It is often the case that a group of resources already have a unique identifier. This might be a local system identifier derived from, e.g. a database key; or a global non-URI identifier, e.g. an ISBN.

Mint URIs that are algorithmically derived from the existing non-URI identifier. This can be as simple as concatenating the existing identifier or key with a suitable base URI. The existing identifier may need to be URL encoded before creating the URI. It is common to combine this technique with Patterned URIs.

The BBC programmes website uses URIs that are derived from its existing "programme id" or pid.

Where resources are already associated with existing keys, it is likely that the creation and management of those identifiers will already be supported by a specific technology or process. There is a need to be able to create global URI based identifiers for these resources without creating unnecessary additional overheads in creating entirely new identifiers and/or mapping between URIs and existing keys.

By deriving the URI from the natural key for the identifier we avoid the need to create a new process for assigning identifiers and largely eliminate the need to have a mapping between the two identification systems.