Identifiers:
Identifier is a "name" which uniquely identifies one particular object - this may be a complex object - within a defined context. An object may have many identifiers in different contexts.
Examples of these in a repository context are:
- Filename
- Author - http://names.mimas.ac.uk/
- Email address - LSE use it as a hidden identifier to disambiguate authors internally
- Handle - a persistent identifier for a digital object regardless of where it is stored
- DOI - http://www.doi.org/
- URI - http://en.wikipedia.org/wiki/Uniform_Resource_Identifier
- Accession / Repository Number - Unique to the repository itself.
- ISBN - http://en.wikipedia.org/wiki/International_Standard_Book_Number
It is important that as many identifiers that are known for an object are recorded in a repository record and metadata. There are also opportunities to record such information within the object itself, such as on a coversheet.
Wherever there are identifiers assigned which are intended to be permanent, such as a fixed url, these should be advertised for citation purposes to guarantee that the object can be found using that identifier.
The inclusion of all an object's related identifiers in metadata potentially permits chains of identifiers to be built, an approach being investigated by the JISC-funded RIDIR project. If, for instance, a repository user discovers a useful video clip and wants to trace the complete video, the inclusion in the clip's metadata of the original video's identifier - even if no other information - gives the user a starting point for tracking it down.
RIDIR aims to demonstrate the potential usefulness of recording the relationship(s) that the user might establish so that a network of relationships can be built up for others to use. On a future occasion another user might discover the same clip but, through the identifier relationships, immediately be able to see, for instance, where the clip originated, further derivatives of it and, maybe, related clips or videos.
Pros
- Obvious, simple and effective.
- Able to use identifiers to cross-link to the same object in another repository.
- Applicable to all object types.
Cons
- Need to be clear about the identifier and the provenance.
- Displaying all identifiers may lead to dead links over time as not all identifiers are persistent.