Everyone knows how the Library of Alexandria ended, or thinks they do — the story of its destruction has more competing versions than the library itself had scrolls. Fewer people know what it actually did while it stood: it ran, as a matter of state policy, a programme to acquire, translate, and copy every text its agents could locate, in every language they could find someone to read.
Completeness as a design goal
This was new. Earlier libraries, mostly temple or palace archives, collected what was locally relevant. Alexandria's founders under Ptolemy I set out to collect everything, on the theory that a text's usefulness couldn't be judged in advance. That premise — collect first, evaluate relevance later — is the same premise behind every modern crawl-and-index search engine, five hundred generations removed.
The cost nobody budgeted for
What Alexandria didn't solve was preservation at scale. Papyrus decays, copies introduce errors, and a single-site collection is a single point of failure — whatever combination of fire, funding cuts, and neglect actually ended it, the deeper lesson is that comprehensiveness without redundancy is fragile by construction. Modern digital archives take this seriously in a way ancient ones structurally couldn't: distributed backups, format migration, checksums against silent corruption.
Alexandria's real failure wasn't the fire. It was believing one building could be the whole of human memory.
Every institution attempting a "complete" digital archive today is quietly answering a two-thousand-year-old design problem.