FAIR Digital Object (FDO)

• FDO is a formal digital abstraction defined by persistent identifiers, typed metadata, and explicit operations that operationalize FAIR principles.
• FDO supports scalable, interoperable infrastructures by enforcing strict metadata profiles and machine-actionable protocols for seamless integration.
• FDO promotes automated discovery and reuse through standardized operations and controlled vocabularies, ensuring data is findable, accessible, interoperable, and reusable.

A FAIR Digital Object (FDO) is a formal, machine-actionable abstraction of a digital resource, defined by a persistent identifier, rigorously typed metadata, and explicit support for operations and relationships. FDOs operationalize the FAIR principles—Findability, Accessibility, Interoperability, and Reusability—by encapsulating digital entities in a domain-independent structure that enables automated reasoning, discovery, and reuse across scientific and industrial contexts (Blumenroehr et al., 2024, Blumenröhr et al., 7 Apr 2025). FDOs have become fundamental in the pursuit of scalable, interoperable, and semantically transparent digital infrastructures.

1. Formal Definition and Structural Elements

FAIR Digital Objects are characterized by a set of essential, typed components:

• Persistent Identifier: Each FDO is assigned a globally unique, resolvable identifier (PID, e.g., DOI or Handle) (Blumenroehr et al., 2024, Soiland-Reyes et al., 2023).
• Information Record: The core metadata of an FDO is a finite set of key–value pairs, each corresponding to a profile-registered attribute (often identified by PID), forming a minimal Information Record $R_f$ that instantiates exactly one Kernel Information Profile (KIP) $p\in P$ (Blumenroehr et al., 2024).
• Kernel Information Profile (KIP): KIPs enumerate the canonical set of typed attributes ($A_p$) that must be instantiated, with a non-empty mandatory subset (profile reference, license, checksum, location, creation date, resource type) (Blumenroehr et al., 2024, Inckmann et al., 22 May 2025).
• Operations: Each FDO supports machine-actionable operations associated through its typed attributes, enabling functions such as retrieve, validate_checksum, or traverse (Blumenroehr et al., 2024, Inckmann et al., 22 May 2025).
• Entity Relationships: FDOs reference other entities (typically other FDOs) through typed attributes with values as PIDs or URIs, forming navigable graphs (Blumenroehr et al., 2024, Blumenröhr et al., 7 Apr 2025).

The formal model can be summarized by the assertions: $$\forall f\in F\;\exists p\in P:\;\mathrm{Instantiate}(f,p)=R_f$$

$$\forall f\in F:\;\Bigl(\forall a_m\in A_m\;\exists v\in V:\;\mathrm{Create}(a_m,v)=\langle a_m,v\rangle\Bigr)\;\land\;\exists!\,i\in I:\;\mathrm{Assign}(i,R_f)=R^f_\mathrm{reg}$$

with operations and entity relationships governed by further formal predicates as specified in (Blumenroehr et al., 2024).

2. Typing Mechanisms and Operation Association

The association of operations with FDOs is governed by structured typing mechanisms, each supporting different levels of granularity, scalability, and interoperability (Blumenröhr et al., 7 Apr 2025, Inckmann et al., 22 May 2025):

• Record Typing (“duck typing”): Operations are associated ad hoc with individual FDOs by explicit reference attributes within their metadata records.
• Profile Typing (“nominal typing”): FDOs are instances of profiles, each carrying a predefined set of allowed operations. All FDOs of the same profile thereby share an operation list.
• Attribute Typing (“structural typing”): Operations specify attribute requirements. Any FDO whose instantiated attributes satisfy the required subset automatically becomes eligible for those operations.

The mathematical specification in (Blumenröhr et al., 7 Apr 2025) formally distinguishes these via graph-based representations, supporting efficient computation of available operations and enabling flexible interoperability across typing strategies.

Model	Simplicity	Efficiency (check)	Flexibility (updates)	Versatility	Interop
Record Typing	lowest	$O(|\mathrm{attr}(f)|)$	add-$o$: $O(|F'|)$, add-$f$: $O(|O'|)$	maximal	moderate
Profile Typing	moderate	$$O(|\mathrm{attr}(f)|+|\mathrm{Ops}_{\mathrm{prof}}|)$$	add-$o$: $O(|\mathrm{profiles~with~o}|)$, add-$f$: $0$	per-profile	high
Attribute Typing	largest	$O(|\mathrm{attr}(f)|+|R(o)|)$	add-$o$/f: $0$	constrained	high

Different usage scenarios motivate hybrid deployments, allowing expressive, scalable, and dynamic assignment of operations.

3. Semantic Modularity and Granular Composition

FDOs extend beyond flat datasets by adopting a modular architecture based on semantic units (Vogt et al., 30 Sep 2025, Vogt, 2023, Vogt et al., 2024):

• Statement Units (Atomic FDOs): The minimal, citable representation of a proposition, observation, or directive—each carrying its own PID, metadata, and content graph.
• Compound Units (Nested FDOs): Collections of statement units organized by subject, granularity, or context—e.g., all assertions about a specific entity aggregated as an “item unit.”
• Granularity Trees: Hierarchies mirroring biological or logical partonomies, with levels from atomic (identifier/datatype) up to item groupings and contextual aggregations.

The FDO meta-model mandates explicit type hierarchies, schema references, and logic declarations, ensuring both semantic transitivity (from natural language tokens to machine-actionable schemas) and format-agnostic interoperability (Vogt et al., 30 Sep 2025, Vogt et al., 2024). Each semantic unit is independently addressable, citable, and composed via graph relationships.

4. Interoperability, Machine-Actionability, and FAIR Services

FDOs are engineered for cross-domain operability and automation (Blumenroehr et al., 2024, Soiland-Reyes et al., 2023, Zoubia et al., 2024):

• Interoperability: All typed attributes and profiles are registered in centralized or federated type registries, with qualifiers referencing controlled vocabularies (typically by PID) (Blumenroehr et al., 2024, Inckmann et al., 22 May 2025).
• Machine-Actionability: Standard interfaces (e.g. DOIP, REST, SPARQL) expose FDO operations, permitting autonomous workflows to discover, validate, and process objects without human mediation (Zoubia et al., 2024). Attribute typing and operation registries further enable dynamic extension of available functions.
• FAIR Services: The ecosystem is supported by three core services (Vogt et al., 2024, Vogt et al., 30 Sep 2025):
  • Terminology Service (manages vocabularies and entity mappings, supporting both ontological and referential alignment),
  • Schema Service (registers data schemas, validates and mediates schema crosswalks),
  • Operations Service (registers, discovers, and deploys executable functions linked to FDO types and schemas).

These services are themselves implemented as collections of FDOs, with APIs designed for both human- and machine-consumability.

5. Implementation Architectures and Case Studies

Multiple operational platforms and reference implementations illustrate the FDO paradigm in practice:

• FDO Manager (Zoubia et al., 2024): Implements a minimal viable FDO infrastructure, with dual-PID management (object and metadata), schema.org/JSON-LD metadata models, and RESTful API access.
• Research Object (RO) Model in Earth Science (Garcia-Silva et al., 2018): Aggregates resources, workflows, annotations, and lifecycle metadata under FDO-compliant DOIs, with automated semantic enrichment and quality assessment pipelines.
• Integrated Data Type and Operations Registry (IDORIS) (Inckmann et al., 22 May 2025): Graph-based registry supporting inheritance, rule-based validation, and dynamic operation association for type-safe, extensible FDO workflows.
• Domain-Specific FAIRification (e.g., High Energy Physics) (Neubauer et al., 2022): Achieves FAIRness through DOI-minted code and metadata artifacts, standardized metadata schemas, and validation pipelines closely integrated into established research practices.

The table below summarizes representative instantiations:

Platform/Model	Core Features	Reference ID
FDO Manager	Dual PID, JSON-LD, REST API, payload preservation	(Zoubia et al., 2024)
Research Objects (RO)	RDF aggregation, DOI, semantic enrichment	(Garcia-Silva et al., 2018)
IDORIS	Type registry, inheritance, operation registry	(Inckmann et al., 22 May 2025)
HEP UFO Model	JSON Schema, DOI, validation, GitHub workflow	(Neubauer et al., 2022)

These implementations emphasize formal metadata commitment, preservation of provenance, and tight coupling between data and executable operations.

6. Evaluation Criteria, Adoption Challenges, and Future Outlook

Evaluation frameworks assess FDOs along dimensions of technical interoperability, semantic richness, compliance with FAIR guidelines, and middleware effectiveness (Soiland-Reyes et al., 2023, Blumenroehr et al., 2024):

• FAIR Compliance: Scored or checked via explicit alignment with principle-derived metrics (unique PID, machine-actionable metadata, schema references, provenance, license, etc.) (Soiland-Reyes et al., 2023, Blumenroehr et al., 2024).
• Interoperability Barriers: Major obstacles include the lack of universally adopted type/profile registries, protocol unfamiliarity (e.g., DOIP), and variability in metadata and licensing standards (Soiland-Reyes et al., 2023). A plausible implication is that tighter community specification and shared infrastructure are prerequisites for global alignment.
• Big Data Considerations: FDOs support scalability across the 5Vs (Volume, Variety, Velocity, Veracity, Value) by abstracting heterogeneous data in uniform, operation-enabled containers (Blumenroehr et al., 2024).
• Semantic Extension: The FAIR 2.0 initiative extends FDOs with explicit semantic interoperability requirements, deepening the granularity of schema and vocabulary mapping and advancing both machine and cognitive (human) interoperability layers (Vogt et al., 2024, Vogt et al., 30 Sep 2025).

Continuous evolution of FDO standards—especially in automated type-operation association, hybrid typing strategies, and adoption of semantic unit modularity—are regarded as key milestones for achieving an open, interoperable, and dynamic Internet of FAIR Data and Services (Blumenröhr et al., 7 Apr 2025, Vogt et al., 30 Sep 2025).