A 1945 essay proposed linking personal records by association

The setup

Vannevar Bush wrote 'As We May Think' as director of the U.S. Office of Scientific Research and Development, the wartime post from which he had coordinated roughly six thousand scientists. The essay ran in The Atlantic's July 1945 issue, addressed to scientists about to be released from war work. Bush's brief was to redirect that talent toward making the accumulated store of research usable again, since existing filing and indexing methods, built for a much smaller volume of printed material, were already failing under wartime and prewar publication.

What the documents show

The essay proposes, but does not build, a desk-sized device Bush names the memex: a mechanized private file and library storing books, records and correspondence on microfilm, operated with levers and a keyboard. The MIT-hosted text of the essay describes its central mechanism as associative indexing: any stored item can be linked to any other, so a reader builds a named trail through material the way memory jumps between associated ideas rather than filing by rigid category. Bush illustrates this with a hypothetical researcher tracing why the Turkish bow outperformed the English longbow, linking encyclopedia entries, history texts and his own inserted comments into one traceable path, then handing a copy of that trail to a colleague. Neither copy of the essay describes a working prototype; every mechanism is speculative engineering, extrapolated from 1945 microfilm and photographic technology.

The friction

The essay is explicit about what it does not solve. Building a trail is manual, one link typed in at a time. Sharing a trail means physically photographing and passing along a film copy to a single friend's own memex, not publishing to any wider audience. And the apparatus depends on hardware Bush admits does not yet exist in the form he needs. Read as a specification rather than a manifesto, the memex is closer to a research proposal with unresolved engineering than a described product.

What changed in the work

Modern linked-notes and personal-knowledge tools regularly cite this essay for their backlink features, and the documents support that lineage at the level of the idea: associative trails between discrete notes, rather than folders, is exactly what Bush describes. What the essay does not support is any claim that Bush anticipated computing or search; the memex is optical and mechanical, not digital. That distinction is editorial but load-bearing: crediting a 1945 essay with today's specific interface choices overstates a proposal that was never built into a working lineage for the software itself.

• Does the tool being credited to this essay actually implement associative trails, or just tags and folders relabeled as "linked notes"?
• What happens to a trail's readability once its author is no longer available to explain the connections?
• Is the claimed lineage to Bush doing marketing work that the feature's actual design does not support?

The value of reading the essay directly, rather than the claims made about it, is that it separates a genuinely old idea, associative linking, from a genuinely new implementation. Bush described a problem and a plausible mechanism; he did not solve either.