People confuse archives and libraries constantly, which is understandable, since both involve shelves and quiet rooms. But an archivist would tell you the two disciplines solve opposite problems. A library organises finished, published objects by subject. An archive preserves unpublished, often unfinished material by origin — keeping the papers of one office, one family, one committee together, in the order they were created, even when that order looks arbitrary to an outsider.
Provenance over subject
This principle, called respect des fonds, exists because the context of creation is itself evidence. Scramble a collection of letters into alphabetical order by subject and you lose the ability to answer who knew what, when. Information architects rediscover this constantly when they normalise a database into subject-clean tables and then can't reconstruct the sequence of events that produced the data in the first place.
The finding aid as schema
Archives don't index every item; that would be impossible at scale. Instead they produce a finding aid — a hierarchical description of a collection's structure, series by series, box by box. It is, in effect, a schema written for humans before schemas were written for machines. Any product team that has tried to document "where does this data live and why" has written a finding aid without calling it one.
An archive keeps the mess on purpose. The mess is the record of how the thing was made.
The lesson for digital systems isn't to become messier for its own sake. It's to recognise that some disorder is evidence, and normalising it away costs you information you didn't know you were storing.