a top-down framework.
OpenAIRE welcomes the SRIA as a landmark strategy for the future of EOSC. This document will be a reference point by member states and pan-European initiatives for developing the EOSC framework and as such we would like to take the opportunity to put forward some observations, in particular, the following areas where we believe gaps exist to fulfil the objectives of EOSC.
1. Beyond FAIR
Open Science in EOSC should encapsulate FAIR data as one element of a wider ecosystem of research elements. An overarching observation from our community is a shift in emphasis is needed to recognise that Open Science should be about knowledge sharing at a wider level, including FAIR data - an extremely valid part of doing Open Science -, software, etc., but also about research integrity, transparency and reproducibility, thus including other elements involved in the ‘knowledge ecosystem’, such as tools and services. In particular, in contradiction with previous Programs and EC supported Open Access initiatives (e.g. PlanS), the SRIA omits publications as valid digital research objects that also need to be FAIR and open. Open Access to publications is the main building block of Open Science. They are still at the heart of the research process and immediate free access is indispensable to researchers’ needs and this should be at the very centre of EOSC’s strategy. The H2020 mandate for Open Access to publications is still in place, aiming for 100% Open Access to all publications in the coming years and this should be reflected in the EOSC agenda to support member states to realise and implement this.
The EOSC ecosystem should include all types of research resource types:
- Research products (digital results of science): experimental data, workflows services, publications, software
- Research services (e-infras, research infrastructures, clusters): services used to manage and process data, using software and other products, possibly in combination with other services (scientific workflows)
Recommendation 1: Clarify the EOSC is about Open Science, in turn, concerned with openness and reproducibility of science. This entails advocating for open access, open data, open-source, etc. as well as establishing common practices (e.g. workflows, standards) and infrastructure (e.g. services, tools) to perform science in such a way the scientific process is reproducible (repeatable, replicable, reusable).
Recommendation: Extend the milestone of objective 2 (Standards, tools and services allow researchers to find, access, reuse and combine results):
- “Making EOSC resources (such as data, software, publications, tools, and workflows) as FAIR as possible across all European research infrastructures will be key for sharing/reusing research results, [add] and for using federated infrastructure for combining and exploiting open data”
2. Interoperability framework
The scope of the Interoperability Framework in the SRIA is (i) focused on a high-level modelling of interoperability by referring to the European Interoperability Framework and (ii) limiting the challenges to content (aka “data”) via FDOs. As such, it does not sufficiently cover all interoperability challenges to be faced to fully implement Open Science. “EOSC Interoperability frameworks” include metadata standards, service APIs, workflow frameworks, legal frameworks, can be discipline specific, cross-disciplines, or discipline-agnostic, such as the frameworks to be adopted in order to comply with the EOSC Core RoPs (e.g. EOSC-Hub/ENHANCE’s Service Description Template). As there is no one-size-fits-all model, the EOSC should anticipate and welcome EOSC users at sharing IFs, describing them, discovering them, and, via the EOSC instrument, find the means to reusing/adopting them. Ideally, the EOSC should facilitate scientists at converging towards common IFs to avoid redundancy, and advocating for differences where these clearly deliver added value.
- The scope of the EOSC IF should be broadened beyond AAI, metadata encoding, PIDs.
- The consultation document has a specific emphasis on FDO, which can only grasp a portion of the grand-challenge and is only one of a number of approaches to data interoperability which is needed or accepted in the EOSC community.
- Innovation should be part of all applicable action areas.
- The document should highlight the importance of an EOSC authority for IFs: (i) the need of communities to endorse specific interoperability frameworks via an EOSC branding, and (ii) the need to define the workflows and criteria through which frameworks are “submitted” and “evaluated” to become EOSC-compliant frameworks.
Recommendation: Take “AA7 – EOSC Interoperability Framework” at a higher level in the document structure to become one of the main pillars of the EOSC Office. Under the EOSC Interoperability Framework any action in the SRIA is called to nourish the set of an EOSC IF resources and therefore to contribute to the definition of a controlled, validated, shared EOSC commons.
3. Broaden User involvement
EOSC should be seen as a community-driven action. At the core should be the researcher, as well as society and academia, all who drive the needs and requirements. These needs and requirements are related not only with research performance but also with monitoring and evaluation, rewards and recognition. Academic-related issues are established by research communities and research performing organisations, vital organisations that contribute to EOSC. Once these are established, EOSC can build the relevant tools needed - for example, monitoring systems - on top of the existing infrastructures. EOSC can underpin and support this change rather than drive it, responding to user needs in an agile manner. More emphasis should be given to this aspect and local support for interacting with EOSC and responding to different levels and needs from local and institutional to national players.
Recommendation: Fully involve research performing organisations, universities and institutions in EOSC development. Clarity on member state involvement between national research infrastructures and interplay with EOSC.
For open science to be achieved and have an impact in all aspects of the research / data lifecycle EOSC needs to establish appropriate metrics and indicators, which will be accepted by all: researchers, research communities, research and e-Infrastructures, research performing organizations, evaluators of careers and of funding. The EOSCC Tinman report has clearly identified the need for an “An open metrics framework, which sets the rules (usage, performance, value for money) for the assessment of EOSC elements, i.e., policies, access framework, services, data, business, funding and usage models. This should include elements to facilitate the incentives and awards mechanism for researchers, as recommended by the EC HLEG on Next Generation Metrics and the EOSC Pilot policy group.” The FAIR Metrics constitutes only one part of this requirement, i.e., data vs. other resources, generation vs. use, providing a narrow view of the research process and the role of EOSC. OpenAIRE community strongly proposes that EOSC supports and implements the Open Metrics Framework, as part of a global initiative, with the following elements:
- Include metrics in EOSC that cover both monitoring and usage of all its resources, not just FAIR data.
- Build consensus on data used for metrics and indicators. Facilitate the interoperability of based on different monitoring services (formats and APIs) across disciplines and national infrastructures (services, open data…)
- Develop and embed in EOSC an open metrics infrastructure around the recommendations of the HLEG Next Generation Metrics and OSPP to combine more intelligent use of more diverse / open / auditable metrics, with qualitative assessment.
Recommendation: Replace "AA3 - FAIR metrics and certification" with "AA3 - Metrics and certification".
Extend its scope to:
- Include broader metrics for open science in EOSC, including both monitoring and usage aspects.
- Develop an EOSC Open Science Monitoring framework for all assets, provide incentives for local developments (national and RI) and facilitate its overall coordination.
- Put in place the appropriate monitoring mechanisms and services and provide the set of metrics specified by the OS Monitor for measuring Open Science
Extend the scope of AA12 - Rewards and recognition to:
- Implement the framework to be achieved in practice via an Open Metrics framework and its implementation
- Define mechanisms to require the use of a standardised metrics system in line with those proposed by research communities, funders, institutions.
5. “EOSC as an Enabler”- EOSC Innovation Radar
More emphasis should be given in the SRIA to supporting diversity and developing an agile EOSC that can adapt to the rapidly changing research landscape. While the SRIA certainly recognises a multi-stakeholder approach is taken across borders and disciplines, it is not necessarily ‘a given’ that the new standard ‘must be the sharing and reuse of data’. Rather it should allow for new standards and approaches to emerge from the communities. The language is very emphatic throughout which - while assertive and clear - should take into account future inevitable changes in the scholarly landscape. The emphasis on reuse of data and metadata could be widened to ‘research knowledge’ in general, to encompass the broad nature of research and science in society. Rather, investments should be made on incentivising current and future EOSC communities and operating a service that identifies and aligns current standards, for example, we cannot ‘demand’ that all repositories and services become interoperable however we can certainly make contents discoverable via data registries. It is yet unknown how many content providers and services will be using and interacting with EOSC but for certain it will be a diverse group with different standards, requirements, and research outputs.
Recommendation: EOSC’s mission and its users will evolve. Make reference to an ongoing roadmap to engage these actors, both new and existing, in the medium term. EOSC can be an environment to reflect and re-evaluate the functions of EOSC and to facilitate the involvement of all players.
6. Tackle legal and ethical aspects
The SRIA seems to rely on concepts such as Data Commons, which are very ambitious, however, it ignores the legal and ethical aspects connected with their realisation. This approach is problematic as it may compromise an otherwise excellent plan. The document only briefly mentions GDPR, and it completely disregards issues of data ownership, extractions and reuse based on copyright, database rights and other related rights (e.g. unfair competition, trade secrets, etc). It also omits to address the important contributions of the PSI legislation which will have an important impact on national open access policies for the reuse of research data. In conclusion, it seems that the document was drafted in absence of a copyright expert and all the issues connected with that have been somehow sidestepped, but it is not clear how they will be addressed.
Recommendation: Review the SRIA with legal aspects in mind and how to address and align national IPR and copyright developments which will affect reuse of data and content providers. GDPR and copyright rules are complex for researchers to grasp and implement at a day to day level and there must be sufficient support and priority given to these legal matters.
7. Business models
Data related services in EOSC currently follow an outdated unit cost model, which often does not agree with present-day usage patterns and cost/ revenue models. To foster a practical EOSC where service providers confidently provide long term services, EOSC must embrace business models for both public and commercial services that define arrangement of funding, money flows, acceptable and comparable cost and revenue models for usage, ensuring that there is room for innovative ones.
EOSC needs to involve overseeing activities for the development, coordination and adoption of cost and revenue streams as part of applicable EOSC business models that provide adequate incentives for users and providers. Furthermore, EOSC should facilitate activities to scout for, understand and pilot emerging, innovative business models that allow EOSC providers to provide their services in the emerging EU Data Spaces.
Recommendation: Expand “AA10 –Business Models” with priorities dealing with a) the definition of acceptable cost/revenue structures for different EOSC use cases, layers of EOSC, types of services, public or commercial, and b) the delivery of an in-depth macroeconomic quantitative assessment of the gains and costs for the implementation of the EOSC Core and Exchange.