Why this question matters
A catalogue may identify a photograph while leaving no record of the checks or transformations applied to its file. The question is easier to resolve when the institution separates the technical action from the archival decision it is meant to support. Preservation works only when the technology is paired with custody, documentation and a process someone is responsible for maintaining. Copies and checks are meaningful because an institution knows which objects they protect and what decisions to make when something changes.
Descriptive metadata helps discovery; preservation metadata records the objects, events, agents and rights needed for long-term stewardship. A useful response therefore needs both a documented process and a way to test whether the chosen action achieved the intended result. The specific checks differ by institution and type of material, but they should always be intelligible to another member of staff.
A practical approach for archival teams
- Establish a reliable baseline. Record the original object and its relationship to derivatives.
- Test the controlled process. Capture meaningful events such as ingest, fixity verification and migration.
- Keep the evidence for the next review. Retain responsible agents and rights information appropriate to the preservation action.
These are suggested local steps, not a quoted requirement from a particular standard. Define who owns each decision, what material can be included, and what happens when the expected outcome is not achieved. A short review with a realistic sample is more revealing than a demonstration with only uncomplicated records.
What evidence should be saved?
| Decision point | Evidence that would make the decision reviewable |
|---|---|
| Baseline and context | Inventory and custodial identifier. Include the object or accession identifier. |
| Review or verification | Verification results and change log. Record the outcome and any unresolved exception. |
| Ongoing stewardship | Recovery or rendering test. Keep enough documentation for a new custodian. |
A closer look at the trade-off
A preservation event is useful only if it says what happened to which object and with what outcome. If a file was migrated, preserve the old and new identifiers, date, method, responsible agent and relevant validation result. A catalogue title alone cannot answer why a derivative exists or whether a failed conversion was rolled back.
Worked example
The following example is hypothetical. After an image migration, a curator needs to establish whether the original survives and who approved the change. Rather than assuming that a technically successful action completes the project, the archive could establish the original context, document the decision and test whether another person can retrieve and explain the result. The institution should turn that experience into a test case. Future staff can then check the same failure mode instead of relying on informal memory.
The strongest improvement would be a procedure that handles the same problem next time without relying on one person's memory. The institution should identify which evidence is authoritative, which material remains uncertain and who can resolve an exception.
Common mistake to avoid
Do not treat a PREMIS-based model as proof that every preservation risk is covered. A future colleague may not recognise today’s software, storage service or filename conventions. A useful record includes the decision, its evidence, the object identifier and a person or role that can interpret it. Clarity for a new custodian is a practical measure of sustainability. This matters when cataloguing, preservation, technology and rights responsibilities are distributed across different people. A record is easier to trust when readers and staff can tell the difference between a verified statement, a provisional inference and a process that has not yet been tested.
Questions to ask before adoption
Can the archive explain what happened to a file between transfer and access?
Capture meaningful events such as ingest, fixity verification and migration. Retain responsible agents and rights information appropriate to the preservation action. If the evidence is incomplete, describe the limit rather than turning an assumption into a published claim.
How would the institution demonstrate a successful outcome?
The team should be able to identify a representative object, reproduce the relevant action, inspect verification results and change log and explain the implications to another qualified colleague. If that cannot be done without the original operator, the workflow needs further documentation or a different control.
Professional guidance and related reading
The Library of Congress, PREMIS Data Dictionary for Preservation Metadata provides a useful professional starting point. PREMIS 3.0 describes Objects, Events, Rights and Agents as core entities for preservation metadata. Use the guidance to inform local policy, but check the current source rather than treating this article as a substitute for an archival standard.
For an example of a platform approach related to this topic, review the standards and interoperability overview. This is an Archively product resource, not an independent authority. Evaluate its relevance with your own collections, requirements and staff responsibilities.
What to do next
Test the process in an environment that permits correction. Record what survives export and what must be rebuilt before expanding the programme. Start with the decisions and evidence above, then record the limits of the pilot and revise the workflow accordingly. The objective is not to accumulate more technical activity: it is to make the records more understandable, accountable and usable over time.
Top comments (0)