FAIR-IMPACT recognises the essential role of governance in the development and maintenance of semantic artefacts. To address this, the project has dedicated a specific task aimed at defining and expanding on the concept of "Semantic Artefact Governance". Through the FAIR-IMPACT Semantic Artefact Governance workshop, we explored a variety of governance strategies and practices across diverse EOSC-related communities. This effort enabled us to generate recommendations for initiatives and communities seeking to implement governance methods throughout their semantic artefact lifecycle.
Semantic Artefact Governance Definition
Despite the acknowledged importance of governance for semantic artefacts, there is no standard definition available in the literature, underscoring the complexity of this task. Governance is a structured framework that ensures quality ,transparency, and consistency, maintaining the accountability, relevance, and usability of resources for their intended purposes. It involves defining policies, standards, and responsibilities to manage the resource's lifecycle and sustain alignment with initiative’s goals. Implementing governance practices in a team requires role clarity, stakeholder engagement, and regular reviews.
By adapting definitions of data governance to ensure theoretical soundness, we developed the following definition of Semantic Artefact Governance:
|
Semantic Artefact Governance Component
Governance Framework
Governance Framework defines a set of high-level rules and agreed-upon specifications to provide management mechanisms, supports the development, maintenance, and use of standard tools, facilitates the application of best practices, and ensures high level of quality. By reusing the pattern established by the BASF Governance Operational Model for Ontologies (GOMO), which is the result of a collaborative effort between industry and academia in the semantic web field, we provide a unified governance structure aimed at enhancing interoperability across initiatives adopting the framework.
Principles in semantic artefact governance can take the form of simple statements or concrete policies that address specific aspects of the semantic artefact lifecycle, such as metadata, versioning, information model, etc. Each principle should be supported by Recommendations, offering guidance on implementing the principle while providing necessary background knowledge for carrying out related activities. Recommendations lead to Standards, derived from best practices, which set norms or requirements to ensure consistent and effective governance. These standards enable human- or software-based evaluations through a Quality Control process to verify correct implementation. Lastly, training programs emphasise the importance of educating stakeholders on using these recommendations and standards effectively, ensuring the principles are applied consistently.
In the following, we provide an example of how the framework can be applied in practice. We take one of the FAIR Principles, specifically I1: '(Meta)data use a formal, accessible, shared, and broadly applicable language for knowledge representation,' and demonstrate its implementation within the framework in the following graph.
Engineering Methodology
Engineering Methodology comprises the development and maintenance processes, methods, and tools for managing semantic artefacts. The Engineering Methodology varies depending on the type of semantic artefacts under development, the strategy which would be either building them from scratch or reusing existing semantic artefacts, the associate domain, and the maturity level of infrastructure, data, tools, and processes in semantic artefact management. Despite this variety, Engineering Methodology structures activities, tasks, and operations into systematic and formal phases necessary for managing SA lifecycle such as conceptualization, encoding, and publication. Initiatives could either develop their own methodology, workflows and related tools, customise all steps of SA lifecycle based on their requirements and objectives, or adapt existing methodology and utilise available tools and services to adhere to best practices and increase interoperability with others initiatives. Further details are referenced in FAIR-IMPACT’s T4.2.1, which produced a revised Linked Open Terms methodology (FAIR by design methodology) through identifying gaps in current practices for ontology engineering and proposing solutions to address them.
In the following, we highlight several key aspects of the semantic artefact lifecycle that should be addressed during the development, publication, maintenance, and consumption phases of a semantic artefact. These aspects represent primary topics that should be formalised through the Governance Framework introduced in the previous section.
- Availability (licencing)
- Access rights and security policy
- Documentation
- Relevant attributes, metadata and provenance
- Unique identification for (meta)data
- Versioning
- Deprecation
- information model
- (Meta)data reuse
- Visualisation
- Mapping
- Modularity and extensibility
- (Meta)data access and semantic repository
- Naming Conventions
- Scope
- Language
- Commitment to collaboration
- Channels for communications and contribution
Organisational Structure
Organisational Structure outlines how specific activities are allocated and directed to achieve the organisation’s goals. It defines decision-making, accountabilities, roles, responsibilities, and the engagement of stakeholders. Operational mechanisms, such as Standard Operating Procedures (SOPs), can be employed to disseminate tasks among actors. Through examining communities of practice, we identified key domains of activity essential for initiatives to recognize. These domains should be addressed according to the initiative’s level of maturity in semantic artefact management and its available human resources.
Editorial: Focuses on developing, maintaining, and updating the principles and standards of the initiative. This includes creating, implementing, and overseeing review processes for these standards, as well as establishing policies that ensure editorial consistency.
Technical: Responsible for the upkeep and administration of the technical infrastructure. This includes managing websites, repositories, APIs, SPARQL endpoints, and other related services and workflows, ensuring they are up-to-date, secure, and accessible.
Outreach: Manages the organisation’s public relations and external communications. This includes monitoring communication channels, producing documentation and educational resources, and representing the organisation at events and other external venues to enhance engagement and awareness.
Coordination: Oversees the consistency and alignment of governance policies with organisational standards, ensuring technical compliance. This area supports best practices and coordinates training for curators, fostering collaborative interactions between teams to reinforce policy coherence.
SA Development: Encompasses the full lifecycle of developing and releasing semantic artefacts, from initial conception to final deployment. This includes iterative development processes, version management, and ensuring that artefacts are released with documentation and clear functionality for end users.
SA Curation: Ensures that semantic artefacts are current and aligned with ongoing requirements. This includes managing lifecycle aspects like versioning, deprecation, and identification, while also engaging with the community, expert networks, and external entities to maintain relevance and foster collaboration.
Expertise: An advisory group of subject-matter experts providing specialised knowledge to support ontology validation and curation. This team promotes and guides the use of semantic artefacts, offering recommendations for content evolution, reuse, and adaptation while validating changes and improvements.
Community: Foster an inclusive, collaborative network of stakeholders involved in the development, use, and governance of semantic artefacts. This involves engaging users, curators, and contributors, feedback loops, and communication channels that facilitate active participation and knowledge sharing.
Semantic Artefact Governance Recommendations
The exploration of semantic artefact governance within the EOSC context (and beyond) highlights its multifaceted and critical role in digital data management and interoperability. While the interaction between actors and their accountability for different domains of the semantic artefact life cycle is crucial, it must be guided by well-defined rules aligned with the organisation's or community's goals. A holistic view is essential, encompassing all requirements for semantic artefact management, from the maturity level of infrastructure and processes to the capabilities of human resources.
Following the structured methodology, we analysed practices, principles, and resources for semantic artefact governance to create tailored recommendations that cater to the specific requirements of distinct target groups.
1) Project-based initiatives managing single-core semantic artefact, such as application ontologies (e.g., EFO and SAREF) or Linked Open Data sets (e.g., AGROVOC). Within this category, semantic artefact is developed to address the need for standard vocabulary in a target domain. These semantic artefacts enable semantic interoperability and support the creation of knowledge graphs by providing domain-specific concepts and properties, enhancing data integration between systems that deploy them.
2) Support-based initiatives direct communities in developing and maintaining semantic artefacts. Individuals or groups of researchers independently build and maintain artefacts in a decentralised manner. Organisations such as INRAE Voc and NFDI4Biodiv operate within this category.
3) Harmonisation initiatives that provide principles and guidelines, along with infrastructure, to harmonise the management of semantic artefact across communities or organisations. Developers can use the upper-level ontologies, modularisation methods, and other standard workflows and practices established by these initiatives to ensure they develop interoperable semantic artefact. The OBO Foundry is an example of this category, serving as a pivotal point in using semantic web technologies in biological and biomedical sciences. BASF also falls into this category but restricts its developed ontologies to internal organisational use.
Here you can find the document:
https://fair-impact.eu/sites/default/files/2024-11/Semantic%20artefacts%20governance%20page.docx.pdf