IDFThe DOI Kernel Metadata Declaration defines the minimum set of metadata that must accompany every Digital Object Identifier (DOI) name. Maintained by the International DOI Foundation (IDF) and formalized as part of ISO 26324, the kernel ensures baseline descriptive and administrative information is always available for any DOI-identified entity, regardless of which Registration Agency issued the identifier or what domain the resource belongs to. Version 2.2 of the kernel XML schema is the current release, with allowed value sets last updated in July 2020.
Background
The Digital Object Identifier system emerged in the late 1990s from the publishing industry's need for persistent identifiers for digital content. The DOI Data Model, which includes the kernel, draws on the indecs (interoperability of data in e-commerce) Framework, an influential metadata interoperability project (1998--2000) that pioneered event-based metadata models for integrating digital transactions across content, author, creator, library, publisher, and rights communities. The DOI system was formalized as ISO 26324 and is governed by the International DOI Foundation, a non-profit organization.
The kernel itself serves two purposes articulated in the DOI Handbook: recognition (making it possible to understand what kind of thing a DOI identifies and to match it with reasonable accuracy to a particular entity) and interoperability (ensuring metadata from different Registration Agencies can be combined or queried by the same software without requiring semantic mapping or transformation).
Purpose & Scope
The DOI Kernel answers fundamental questions about any DOI-identified entity. Its 17 elements are divided into two groups.
Descriptive Elements (Table 4.1 of DOI Handbook)
| Element | Occurs | Description |
|---|---|---|
| DOI name | 1 | The specific DOI name allocated to the referent |
| referentIdentifier(s) | 0-n | Other identifiers for the same referent (ISBN, ISSN, ISRC, ISAN, ISNI, etc.) |
| referentName(s) | 0-n | Names or titles by which the referent is usually known |
| primaryReferentType | 1 | Whether the referent is a creation, party, or event (open list) |
| structuralType | 1 | For creations: physical, digital, performance, or abstraction; for parties: person, animal, or organization (closed lists) |
| mode | 0-n | Principal sensory modes of perception: audio, visual, tangible, olfactory, tasteable, none (closed list) |
| character | 0-n | Fundamental communication form: music, language, image, other (closed list) |
| referentType | 0-n | More specific type classifications (open list) |
| linkedCreation | 0-n | Associations with other creations, including role relationships |
| linkedParty | 0-n | Associations with other parties |
| principalAgent | 0-n | Entities principally responsible for creation or publication |
| dateOfBirthOrFormation | 0-1 | For parties only |
| dateOfDeathOrDissolution | 0-1 | For parties only |
| associatedTerritory | 0-n | For parties only, using ISO 3166a2 codes |
Administrative Elements (Table 4.2)
| Element | Occurs | Description |
|---|---|---|
| registrationAuthorityCode | 1 | Code identifying the issuing Registration Agency |
| issueDate | 1 | Date when the DOI name was issued |
| issueNumber | 0-1 | Version designation of the kernel declaration |
Controlled Vocabularies and Allowed Value Sets
The kernel uses both closed and open controlled vocabularies. Closed lists (structuralType, mode, character, language via ISO 639-2, territory via ISO 3166a2) cannot be modified by Registration Agencies. Open lists (primaryReferentType, agentRole, creationType, creationIdentifierType, and others) can be extended by any RA through registration in the DOI Data Dictionary.
Version 2.2 expanded several open lists at the request of the EIDR Registration Agency, adding creation types (Film, TVProgramme, WebResource, Season, Clip, and others), new creation-to-creation link roles (Episode, Season, Edit, Performance, Derivation, and others), and additional creation identifier types (Ad-ID, SMPTE-UMID, UUID, URI, and others). It also introduced non-ISO territory codes for defunct territories (DDR, USSR, Yugoslavia) and restructured creation identifiers to support both resolvable URIs and non-resolvable values.
The allowed value sets are maintained in a separate namespace at http://doi.org/10.1000/282, introduced in version 2.2 to make them available to schemas beyond the kernel.
Serialization & Technical Format
The DOI Kernel is expressed as an XML Schema (XSD), maintained by the IDF at http://doi.org/10.1000/276. Registration Agencies may implement the kernel either by producing declarations directly against the kernel XSD or by incorporating kernel elements into their own wider metadata schemas. The DOI Data Dictionary, implemented within the Vocabulary Mapping Framework (VMF), provides the ontological underpinning and supports mapping between kernel terms and those of other metadata standards (ONIX, DDEX, MARC21, Dublin Core, RDA, CIDOC-CRM, FRBR, LOM).
Governance & Maintenance
The IDF governs the kernel through the DOI-RATech group, which coordinates schema updates. Any member of the group may propose changes; routine updates are typically implemented within two weeks. The update procedure involves proposal, consensus review, draft implementation, review period, and release. The DOI Foundation site content is licensed under CC BY 4.0.
ISO 26324 mandates that the data dictionary serve as the repository for all data elements and allowed values, and that all allowed values used in kernel metadata be registered there to facilitate integration of DOI data from different sources.
Notable Implementations
Every DOI ever issued carries kernel metadata. The most prominent Registration Agencies implementing the kernel include:
- Crossref -- scholarly articles, books, conference proceedings
- DataCite -- research datasets and other research outputs
- mEDRA -- European publications
- EIDR -- entertainment industry identifiers for audiovisual works (the primary driver of version 2.2 changes)
- KISTI -- Korean scientific and technical information
As of the mid-2020s, more than 300 million DOIs have been registered, each with an associated kernel metadata declaration.
Related Standards
- indecs Framework -- the semantic interoperability model underlying the DOI Data Model
- ISO 26324 -- the ISO standard for the DOI system as a whole
- Vocabulary Mapping Framework (VMF) -- mapping tool supporting semantic interoperability across DOI and non-DOI metadata standards
- Crossref Metadata Schema -- an RA-specific extension building on DOI metadata
- DataCite Metadata Schema -- another major RA-level schema built on the DOI foundation