Why Digital Cultural Records Need More Than Backups
Backups help recover data but do not ensure long-term access. Cultural institutions also need to check file integrity, formats and rendering dependencies, rights and context, and who is responsible for review.
Publication & corrections: infostoa · Editorial standards
Individual author and specialist reviewer information is not provided for this article.
Keeping several copies of a digital file lowers the risk that one device failure will destroy it. Yet over time, a file can become damaged, its format can be forgotten, or the software and rights information needed to use it can disappear. Having copies is different from being able to understand the content in the future.
Backups and preservation answer different questions
A backup is a copy used to recover data after an error or system failure. Digital preservation is a continuing effort to keep material understandable and usable while retaining the characteristics that matter. For a cultural collection, those characteristics may include the image, sound, or video itself as well as its description, creator, date, rights, and relationship to other records.
Several copies can still share the same risk. They may all have been made from a damaged original, depend on one account or administrator, or sit in locations exposed to the same incident. Checking that a file opens today can miss the case where the wrong or incomplete file was copied. A preservation plan therefore needs integrity checks, documentation, named responsibility, and a way to recover from a verified copy.
Integrity checks reveal unexpected change
A checksum is a value calculated from a file’s bytes. An institution can record one when it receives or creates a file and compare it with later values to detect accidental changes or corruption. Regular checks, documented results, and a response to mismatches turn the value into a working process rather than an isolated number.
When files are moved or converted, keeping before-and-after checksums with a record of the action makes the timing and scope of validation easier to trace.
Matching checksums do not prove that a file was correct when first collected. They do not explain the cultural significance of the content or whether color, audio, captions, and other important properties will remain interpretable. Link integrity records to descriptive and technical metadata, and decide which trusted copy will be used if a check finds a difference.
File formats depend on future tools
The Library of Congress discusses factors that affect whether a format can remain usable: how fully it is documented, how widely it is adopted, whether it is transparent and self-describing, its external dependencies, and whether patents or technical protections constrain use. An openly documented and widely used format may be easier to validate or render in the future, but a format label alone does not settle the preservation question.
Significant properties depend on the material. Color and resolution may matter for a still image; time, sound, and captions for a video; sequence and performance context for an event recording. An institution may retain a source file and create an access copy, but conversion should not silently replace the source. Record what changed, what was intended to remain, and which version supports which use.
Expert perspective: NDSA levels as an institutional review framework
The National Digital Stewardship Alliance’s Levels of Digital Preservation is a staged framework institutions can use to examine practice and identify areas for improvement. It encourages a broader view than purchasing a storage tool: staff can ask how files are stored and checked, who controls them, what metadata is kept, and how content is managed. An organization can start with practices it can sustain and plan improvements around its collections, staffing, and risks. For a review to lead to action, record the practice checked and who owns the next improvement, not only the level label.
Test the plan with one important record
Choose a record and write down where the source and copies are kept, when integrity was last checked, the file format, the software needed to open it, its description and rights, and who owns the next review. Then ask a colleague to locate and interpret the file using only those notes. The exercise often exposes gaps more clearly than another general policy document.
If the checksum matches but no one can find a reader or explain the rights and context, another backup does not address the problem. The next step may be documenting software dependencies, adding missing descriptive information, or transferring responsibility when staff roles change. Rather than trying to perfect every file at once, set priorities according to cultural value and potential loss, and use each check to plan the next one.