A Brief History of Timekeeping

The oldest timekeeping devices didn't measure time so much as track the sun's position. A sundial's shadow is really a proxy for the Earth's rotation, which made it useless at night and unreliable on a cloudy day. Water clocks and hourglasses solved the "no sun" problem by measuring a steady flow instead of a shadow, and they're old enough that versions of both show up in ancient Egyptian and Chinese records, independently invented.

Mechanical clocks, starting around the 14th century in Europe, were the first devices to divide a day into genuinely equal hours rather than the "daylight hours" that stretched and shrank with the seasons. That single change — equal hours, all year — is easy to take for granted now, but it's a prerequisite for almost everything that came after, including train schedules and, eventually, computing.

Why clocks needed to agree with each other

For most of history, "noon" was a local fact — it was whenever the sun was directly overhead in your specific town, which meant every town's clock disagreed slightly with its neighbor's. This was a minor inconvenience until railways made it possible to travel faster than the time difference could quietly correct itself, and train schedules across even a single country became a genuine mess. The result was the standardized timezone system, first adopted by railways in the 1840s-80s and then formalized internationally, which is the direct ancestor of the IANA timezone database your browser uses today.

Coordinated Universal Time and the atomic clock

Mechanical and even early electric clocks drift. Atomic clocks, developed in the mid-20th century, measure time using the extremely consistent vibration of atoms (cesium, specifically, for the modern definition of a second), and they're accurate enough that they revealed the Earth's rotation itself is slightly irregular. That mismatch between atomic time and the Earth's actual rotation is the reason leap seconds exist at all — a patch to keep clock time and sunrise time from drifting apart.

Why Unix time counts from 1970

By the time Unix was being developed in the early 1970s, engineers needed a simple, storage-efficient way for computers to represent a moment in time — and counting seconds from a fixed, recent starting point turned out to be far simpler than storing a full calendar date in every timestamp. January 1, 1970 wasn't chosen for any deep reason; it was just a convenient, round, recent date at the time the standard was being written. It stuck purely because so much software was built on top of it afterward.

That single number — seconds since 1970 — is still how almost every computer system represents a moment internally, even though the humans reading the result see a calendar date. If you want to see it in action, the timestamp converter on the home page shows the conversion both ways.