OMI-SPEC-005: Citation Model
Status
Draft
Version: 0.1
Purpose
The Citation Model defines how scholarly references are represented within the Open Manuscript Initiative (OMI).
Rather than storing citations as formatted text, OMI models them as semantic objects connected to bibliographic metadata.
This separation enables flexible citation styles, machine processing, validation, and interoperability.
Design Principles
The Citation Model follows these principles:
- Semantic first
- Presentation independent
- Persistent
- Machine readable
- Interoperable
- Extensible
- Style independent
What Is a Citation?
A citation represents a semantic relationship between one scholarly object and another.
It is not a formatted string.
Instead, it references a bibliographic object through a persistent identifier.
Citation Architecture
Citation
│
├── Anchor
├── Bibliographic Record
├── Citation Style
└── Rendering
The Citation object remains independent from visual formatting.
Citation Components
Each citation contains:
- Identifier
- Citation Type
- Bibliographic Target
- Anchor
- Locator
- Prefix
- Suffix
- Notes
- Version
Citation Targets
A citation may reference:
- Journal Article
- Book
- Book Chapter
- Conference Paper
- Thesis
- Dissertation
- Dataset
- Software
- Preprint
- Standard
- Web Resource
- Archival Source
- Manuscript
- Image
- Map
- Audio
- Video
The model is extensible through plugins.
Persistent Identifiers
Whenever available, citations should use persistent identifiers.
Supported identifiers include:
- DOI
- ISBN
- ISSN
- ORCID
- ROR
- Handle
- ARK
- URL
A citation may contain multiple identifiers.
Bibliographic Records
Bibliographic metadata is stored independently from citations.
A single bibliographic record may be referenced by multiple Citation objects.
Bibliographic Record
↓
Citation A
Citation B
Citation C
This eliminates duplicated metadata.
Locators
Citations may include precise locators.
Examples:
- page
- page range
- chapter
- section
- paragraph
- figure
- table
- appendix
- line
- folio
Examples:
p. 24
pp. 24–31
Chapter 5
Figure 3
Citation Relationships
Citations may express different scholarly relationships.
Examples:
- supports
- refutes
- extends
- discusses
- compares
- reproduces
- translates
- critiques
Future versions may introduce explicit citation semantics.
Citation Styles
Citation styles are presentation profiles.
Examples include:
- APA
- Chicago
- MLA
- Harvard
- IEEE
- Vancouver
- OSCOLA
- Turabian
Style selection never changes the underlying Citation object.
Discipline Profiles
Different disciplines may introduce specialised citation types.
History
- Archival Source
- Charter
- Manuscript
- Regest
Law
- Case Law
- Legislation
- Treaty
Medicine
- Clinical Guideline
- Trial Registry
Computer Science
- Software Package
- Repository
- API Documentation
Citation Rendering
Rendering depends on the publication profile.
Example:
HTML
↓
Inline citation
↓
Footnote
EPUB
↓
Hyperlinked citation
JATS XML
↓
Structured XML
Validation
The system may validate:
- DOI format
- ISBN checksum
- ISSN checksum
- ORCID format
- duplicate references
- incomplete metadata
- broken URLs
Validation rules are configurable.
Interoperability
Future mappings include:
- CSL JSON
- BibTeX
- RIS
- Crossref
- DataCite
- JATS XML
- Dublin Core
- Zotero
- EndNote
- Mendeley
Versioning
Citations participate in document version control.
Changes to citation metadata are tracked independently from manuscript text.
Plugin Extensions
Plugins may introduce:
- citation types
- validation rules
- metadata fields
- rendering rules
without modifying the OMI Core.
Future Work
Future specifications will define:
- Citation Graph
- Linked Open Citations
- Citation Intent Model
- Authority Control
- Reference Resolver
Summary
The OMI Citation Model represents citations as semantic, reusable, and presentation-independent scholarly objects.
By separating citation semantics from rendering, OMI enables consistent citation management, automatic style generation, metadata validation, and interoperability across the scholarly publishing ecosystem.