P sure I talked about this a bit before, but it came back to my mind and I had to actually write it out.

A Better Temperature System https://www.xanthir.com/b58m0 

tldr: Freezing point of water is 20°T, absolute zero is -500°T. 0°T and 100°T are roughly as cold and as hot as temperate climates reach. Body temp is 90°T. Water boils at 210°T.

The system's 0/100 are close to Fahrenheit's, giving the same (superior) weather-based advantages. But freezing point is a round number, and converting from human-friendly to physics-friendly scale variants is just a ±500, rather than the awkward/bad ±273.16 that Celsius does.

I'm being a little unfair to Fahrenheit in the post, after some thought. His 0 was placed at a spot that's very easy to accurately calibrate to.

Getting sufficiently pure water to get an accurate 0°C could be annoying back in the day. But half-water-half-salt? Easy.

Minor impurities could easily shift 0°C by a noticeable fraction of a degree, bad if you're trying to calibrate scientific instruments. But when your solution is *half salt*, those impurities just don't matter anymore, so 0°F is gonna be pretty spot-on.

His high-point was still pretty arbitrary nonsense, but there's historical logic to it - he was adapting from the earlier Roemer scale, and just cut the Roemer degree into fourths because it was too big.

That put boiling water at 240°F, but later changes brought it slightly down

(There was probably some minor numerology involved in choosing pure-water's freezing point to be 32, especially considering that made body temp at 96, an exact multiple.)

Keeping this thread linked: https://twitter.com/tabatkins/status/1308226259374149633