Rec. 8: Cross-disciplinary FAIRness

[sjDCC]

Interoperability frameworks should be articulated in common ways and adopt global standards where relevant to enable interdisciplinary research. Common standards, intelligent crosswalks, brokering mechanisms and machine-learning should all be explored to break down silos.

• Efforts should be made to identify information and practices that apply across research communities and articulate these in common standards that provide a baseline for FAIR. Stakeholders: Standards bodies; Research communities.

• Case studies for cross-disciplinary data sharing and reuse should be collected. Based on these case studies, mechanisms that facilitate the development of frameworks for interoperability and reuse should be developed. Stakeholders: Global coordination fora; Data stewards.

• The components of the FAIR ecosystem should adhere to common standards to support disciplinary frameworks and to promote interoperability and reuse of data across disciplines Stakeholders: Data services; Research communities; Global coordination fora.

[asconrad]

This is to a large extent what EOSC is about - if I have understood it rightly. Not sure if it is wise to extend this to be part of FAIR action plan - or to solve in the framework of EOSC implementation.

[ghost]

4TU.Centre for Research Data position: The way the interoperability frameworks are described here, makes the impression to us, that they should relate to the policy document, and also is connected to the DMP. Having these frameworks linked to the FAIR components, and machine-actionable seems reasonable. Adding this element to the FAIR components, creates a very complex and high-maintenance environment.

[AlasdairGray]

The term "intelligent crosswalk" needs to be properly defined

[AlasdairGray]

Machine learning is a specific set of tools to achieve some goal. The recommendation should state the goal not the approach by which it is achieved.

[katerbow]

DFG position: In principle, there also is agreement to this recommendation, in particular, with respect to foster cross- and interdisciplinary research. It is in full alignment with DFGs funding principles to support sciences und humanities, in particular in exploring new ways in research. That includes developing of interoperability frameworks including standards, brokering mechanisms, machine-learning and other information technologies and methods with the support of research and science.

[eiroforum-it-wg]

EIROforum has published its input for the consultation which is available online (20180724-EIROforum-position-paper-EOSC.pdf). The paper highlights a number of practical points EIROforum members consider essential to ensure the EOSC can effectively interlink People, Data, Services and Training, Publications, Projects and Organisations, including aspects related to Rec. #8 “Cross-disciplinary FAIRness”.

[bertocco]

INAF (astronomy) position: There are examples, like RDA, to help bridging disciplines boundaries.

[pkdoorn]

This boils down, IMHO, to stimulating Linked Open Data and semantic web technologies. Why not mention this explicitly here?

[mromanie]

ESO position The recommendation is certainly agreeable in principle, but it should not results in a system that is so complex that it doesn't work for individual disciplines.

Question for clarification (and curiosity): in what ways does machine-learning help in breaking down silos?

[gtoneill]

Some overlap with Recommendation 7 related to disciplines and interoperability. Perhaps merge?

[MSoareses]

On item 2 of this recommendation data services and publishers are stakeholders. Mechanisms that facilitate interoperability and data re-use can be supported through a diverse and complementary ecosystem of repositories/data services from discipline-specific ones to broad scope repositories (at Elsevier Mendeley Data serves the scientific community broadly). Not necessarily along the complete data journey but certainly when it regards peer-review then journals can be expected to involved and establish links between peer reviewed articles and data services where underlying results are deposited.

On item 3 of this recommendation I would remit to the comment in Rec. https://github.com/FAIR-Data-EG/Action-Plan/issues/4. Publishers alongside with data services are part of the FAIR ecosystem that can support common standards, disciplinary frameworks and promote interoperability.

[npch]

SSI position:

We agree with these recommendations.

An example of a crosswalk in the research software metadata area is CodeMeta.

As an additional comment, if software (e.g. machine learning) is to be used to break down silos between disciplinary data holdings then that software needs to be well managed. It will be important to understand the provenance of the decisions made, to test correctness and test improvement between different versions.

[rachbruce]

@jisc fully support the recommendation on the importance of cross-disciplinary FAIRness. In research practice, cross-disciplinary FAIRness can ultimately contribute to the growth of new, interdisciplinary research. The identification of cross-disciplinary, common needs can also facilitate the development of more sustainable services, which support a broader group of users than discipline-specific solutions. However, until such services start to emerge, continued funding for ongoing and new work on cross-disciplinarity is needed. In addition there are issues that still need to be addressed within interdisciplinary research, and ways to support re-use and translation between disciplines, and also use by wider non-academic audiences requires much further investment.