The CITE Object URN specification, version 0.1

Content: This document defines version 0.1 of the Canonical Text Services URN.

Editors: Christopher Blackwell and Neel Smith, Center for Hellenic Studies Technical Working Group leads.

Date: April, 2015.

License: This specification is available under the terms of the Creative Commons Attribution-ShareAlike 4.0 International License,


RFC2141 defines the goal of URNs as follows:

Uniform Resource Names (URNs) are intended to serve as persistent, location-independent, resource identifiers.

The CITE Object URN (or more briefly, "CITE URN") scheme defined here closely follows the syntactic requirements of a URN, but extends its goal to provide a system of persistent, technology-independent identifiers for texts and passages of texts. Specifically, the semantics of the CITE Object URN scheme are recognized by implementations of the CITE Collection Services protocol, but CITE URNs can be used in any system, digital or other, that can parse the syntax of CITE URN values according to the syntax and semantics of this specification.

Semantics of a CITE Object citation

There are two major components to a citation with CITE URN version 0.1: a component identifying a naming authority, and a component uniquely identifying an object.

The naming authority component is a single identifer; systems using CITE URNs must be able to resolve this value to a unique URI.

The object component is a hierarchy identifying sets of unique objects in versioned representations. The sets may be ordered. Conceptually, the object hierarchy is similar to code management systems such as maven that organize versioned packages in groups. The object component begins with an identifier for a collection, which is a set of objects sharing a common set of properties. Collections can represent any kind of physical objects (e.g., paintings in a museum) or conceptual objects (e.g., episodes in a prose narrative). This identifier may be followed by an identifier for a unique notional object within that collection. This in turn may be followed with an identifier for a specific version of that object.

If the object component of the CITE URN is at the notional object or version level, reference to a single object may be extended with additional, type-specific citation of some aspect or component of the object.

Syntax of a CITE URN

Overall structure of a CITE URN

All URNs must begin with the string urn: followed by a protocol identifier (the "urn namespace identifier"), followed by a colon (":", Unicode x003a). The protocol identifier for the CITE Object URN must be cite. The rest of the URN consists of a CITE namespace, and an object component.

The namespace must be separated from the object component by a colon. A value for the CITE namespace and for the object component is required.

The top-level structure of a CITE URN could be summarized as


The object component

The work component consists of one to three parts. Each part must be separated from a following part by a full stop (".", Unicode x002e). A full stop must not be appended to the part value if there is no following part. The value for the first part of the work component is mandatory, and must be a valid value for a collection in the URN's CITE namespace. The second part of the work component is optional; its value must be a valid value for an object within the collection. The third part of the work component is optional; its value must be a valid value for a version of the parent object.

The structure of the work component of a CTS URN could be summarized as


Extended references

References to specific objects (either at the notional object level, or the version level) may be qualified by type-specific extended references.
The extended reference must be separated from the canonical reference by the "at" sign ("@", Unicode x0040). If URN includes an extended reference, the value of the extended reference must not be null.

The overall structure of a node reference could be summarized as


Ranges of objects in ordered collections

Ranges of two or more nodes are identified by an ordered pair of object references as previously defined. The value for the first reference must identify the first object in the range; the value for the second node reference must identify the second object in the range. The two node references must be separated by the hyphen-minus sign ("-", Unicode x002D).

The structure of a range reference could be summarized as


where each of OBJECT1 and OBJECT2 a valid reference to a citable object as previously defined, and the hyphen is literal.

Character set

CITE URNs must be expressed in the Unicode character set.

Reserved code points

Following the generic URN syntax, CITE URN reserves the use of the following code points: they must not be included in the data components of a CITE URN.

Excluded code points

Following the generic URN specification, CITE URNs exclude the following code points: