Between 1960 and 1983 the metre was defined as being equal to 1 650 763.73 wavelengths of an orange emission line of krypton-86. How did they actually use that definition? Was it possible to move the mirror of an interferometer slowly over a distance of 1 metre, perhaps by a very smooth screw, and count the appearance of 1.6 million fringes without mistakes? Was the coherence length of the krypton light more than 1 metre?
There are much easier ways to measure a wavelength than by "tromboning" an interferometer. A simple two-slit grating, for example, will give you a pretty decent measurement. You don't need to measure a "meter's worth" of light, just run an experiment where the output data (such as a diffraction pattern) can be used to calculate the wavelength to high precision and accuracy.
In any case, the definition was basically "in reverse." Using existing measurement techniques, NIST or ISO or whomever took the current measurement of this narrow (no broadening, no fine-structure splitting) line and used that to define the meter.
It was actually done with a "tromboning" Michelson Interferometer
In 1892 by Michelson. He measured a distance of 10cm to an accuracy of ± 0.00004 mm using a Cadmium lamp . The 1960 re-definition with Krypton used the same data - it was just a matter of measuring the ratio of the wavelengths of the Cadmium and Krypton lines
Michelson's 1907 Noble prize speech describes it http://qsd.magnet.fsu.edu/course/exp3802/optics/AAMichelson1907.pdf