Next club
ReadtheLOD 2024-07-15 Meeting Minutes
Index
Meta
- Paper: Converging on a Semantic Interoperability Framework for the European Data Space for Science, Research, and Innovation (EOSC) by Romain David et al.
- Date: July 15th, 2024 (14:00 - 17:00)
- Location: ONLINE
- Attending Members: Jorge Mendes, Marc Portier, Romain David (author of the paper)
Understanding the Paper
- (JM): The goal of the paper is not really clear. What value is it providing?
- This was a “teaser paper” aiming to fuel discussion within the EOSC-A, coming from the Semantic Interoperability Task Force. It also aggregated key takeaways from two years of task force work.
- (MP): It might be useful for readers to know some core concepts and existing material, such as:
- The four layers in the interoperability framework.
- FAIR-Digital Objects.
- Experience with the FIPs and DS-Wizard tool.
Semantic Business Objects
- Originating from GoFAIR, the idea is to ensure published semantic artefacts are published as FAIR Digital Objects (FDOs). However, how this translates into technology remains unclear.
Critique
- (JM): The paper could have been presented better. The writing style of different sections seems inconsistent, and there seems to be insufficient challenges from external readers. The paper lists future work but is unclear about existing results. As a teaser paper, its ambition was more about resolving conflicts through “radical collaboration,” a technique used in COVID data-sharing strategies to achieve consensus quickly.
- (MP):
- Detail: There are as many SI implementations as there are research communities, which might be too optimistic; the reality is worse.
- Focus on semantic artefacts: Reflecting Alan Kay’s quote, the paper explicitly states that a statistically relevant base is not available or achievable.
- Mapping strategy is inherently not scalable. A central focusing standard for mappings might be more effective.
- Friction to polish: Some friction might lead to quicker interoperability.
- DS-Wizard usage should come with a caveat: The semantic interpretation of questions depends on the interviewee, and the paper should mention this limitation.
Meeting Expectations
- (MP):
- On the positive side, the goal is to have machines assist humans.
- Qualifiers are not boolean, as mentioned in the definition box about “machine actionability.” This statement should be repeated, emphasizing that the unit to measure interoperability is lost money and time (euros).
- (JM): The word “convergence” is missing in the title. The promise of narrowing the field and tuning into an actual solution is not achieved.
- (MP): The paper reads more like a post-rationalization for its strategy rather than providing a good overview of available strategies. It misses motivation for not pursuing other strategies and lacks suggestions for closing identified gaps, such as making a statistically relevant base, ensuring shared methodology, and increasing expertise in formal modeling and semantic technology.
- (Romain): Wondering if the reading was a waste of time.
- (MP): Surely not. It provides a good overview, has solid plans, and has led to a worthwhile discussion.
Applying the Knowledge
- (MP): We should look into this MSCR from FAIRCORE4EOSC. MSCR = Metadata Schema and Crosswalk Registry with the Open-Science team at VLIZ.
- (JM): The cost of FAIRness is too high. Incentives are essential to engage people in applying more semantics in a broader and interoperable way.
Romain suggests checking a co-authored paper with Carole Goble: “Be sustainable: EOSC-Life recommendations for implementing FAIR principles in life science data handling.”
- (MP & JM): Consider alternative strategies and components not mentioned in the paper:
- Start with landscaping the needs and requirements to gauge the gap from actual needs.
- Regularly update and expose daring standards to quantify the gap.
- Introduce supported semantic-artefact-publication processes and pipelines.
- Invest in a technical interoperability web-based open publication to compare and quantify progress.
Any Other Business, Next Steps
- Next club meeting: Select a repository for further work and discussions.