Workbook on Digital Private Papers > Administrative and preservation metadata > Preservation metadata
Preservation metadata
Assessing the use of PREMIS semantic units and PREMIS XML schemas in the preservation of personal archives
The <agent> entity
The purpose of the agent entity is to supply information about the agents (persons, organisations, or software) associated with preservation events or with the management of rights related to one or more objects in the repository. The provision of a unique identifier serves to distinguish between agents and can be used to link an agent with one or more events. Groups of agent types, which can be allocated rights to authorise or undertake certain events, may be created.
Recording agent metadata linked to event metadata is useful in the following circumstances:
- Where a change is made to a object, we might want to record the human and/or software agent responsible.
- Where an item is removed from a repository, we might want to record the details of the responsible agent.
If the agent metadata is to be used for troubleshooting, it may be necessary to record information about agents in some detail, for example the environment details for virus checking software may impact on the success of the virus check, and if the check fails to turn up information about a subsequently discovered virus, it may be helpful to have details of the virus checker and its environment available.
At present, the agent entity is very simple consisting of very few semantic units. Karen Coyle has suggested changes to the entity, for better recording of rights-related information.
3.1 <agentIdentifier> Used to record the type and value of the agent identifier.
3.2 <agentName> Used to record a reader-friendly descriptor of the agent.
3.3 <agentType> Used to record a value from a controlled vocabulary to denote the type of agent; PREMIS suggests person, organization and software. It might be useful to have more granularity than this, e.g. software - virus checker.
