FAIR Digital Objects (FDOs)

• FAIR Digital Objects (FDOs) are self-contained, machine-actionable representations that encapsulate both digital data and essential metadata.
• They adhere to FAIR principles—findability, accessibility, interoperability, and reusability—using persistent identifiers and standardized metadata schemas.
• Their structured workflows enable end-to-end provenance capture and cross-domain applicability in fields like materials science, computational chemistry, and imaging.

A FAIR Digital Object (FDO) is a self-contained, machine-actionable entity that encapsulates both data and all metadata required to make that data Findable, Accessible, Interoperable, and Reusable for both human and computational agents. Each FDO is globally identified and described by a formal metadata schema that records provenance, operational environment, computational workflow, and output properties. The concept is underpinned by strict adherence to the FAIR principles, ensuring objects are discoverable, retrievable, integrable, and replicable across heterogeneous domains (Shoghi et al., 6 Aug 2024).

1. Conceptual Foundations and FAIR Principles

FAIR Digital Objects are formally defined as machine-actionable representations of any digital resource that bundle a minimal, typed metadata record and expose a uniform interface for discovery, access, and operations. FDOs are registered under persistent, resolvable identifiers and interrelate through typed entity relationships, forming directed graphs of digital resources (Blumenroehr et al., 27 Nov 2024).

Four FAIR Principles:

• Findable: Each FDO has a persistent unique identifier (PUID), typically generated by hashing all mandatory metadata fields, and is indexed with rich metadata (keywords, subject, description).
• Accessible: The data and its metadata are retrievable via standardized protocols (e.g., HTTP, DOIP), with explicit access and authentication encoded in metadata.
• Interoperable: Metadata fields utilize controlled vocabularies and formal schemas (e.g., JSON Schema, crosswalks to Dublin Core/DataCite/CodeMeta) to integrate seamlessly with other objects and systems.
• Reusable: Provenance, licensing, versioning, and context information are captured to support experimental replication and derivative use (Shoghi et al., 6 Aug 2024).

2. Structural Model and Metadata Schema Composition

FDOs are structured around a compositional metadata schema encompassing five primary entities:

• User (U): Includes creator, contributors, ORCID, affiliations, license, access rights.
• System (S): Captures software, operating system, and processor details elucidating the computational environment.
• Job (J): Dissects workflow into geometry (e.g., RVE_size, discretization), material model (e.g., constitutive laws, orientations), and boundary conditions (mechanical, thermal).
• Property (P): Catalogs homogenized simulation or experimental outcomes such as stress/strain tensors and derived quantities.
• Units: Maps every physical property to standardized units (MPa, mm, K).

The metadata schema implements strict relationships among these components, enforceable via class diagrams and formal schema definitions. For computational workflows, each execution instance yields a unique FDO by aggregating (U, S, J, P) and outputting a record whose ID is given by hashing all mandatory fields: $$ID = \mathrm{HDF5\_hash}\left(\bigcup_{\text{mandatory fields}} \text{value}\right)$$ Full formalization covers JSON-schema fragments for geometry and boundary conditions. For example, geometry metadata is encoded as: $$\texttt{Geometry}:\{~\texttt{rve\_size}:[\mathbb{R}]^3,~\texttt{rve\_continuity}:\{\text{"True","False"}\},~\texttt{discretization\_type}:\{\text{"Structured","Unstructured"}\},~\texttt{discretization\_unit\_size}:[\mathbb{R}]^3,~\texttt{discretization\_count}:\mathbb{N}\}$$ (Shoghi et al., 6 Aug 2024).

3. FDO Generation Workflows and Provenance Capture

Workflow-centric FDO creation involves:

1. Workflow submission: Automated capture of user and system metadata (e.g., ORCID, software versions).
2. Job parsing: Extraction of geometry, materials, and boundary condition details.
3. Execution: Computational solver produces raw output data.
4. Post-processing: Calculation and normalization of physical properties.
5. Assembly: Aggregation of metadata, hashing for ID creation, attachment of units.
6. Registry entry: FDO record deposited into a machine-readable datastore.

Each workflow run produces a unique FDO described by

$\mathrm{FDO}_i = (U_i, S_i, J_i, P_i, ID_i)$

with provenance recorded end-to-end (user, system, job, result). This structure assures traceability and full redeployment of computational experiments (Shoghi et al., 6 Aug 2024).

4. Domain-Agnostic Generalization and Data Ecosystem Integration

Although motivated in micromechanics and simulation-based materials science, the workflow-centric FDO infrastructure generalizes provided:

1. The process consists of well-defined workflow steps.
2. Each step’s input, environment, and output can be captured as formal metadata.
3. Adopted schema allows formulation of “Job” and “Property” analogues.

Examples include:

• Computational chemistry: FDOs encode molecule geometry, basis sets, and spectral outputs.
• Imaging experiments: Metadata covers microscope parameters, sample preparation, and image outputs.
• Additive manufacturing: FDOs represent machine settings and microstructure results.

This abstraction ensures cross-domain interoperability and uniform data re-use patterns (Shoghi et al., 6 Aug 2024).

5. Mechanisms for Associating FDOs with Operations

Operationalization of FDOs is governed by three emergent typing models (Blumenröhr et al., 7 Apr 2025):

• Record Typing: Each FDO information record embeds direct pointers to executable operations—optimal for static operation sets but brittle to change.
• Profile Typing: FDOs reference profiles; each profile maps to a standard operation suite. This model is suited for domain classes with stable operation sets.
• Attribute Typing: Operation discovery is driven by attribute–type definitions in the data FDO and dependency-matching via operation FDO metadata, enabling zero-maintenance, automated matching in metadata-rich domains.

Each model is formalized as a directed graph and evaluated for simplicity, lookup efficiency, update flexibility, granularity, required client knowledge, and overall versatility. The optimal association mechanism depends on the scale of FDO deployment, the dynamism of operations, and metadata richness.

6. FAIRness Metrics, Validation, and Cross-Domain Adoption Criteria

FAIRness assessment integrates both qualitative and formal quantitative metrics:

• Completeness: Fraction of required metadata present.
• Identifier resolution: Fraction of PUIDs that resolve correctly.
• Interoperability: Adherence of metadata to prescribed schemas and controlled vocabularies.

Mandatory adoption criteria include:

• Community-wide kernel profile and attribute governance.
• Persistent ID services (Handle, DOI) supporting machine-actionable records.
• Automated validation tools for ingestion and object compliance.
• Protocols for PID resolution and operation invocation (e.g., REST/DOIP).

Limitations arise from inconsistent schema extension policies, lack of unified operation linking mechanisms, and out-of-scope security considerations (Blumenroehr et al., 27 Nov 2024).

Typing Model	Lookup Efficiency	Flexibility	Granularity
Record Typing	Best (O(1))	Worst	Highest per-FDO
Profile Typing	Moderate	Moderate	Bundled by class
Attribute Typing	Expensive	Best	Fine-grained metadata

7. Impact, Best Practices, and Future Outlook

FDO frameworks enable scalable, provenance-rich digital ecosystems supporting global alignment of data spaces. By decoupling operational semantics, metadata, and bit-stream representations, FDOs provide a pathway for robust research data stewardship spanning scientific, technical, and industrial contexts.

Best practices include:

• Publishing digital objects with persistent, machine-resolvable identifiers.
• Maintaining extensible, domain-agnostic metadata schemas, mapped to community standards.
• Supporting transparent provenance, licensing, and versioning for each FDO.
• Integrating dataset and workflow publication with open protocols and reusable operation registries.
• Ensuring the ability to automate end-to-end workflows and re-use scenarios with minimal human intervention.

Continued work is required to harmonize metadata profiles, operation registries, and validation regimes to consolidate the global FDO ecosystem (Blumenroehr et al., 27 Nov 2024, Shoghi et al., 6 Aug 2024, Blumenröhr et al., 7 Apr 2025).