Preservation
Two things we tend to conflate
A backup answers the question “can I get my files back if this drive dies tonight?” An archive answers “will this body of work still be readable, identifiable, and understandable in twenty years — including by someone who isn’t me?”
These aren’t two levels of rigour applied to the same task. They’re separate problems, and a flawless backup can leave the second one untouched: a drive holding 80,000 files named IMG_4471.CR2, with no metadata and no structure, is backed up and already lost.
What archiving adds
- Identification. Every image carries a stable identifier that survives being moved or renamed.
- Context. Who, what, where, when, in which project, under which rights. Without it, a photograph becomes a mute object.
- Verifiable integrity. A checksum computed on entry and re-checked periodically, so you can state that a file has not silently changed.
- Durable formats. Containers and encodings you’ll still be able to open without the original software.
- Documentation of the whole. An inventory readable independently of the tool that produced it.
Silent corruption, the risk specific to archives
Files don’t always fail loudly. A flipped bit on a disk, an interrupted copy, a filesystem that lies about a write: the file still opens, the preview still renders, and a region of the image is dead. Nothing warns you.
A backup faithfully copies that corruption. That is precisely its job. Only checksum verification can detect it — and tell you which copy is still sound.
What “long term” actually demands
Long term isn’t a duration, it’s a constraint: your archive has to outlive its tools. If understanding your holdings requires opening one specific application, in one specific version, on one specific operating system, then its lifespan is the lifespan of that chain — years, not decades.
Hence a simple rule: everything that matters must be exportable in the clear. Metadata in standard formats, an inventory in readable CSV or XML, an explicit folder structure.
In practice
- Separate what is in progress (backed up) from what is frozen (archived).
- Give every image a stable identifier, independent of its filename.
- Compute a checksum on entry and re-check it periodically.
- Confirm your inventory stays readable without your cataloging software.
- Document context while you still remember it, not ten years later.
Within the Obscura Flow suite, these points are shared between two apps: Obscura Vault takes the SHA-256 checksum of each original at import and re-checks it through fixity audits, and its archives stay readable without it; Obscura Catalog writes metadata as XMP and exports the whole library in open formats — so the archive doesn’t depend on the software that built it.