(Version 1.0)
Introduction
The MARTI (Metadata for AI Responsibility, Transparency, and Integrity) Framework employs a flexible and scalable object model to manage metadata across diverse disciplines. This model ensures consistency, transparency, and interoperability while allowing for customization to meet specific domain needs. By treating metadata as structured objects with core attributes, relationships, and dynamic behaviors, the object model provides a robust foundation for creating, managing, and sharing metadata.
For details on tailoring this framework to specific fields, see the companion document: Domain-Specific Attributes.

Core Principles of the Object Model
- Modularity: Metadata is composed of discrete, reusable units (objects) that can be customized or extended.
- Scalability: Supports use cases ranging from small datasets to large, complex systems.
- Interoperability: Designed to integrate seamlessly with existing metadata standards and systems.
Structure of an Object
Attributes
Definition: Descriptive fields that define the object (e.g., Title, Creator, Provenance).
Examples:
- Core Attributes: Universal across all disciplines (e.g., Title, Creator).
- Domain-Specific Attributes: Tailored to specific use cases (see Domain-Specific Attributes).
Relations
Parent-Child: Objects inherit and extend metadata from their parent (e.g., a book and its chapters).
Cross-Object Links: Connections between objects that share meaningful relationships (e.g., translations, adaptations).
Behaviors
Definition: Dynamic actions associated with the object (e.g., versioning, linking, cascading permissions).
Example: A digital book object updates all linked chapter objects when its metadata changes.
Identifiers
Definition: Unique keys for traceability and referencing (e.g., DOI, UUID).
Example: Persistent URL for online access.
Object Relationships
Hierarchical Model:
- Root Object: The primary metadata container (e.g., a “Work”).
- Child Objects: Subordinate components (e.g., chapters in a book).
Cross-Linking: Links between related objects, such as translations or adaptations.
Core Attributes
- Title: The name of the object.
- Creator: The individual or entity responsible for creating the object.
- Provenance: Tracks the object’s origin and history (creation, modifications, ownership transfers).
- RightsPermissions: Defines legal and usage constraints, including access restrictions and license types.
For a deeper dive into domain-specific extensions, see Domain-Specific Attributes.
Implementation Example
Object Type: Digital Book
Root Object Attributes:
- Title: Leaves of Grass
- Creator: Walt Whitman
- Provenance:
- {
- “creator”: “Walt Whitman”,
- “creationDate”: “1855”,
- “transfers”: [
- {“to”: “Library of Congress”, “date”: “1942”}
- ]
- }
- RightsPermissions:
- {
- “rightsHolder”: “Public Domain”,
- “license”: “No Restrictions”
- }
Child Objects:
- Poems: Individual poems in the collection, each with attributes:
- Title: Song of Myself
- Page Range: 15–37
- Unique Identifier: UUID-1234-5678
Cross-Object Links: Related editions and translations linked via unique identifiers.
Conclusion
The MARTI object model serves as the foundation for metadata management across disciplines. Its modular, scalable, and interoperable design ensures it can address diverse requirements while maintaining consistency.
For practical examples of how the model adapts to specific fields, see the companion document: Domain-Specific Attributes.
MARTI Record for This Document
{
“objectType”: “Documentation”,
“title”: “Object Model Overview”,
“creator”: “Carrie Bickner”,
“creationDate”: “2024-11-19”,
“provenance”: {
“creator”: “Carrie Bickner”,
“creationDate”: “2024-11-19”,
“modifications”: []
},
“rightsPermissions”: {
“rightsHolder”: “Carrie Bickner”,
“license”: “CC BY-SA 4.0”
},
“attributes”: {
“version”: “1.0”,
“documentType”: “XHTML”,
“relatedDocuments”: [“Domain-Specific Attributes”]
}
}
MARTI Record for This Document
{
“objectType”: “Documentation”,
“title”: “Object Model Overview”,
“contributors”: [
{“role”: “Document Architect”, “name”: “Carrie Bickner”},
{“role”: “Collaborator”, “name”: “ChatGPT”}
],
“creationDate”: “2024-11-19”,
“provenance”: {
“creator”: “Carrie Bickner”,
“creationDate”: “2024-11-19”,
“modifications”: [
{“modifiedBy”: “ChatGPT”, “date”: “2024-11-19”, “change”: “Enhanced content and structure”}
]
},
“rightsPermissions”: {
“rightsHolder”: “Carrie Bickner”,
“license”: “CC BY-SA 4.0”
},
“attributes”: {
“version”: “1.0”,
“documentType”: “XHTML”,
“relatedDocuments”: [“Domain-Specific Attributes”]
}
}
MARTI Record for This Document
{
“objectType”: “Documentation”,
“title”: “Object Model Overview”,
“contributors”: [
{“role”: “Document Architect”, “name”: “Carrie Bickner”},
{“role”: “Collaborator”, “name”: “ChatGPT”}
],
“creationDate”: “2024-11-19”,
“provenance”: {
“creator”: “Carrie Bickner”,
“creationDate”: “2024-11-19”,
“modifications”: [
{“modifiedBy”: “ChatGPT”, “date”: “2024-11-19”, “change”: “Enhanced content and structure”}
]
},
“rightsPermissions”: {
“rightsHolder”: “Carrie Bickner”,
“license”: “CC BY-SA 4.0”
},
“attributes”: {
“version”: “1.0”,
“versionHistory”: [
{“version”: “1.0”, “date”: “2024-11-19”, “description”: “Initial release of the Object Model Overview”}
],
“documentType”: “XHTML”,
“relatedDocuments”: [“Domain-Specific Attributes”]
}
}
Leave a reply to Understanding MARTI: A New Metadata Framework for AI – Zeldman on Web and Interaction Design Cancel reply