If you've spent any time online, you've hit a 404 page. And there's a decent chance someone has told you, with total confidence, where the number comes from: supposedly, at CERN, the birthplace of the web, there was a physical Room 404 that housed the web's first server — and when a file wasn't found on it, the system would report that "room 404" couldn't locate it. Some versions of the story claim it was Tim Berners-Lee's office, and that colleagues trying to find him would be told he wasn't in.
It's a satisfying story. It's also false, and it has been publicly and repeatedly debunked by the people who were actually there.
Why the room story doesn't hold up
Robert Cailliau, who worked alongside Berners-Lee at CERN and co-developed the early web with him, has said flatly that there was no Room 404 involved in the web's creation. CERN's own building numbering at the relevant site didn't even include a room by that number in the building where the first web server ran — accounts of the actual room layout put the first numbered office on that floor considerably higher than 404. Cailliau has pushed back on the myth in interviews for years, to little effect; it keeps circulating because it's a better story than the truth, and because "a mysterious missing room" is more fun to repeat than "a numbering convention."
Where 404 actually comes from
The real explanation is simpler, if less cinematic. HTTP, the protocol that governs how web browsers and servers talk to each other, organizes its status codes into ranges by category. Codes starting with 2 mean success. Codes starting with 3 mean redirection — the page moved somewhere else. Codes starting with 4 mean a client-side error — something wrong with the request itself, as opposed to the server. Within the 4xx range, specific numbers were assigned to specific problems as the protocol was formalized in the early-to-mid 1990s: 400 for a bad request, 401 for unauthorized access, 403 for forbidden access, and 404 for a resource that simply doesn't exist at the address requested.
In other words, 404 isn't a haunted room number — it's a slot in a systematic numbering scheme, the same way 401 and 403 are, chosen for logical organization rather than any physical location. The people building HTTP needed a large, extensible set of codes that servers and browsers everywhere could agree on, and they built one methodically. It's the same reason your oven has numbered settings instead of named ones: it scales better.
Why the myth won't die
Part of the reason the CERN room story endures is that the truth is almost too clean. "A committee assigned sequential numbers to categories of network errors" doesn't spread the way "there was a secret room and the web remembers it" does. The myth gives a impersonal technical standard a face and a place, which is exactly the kind of thing internet folklore loves to do — see also: the equally persistent, equally wrong idea that "spam" was named after a can that fell off a shelf, or that the @ symbol was chosen for some deep symbolic reason rather than simply being an unused character on a keyboard.
The next time someone tells you about Room 404, you now have the actual story ready: no missing room, no locked office — just an orderly protocol doing exactly what it was designed to do, decades before anyone thought a 404 page would need its own personality, jokes, or dancing GIFs.