2. Persistent Identifier Syntax

NRS persistent identifiers have three colon-separated components: a namespace identifier (NID), which is always the string "urn-3"; a naming authority; and a resource name:

urn-3:naming-authority:resource-name

example: urn-3:rad.arch:15009

What is a namespace?

Since many different naming services exist today, the identifiers created under each service are grouped into different namespaces, or sets of syntactic rules defined for each service and specified by namespace identifiers (NID). The namespace to which a particular persistent identifier belongs is specified by the required NID component of the identifier. Whenever you see a persistent identifier with the NID prefix "urn-3":

   urn-3:naming-authority:resource-name

you know that identifier belongs to Harvard's Name Resolution Service. Harvard's NID "urn-3" has been registered with the appropriate Internet standards body to identify NRS persistent identifiers uniquely and globally.

What is a naming-authority?

The naming authority indicates under whose auspices the identifier was created. For example, all persistent identifiers created for the Radcliffe Institute Archives will be of the form:

   urn-3:rad.arch:resource-name

An naming-authority can be composed of alphanumeric characters [A-Za-z0-9]. 

Note that although both upper and lower-case characters are allowed in forming a naming-authority, all persistent identifier comparisons in NRS are case-independent. The naming authorities RAD.ARCH and rad.arch represent the same naming authority.

What is a resource name?

The resource-name portion of a persistent identifier is a string composed of alphanumeric characters [A-Za-z0-9] that uniquely identifies the named resource. The resource name might be an existing identifier, such as catalog record number or accession number, or a unique random string of characters. For example, in the following persistent identifier:

   urn-3:RAD.ARCH:15009

"15009" is the resource name.

Note that a resource-name has to be unique only relative to the authority path. So the resource name "15009" can be used more than once, as long as each instance is part of a different authority path. For example, these persistent identifiers are both valid:

•    urn-3:RAD.ARCH:15009
•    urn-3:HUL.OIS:15009

Note that persistent identifiers are not a substitute for, or a supplement of descriptive metadata about named resources. Such semantically rich descriptive information should be restricted to its proper place in resource catalogs, finding aids, or other discovery services, and should not be encoded into resource names.

For example, the identifier:

   urn-3:HUAM:Burne-Jones,Edward:The_Days_of_Creation:the_First_Day(1870-76)

would be a poor choice for identifier of a digital representation of Edward Burne-Jones painting, The Days of Creation: the First Day (1870-1876), in the Fogg Art Museum. At some time in the future, the work might be re-attributed to a different artist, or subsequent research might uncover a more suitable title or creation date. A good choice for an identifier would use the museum's accession number:

   urn-3:HUAM:1943454

which identifies the work in a semantically neutral manner.