OMI-SPEC-003: Anchor Model
Status
Draft
Version: 0.1
Purpose
The Anchor Model defines how semantic objects are attached to scholarly content.
Unlike traditional document models that reference page numbers or character positions, OMI introduces persistent semantic anchors.
Anchors provide stable references that survive editing, publishing, and format conversion.
They form the connective layer of the Open Manuscript Initiative.
Design Principles
The Anchor Model is based on the following principles:
- Semantic persistence
- Layout independence
- Stable identity
- Multiple target support
- Version awareness
- Machine readability
- Long-term preservation
What Is an Anchor?
An Anchor is a persistent semantic reference that identifies one or more targets within a manuscript.
An Anchor does not contain content itself.
Instead, it connects semantic objects to meaningful parts of a manuscript.
Why Anchors?
Traditional publishing relies on page numbers and character positions.
These references break whenever text changes.
OMI Anchors remain attached to the semantic content rather than its visual position.
This enables reliable collaboration throughout the manuscript lifecycle.
Anchor Targets
An Anchor may refer to:
- a character range
- a word
- multiple words
- a sentence
- a paragraph
- multiple paragraphs
- a section
- a figure
- a table
- a formula
- a citation
- a bibliography entry
- metadata
- the entire manuscript
Multiple Targets
One Anchor may reference multiple locations.
Example:
Anchor A
├── Paragraph 2
├── Figure 4
└── Table 1
This allows a single semantic object to describe relationships spanning different parts of the manuscript.
Semantic Relationships
Anchors connect scholarly objects.
Examples include:
Annotation ↓
Anchor
↓
Text Range
or
Citation
↓
Anchor
↓
Bibliography Entry
or
Translation
↓
Anchor
↓
Original Paragraph
Objects Using Anchors
Nearly every semantic object within OMI may reference one or more Anchors.
Examples include:
- annotations
- citations
- figures
- tables
- glossary entries
- translations
- reviewer comments
- editorial notes
- AI suggestions
- tasks
- cross references
- hyperlinks
Anchor Metadata
Each Anchor contains:
- Identifier
- Type
- Target
- Creator
- Creation Timestamp
- Modification History
- Status
Stable Identifiers
Every Anchor receives a globally unique identifier.
Example:
anchor:3d91f88f-5d61-4d7e-a857-bf5b44d9f8a1
Persistent identifiers ensure reliable references across versions.
Editing Behaviour
Anchors should survive ordinary editing operations.
Examples include:
- inserting text
- deleting text
- moving paragraphs
- splitting sections
- merging sections
The system continuously maintains Anchor integrity.
Anchor Resolution
If the original target changes substantially, the system attempts semantic reattachment.
Possible strategies include:
- exact match
- structural match
- contextual match
- semantic similarity
Implementations may combine multiple strategies.
Visibility
Anchors themselves are invisible.
Users interact with the semantic objects attached to them.
For example:
Footnote
↓
Anchor
↓
Selected Text
Rendering
Different publication formats render Anchor-based objects differently.
Example:
↓
Footnote
HTML
↓
Tooltip
EPUB
↓
Endnote
JATS XML
↓
Semantic element
The Anchor itself is never rendered.
Only the attached semantic object is displayed.
Versioning
Anchors participate in document version control.
The same Anchor identifier should remain valid across manuscript revisions whenever possible.
This enables reliable comparison between versions.
Interoperability
Future mappings may include:
- Web Annotation Data Model
- W3C Selectors
- JATS XML
- HTML Fragment Identifiers
- DOCX Comments
- PDF Annotations
Plugin Extensions
Plugins may define additional Anchor types.
Examples:
History Profile
- Archival Document
- Charter Fragment
Mathematics Profile
- Proof Step
- Equation Component
Biology Profile
- Gene Sequence
- Taxonomic Node
Future Work
Future specifications will define:
- Anchor Resolution Algorithms
- Cross-document Anchors
- Persistent Global Anchors
- Linked Data Integration
- Collaborative Conflict Resolution
Summary
The Anchor Model is the semantic foundation of OMI.
Rather than relying on page numbers or visual positions, Anchors connect scholarly objects directly to meaning.
This allows annotations, citations, peer review, AI assistance, translations, and future scholarly services to remain stable across editing, publishing, and preservation.
Anchors transform a manuscript from a formatted document into an interconnected knowledge object.