4. Link Relation Types

Link relation types identify the semantics of links between resources.

Purpose

Semantic Identification

• Defines the meaning of the relationship
• Example: "copyright" indicates target contains copyright terms
• Can indicate target resource attributes or behaviors
• Example: "service" implies a protocol interface

Not Media Types

Relation types are distinct from media types:

• Don't identify representation format
• Only describe how resources are related
• Media type determined by Content-Type header

Types of Relation Types

4.1. Registered Relation Types

Characteristics:

• Well-defined tokens registered with IANA
• Promote reuse across applications
• Case-insensitive comparison
• Must conform to reg-rel-type ABNF rule

Naming Guidelines:

• Appropriate specificity level
• Application-specific semantics → specific names
• General semantics → general names

Constraints:

• MUST NOT constrain context IRI media type
• MUST NOT constrain target IRI media types
• CAN specify target resource behaviors
• CAN specify allowable HTTP methods/media types

Application Data:

• Registry can include application-specific data
• Example: browser archiving behavior
• Registered via Link Relation Application Data registry

4.2. Extension Relation Types

Definition:

• URI that uniquely identifies the relation type
• For applications not using registered types
• URI may point to semantic definition

Usage Guidelines:

• Clients SHOULD NOT automatically access URI
• Avoids server burden
• Comparison: case-insensitive string match
• Lowercase URIs recommended

Serialization:

• Required to be absolute URIs in Link headers
• Must be quoted if containing semicolon (;) or comma (,)
• Other serializations may use different forms (e.g., CURIEs)

---

Registry: IANA Link Relation Types

---