A guide to registry services

Meaning of a registry entry

An entry in a Registry means that the registrar asserts the existence of an entity. Every entry in a registry must therefore be accompanied by at least one attestation. For registries intended to be used with Reference Indexing services, the attestation should be expressed in the CTS URN format. We maintain a Registry of CTS Namespaces identifying URN namespace prefixes for registered CTS servers; in that case, attestations should be expressed as URIs.

Data structure

The main data contents of an identifier entry in a registry looks like:

<identifier @id= ,@status=, @target=?>
  <description>human readable description or label</description>
  <identifier/>*
</identifier>

The id attribute is a value unique within its hierarchical level. In the registry of Greek literary works, for exmample, the text group Homer might be identified with the unique value tlg0012; the Iliad might be contained within that text group, and identified as work tlg001. Other works in the registry could also be identified as tlg001 within separate workgroups but only one identifier in each workgroup could have that identifier.

The status has one of the following values:

• validated: this identifies an editorially approved entry guaranteed never to be removed from the service
• provisional: a value nominated and expected to be added but not yet validated
• experimental: a temporary value not expected to be added
• redirected: marks an identifier that was erroneously validated, and therefore cannot be removed from the service; while the object it names exists in the registry, this identifier is incorrect. This must be accompanied by a target attribute giving the correct identifier
• invalid : marks an identifier that was erroneously validated, and therefore cannot be removed from the service, but no such object exists in the registry.

Methods

Three methods are available to use a registry:

1. getID(id value, validated (default true)?, provisional (default true)?, experimental (default false)?, redirected (default true)?, invalid (default true)? , separator?):
   This method searches the name service for the value given in the ID parameter. It excludes or includes entries of any status named in optional boolean parameters: the default is to include all but epxerimental entries. The optional separator parameter is a single character used to defin segmentation of multi-tier values. Its default value is a period ('.') so that a query for textgroup Homer and work Iliad as mentioned above would be expressed as tlg0012.tlg001.
   This method returns either an identifier structure or an error message if no entry has the ID value.
2. getIndex(validated (default true)?, provisional (default true)?, experimental (default false)?, redirected (default true)?, invalid (default true)?)
   Retrieves a copy of the entire registry. Optionally, entries of a particular status may be excluded or explicitly included following the same syntax as for the getID() method.
3. getRedirects(id value, separator?)
   This method searches the name service for identifiers redirected to the requested ID value. It allows an optional separator parameter with the same meaning as in getID().

Metadata

The Registry as a whole includes metadata identifying the registrars, and identifying a unique namespace prefix for the registry.

Individual records include an editorial history indicating names of submitter and date of initial submission, and names of editor and dates of any changes to status or to text description.

Links

• Formal specifications: the structure of Registry data is defined by a Relax NG schema; the current draft of the schema is available from the project version control system
• Implementations: for an implementation as a web application using the XML:DB API and the eXist XML database, see the project version control system