A length quoted without a temperature is not a length. Since 1931 the engineering world has agreed that a stated dimension means the size at twenty degrees, so a component gauged in a warm workshop has not been gauged against the drawing at all. That convention sits at the small end of a much larger arrangement.

Julian Ashworth traces the arrangement upward from the workshop bench. Steel gauge blocks adhere to one another so firmly that the contact itself becomes part of what has been measured. A helium-neon laser is held against an iodine absorption line and checked back to the second with a frequency comb. A tunnel is held within a fifth of a degree, because otherwise a bench of sixty-one metres would not be sixty-one metres for long. Two radio telescopes, oceans apart, recover their separation from the moments at which a quasar's signal reaches each of them.

At the top of the arrangement sits a decision rather than a measurement. In 1983 the speed of light was fixed by declaration at 299,792,458 metres per second, and distance became a consequence of time. This was not sleight of hand. Clocks had improved so far past any physical ruler that letting time carry length was the conservative move, not the radical one.

Two claims are worth retiring. The metre has not depended on the Earth's meridian for two hundred years. And 1960, not some date within living memory, is when the platinum bar outside Paris was retired as the definition.

Where a link checker could not reach a source, the entry appears without a link rather than as something confirmed.