Appendix A: Glossary of Terms¶
Terms are listed alphabetically. This glossary covers terminology, acronyms, and external standards used in the Blueprint without an in-text explanation. Sources are provided for each entry and hyperlinked where available. Definitions have been prepared with AI assistance and represent a combination of direct quotations, paraphrases, and working definitions informed by the cited sources. Where precise or authoritative definitions are required, readers are encouraged to consult the cited source directly.
| Term | Definition (with source) |
|---|---|
| Anonymisation and Pseudonymisation | Anonymisation removes or alters personal data so that individuals can no longer be identified, even indirectly. Pseudonymisation replaces identifying details with artificial identifiers, so that re-identification remains possible only with additional information held separately and securely. (Source: Information Commissioner's Office) |
| API (Application Programming Interface) | A defined set of rules and protocols that allows different software systems to communicate with one another, for example, enabling the Data Director to exchange information with a repository, registry or identifier service. (Source: MDN Web Docs) |
| CARE Principles | Collective Benefit, Authority to Control, Responsibility and Ethics: principles developed to ensure Indigenous Peoples retain control over how data about their communities, lands and knowledge is collected, used and shared, complementing the FAIR principles. (Source: Carroll, S.R. et al. (2020). The CARE Principles for Indigenous Data Governance. Data Science Journal, 19(1), p.43. https://doi.org/10.5334/dsj-2020-043) |
| CRediT (Contributor Roles Taxonomy) | A standardised taxonomy of contributor roles used to describe the specific nature of each person's contribution to a research output. Referenced as the basis for recording human roles in provenance records. (Source: CRediT) |
| CRIS (Current Research Information System) | An institutional database system used to capture, manage and report on information about research activity, outputs, funding and personnel. (Source: euroCRIS) |
| Crosswalk | A mapping that shows how fields, terms or elements in one metadata schema correspond to equivalent fields in another, enabling records created for one system to be understood and reused in another. (Source: Library of Congress) |
| Data Dictionary | A document that lists and describes the variables, fields and codes used in a dataset, including their definitions, units and permitted values, helping others understand and reuse the data. (Source: The Turing Way) |
| Deposit (Research Data) | The act of submitting a research dataset to a repository for long-term storage, preservation, and access. Distinct from publication in that deposit is the researcher's act of submitting data, while making data publicly available is typically carried out by the repository following the deposit. The Data Director supports researchers in preparing for and completing deposits; it does not itself publish or make data publicly available. (Source: OECD Principles and Guidelines for Access to Research Data from Public Funding, 2007; see also COAR Controlled Vocabulary for Resource Type Genres) |
| Digital Object Identifier (DOI) | A persistent, internationally standardised identifier used to uniquely and permanently identify a digital object such as a journal article, dataset or report, ensuring it remains findable even if its web location changes. (Source: DOI Foundation; see also ISO 26324:2022, Information and documentation - Digital object identifier system)" |
| Dublin Core | A widely used, general-purpose metadata standard consisting of a small set of core elements, such as title, creator, date, and subject, for describing digital and physical resources, designed to support resource discovery across domains. (Source: Dublin Core Metadata Initiative) |
| Edge Deployment | An approach to deploying computing infrastructure close to where data is generated or used, such as on a local server or device, rather than relying on centralised cloud data centres, reducing latency and supporting data residency requirements. (Source: ISO/IEC TR 23188:2020, Information technology - Cloud computing - Edge computing landscape, ISO/IEC Joint Technical Committee 1, Subcommittee 38) |
| Embargo | The set period of time by which investigators retain exclusive rights to their dataset while it has been deposited in a repository but before it is made publicly available. (Source: SPARC Portal documentation, https://docs.sparc.science/docs/embargoed-data, adapted for general use) |
| FAIR Digital Object (FDO) framework | A machine-actionable digital object that bundles data, metadata, type information and a persistent identifier together, enabling software systems to automatically interpret, validate and act on its content and structure. (Source: FAIR Digital Object Forum) |
| FAIR Principles | Findable, Accessible, Interoperable and Reusable: a set of guiding principles for making research data, software and other digital objects usable by both humans and machines, and a cornerstone of open science policy. (Source: Wilkinson, M.D. et al. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data, 3, 160018. https://doi.org/10.1038/sdata.2016.18) |
| Federation | A model for connecting multiple independent research data systems, repositories or infrastructure nodes so that resources, services and datasets can be discovered, accessed and used across them without being centralised or merged into a single system. Each participant retains ownership and control of their own resources. (Source: European Open Science Cloud (EOSC) Architecture Documentation%22). 'The European Open Science Cloud (EOSC) is a federated system designed to enable seamless sharing, discovery, and access to research resources across Europe. It consists of interconnected EOSC Nodes, which can be national, regional, or thematic infrastructures that provide services, datasets, and computational resources to the research community.' |
| GUPRI (Globally Unique, Persistent and Resolvable Identifier) | An identifier that is globally unique, remains valid and resolvable over time, and can be used by humans and machines to access information about the digital object it identifies. As described in the EOSC PID Policy, a resource that is FAIR must be assigned an identifier that meets all three criteria. (Source: EOSC FAIR Working Group and Architecture Working Group. A Persistent Identifier (PID) policy for the European Open Science Cloud. European Commission Publications Office, October 2020. See also FAIRsharing documentation on GUPRIs. |
| Hallucination (AI) | A term used to describe instances where an AI system generates content that is confidently stated but factually incorrect, erroneous, or not grounded in verifiable evidence. The term is contested: the output may be consistent with patterns in training data while still being factually wrong or insufficiently evidenced for the purposes for which it is used. NIST AI 600-1 uses the more precise term confabulation to describe the production of confidently stated but erroneous or false content. (Source: NIST AI 600-1, Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile, National Institute of Standards and Technology, 2024) |
| Handle System | A system for assigning and resolving persistent digital identifiers, known as Handles, which provide a stable reference to a digital resource regardless of changes to its location. (Source: Handle.net Registry) |
| Infrastructure as Code (IaC) | The practice of defining and managing computing infrastructure, such as servers, networks and storage, through version-controlled configuration files rather than manual processes, enabling consistent, repeatable and auditable deployment. (Source: Microsoft Learn) |
| IRB / REC (Institutional Review Board / Research Ethics Committee) | A formally constituted committee responsible for reviewing and approving the ethical aspects of research involving human participants, animals or sensitive data, ensuring that research meets internationally recognised ethical standards before it proceeds. Known as an Institutional Review Board (IRB) in the United States and a Research Ethics Committee (REC) in many other countries. (Source: World Health Organization. Standards and operational guidance for ethics review of health-related research with human participants. WHO, 2011. See also CIOMS International Ethical Guidelines for Health-related Research Involving Humans, 2016, developed in collaboration with WHO) |
| JSON Schema | A vocabulary for describing the structure, content and validation rules of JSON (JavaScript Object Notation) data, used to check that a metadata record or dataset conforms to an expected format. (Source: JSON Schema) |
| Knowledge Graph | A structured network of entities, such as datasets, people, organisations or concepts, and the relationships between them, represented in a machine-readable form that supports search, discovery and reasoning. (Source: W3C) |
| Machine-readable | Describes information that is structured and formatted so that it can be automatically read, parsed and processed by computer systems, as distinct from formats intended primarily for human reading. (Source: UNESCO Recommendation on Open Science) |
| Metadata | Structured information that describes, explains, or locates a resource, enabling it to be found, understood, and reused. (Source: Dublin Core Metadata Initiative / ISO 15836). For the purposes of this Blueprint, metadata is understood in its full scope, encompassing not only discovery fields but variable definitions, data structure, provenance, quality assessments, access conditions, and applicable governance constraints. See Section 3.1.1. |
| MLCommons Croissant | A community-developed, open metadata format for describing machine learning datasets in a structured, machine-readable way, designed to make datasets easier to find, load and use across different tools. (Source: MLCommons) |
| Model Card / Dataset Card | A short, structured document accompanying a machine learning model or dataset that records its intended use, composition, known limitations and evaluation results, intended to improve transparency and reusability. (Source: Hugging Face) |
| Nanopublication | A small, citable unit of scientific knowledge expressed as a single machine-readable assertion together with information about its provenance and supporting evidence, designed to be linked together to form larger knowledge graphs. (Source: Nanopublications) |
| NIST (National Institute of Standards and Technology) | A US government agency that develops widely adopted standards and frameworks, including in cybersecurity and AI risk management, often used as a reference baseline for security practice internationally. (Source: NIST) |
| ODRL (Open Digital Rights Language) | A standardised, machine-readable language for expressing permissions, prohibitions and obligations relating to the use of digital content and data, such as licensing or access conditions. (Source: W3C) |
| Ontology | A formal, machine-readable representation of the concepts within a domain and the relationships between them, used to support consistent classification, search and integration of data across systems. (Source: W3C Semantic Web) |
| Open Science | A broad movement and set of practices aimed at making the processes, outputs and tools of research, including publications, data, methods and software, openly available, accessible and reusable for the benefit of science and society. (Source: UNESCO Recommendation on Open Science) |
| Open-Source Licence (e.g. MIT, Apache 2.0) | A type of software licence that grants users the right to freely use, modify and distribute software. The MIT licence is short and highly permissive; the Apache 2.0 licence is similarly permissive but additionally includes an express patent licence. (Source: Open Source Initiative) |
| ORCID | A free, persistent identifier that researchers use throughout their career to distinguish themselves from others and to link their identity to their research outputs, affiliations and activities. (Source: ORCID) |
| Persistent Identifier (PID) | An identifier assigned to a digital resource, person or organisation that is globally unique, remains valid over time, and can be resolved by both humans and machines to locate or access the identified entity. A resource that is FAIR must be assigned a PID that meets these three criteria of global uniqueness, persistence, and resolvability. (Source: EOSC FAIR Working Group and Architecture Working Group. A Persistent Identifier (PID) policy for the European Open Science Cloud. European Commission Publications Office, October 2020 |
| PROV-O (PROV Ontology) | A standard vocabulary for recording provenance information in a machine-readable form, describing the entities, activities and people involved in producing, influencing or delivering a piece of data. (Source: W3C) |
| RAG (Retrieval-Augmented Generation) | An AI technique in which a language model's response is grounded by first retrieving relevant information from an external knowledge source, such as a document collection or database, and supplying it as additional context before generating an answer. (Source: AWS — What is RAG?) |
| Repository (Research Data Repository) | An online platform for storing, preserving, describing and providing access to research data and other outputs over the long term, often assigning persistent identifiers and supporting discovery through metadata. (Source: The Turing Way) |
| ROR (Research Organization Registry) | An open, community-led registry providing persistent identifiers for research-performing organisations, used to unambiguously link research outputs, funders and institutions. (Source: ROR) |
| SHACL (Shapes Constraint Language) | A standard language for describing and validating the structure of data, used to check that metadata conforms to a defined shape or set of rules. (Source: W3C) |
| WCAG (Web Content Accessibility Guidelines) | An internationally recognised set of guidelines defining how to make web content more accessible to people with disabilities. (Source: W3C Web Accessibility Initiative) |
| XSD (XML Schema Definition) | A standard for describing the structure, content and permitted data types of an XML document, used to validate that an XML file conforms to an expected format. (Source: W3C) |