In my new scale, °X, 0 is Earths' record lowest surface temperature, 50 is the global average, and 100 is the record highest, with a linear scale between each point and adjustment every year as needed.
After joking about this at work, I landed on the most cursed scale I could think of... pT = log10 FPW.
Pros: no bottom to scale, increasing negative values asymptotically approach absolute zero. Water freezes at zero.
1 pT is almost exactly the melting point of iridium. Lightning bolts are around 2 pT. Boiling points of neon and helium are in the neighborhood of -1 and -2.
At least that follows some mathematical logic. Mohs scale of hardness is pretty close to pT scale in that sense, but there’s no mathematics or logic involved. It’s just a list of standard materials that define specific points on the scale. When you compare the results with a more logical scale, it looks neatly non-linar at first glance, but the closer you look, the less sense it makes. It’s just a list of exceptions to whatever rule you may have had in mind.
Doesn’t mean it’s a useless scale. You can totally use it for qualitative assessment of hardness, but steer clear of it when numbers and decimals actually matter.
You could totally make an extra cursed temperature scale. Randall proposed the °X scale, but maybe we can do better than that. That was pretty cursed because it defines three points based on statistics observed on of Earth and uses linear interpolation to connect the dots.
I propose an extra cursed system that uses completely fictional values. Let’s take -π as the melting point of unicorns and +GrahamsNumber as the peak temperature in the core of the hypothetical planet Vulcan. Between the two points you can fit any seventh degree polynomial you like in order to get the values that fit your needs. On Wednesdays you can use a sine wave too.
Yeah, well that’s a cultural thing really. Celcius and fahrenheit scales are both quite arbitrary. The kelvin scale uses absolute zero, which totally makes sense, but the other fixed point is pretty arbitrary when you think of it. The fahrenheit scale makes sense for the human experience of weather, while the celcius scale makes sense for generally life on Earth where water plays an important role. Neither of them are particularly universal, and they both suck in their unique ways.
Converting between Kelvin and Celsius is simple addition; converting between Rankine and Fahrenheit is simple addition. Converting between the two groups requires multiplication, and pre calculator, that's notably harder.
Also, all your kJ/kg/°C or BTU/lb/°F tables and factors are identical when you swap to referencing absolute zero. If you change to the other unit system, all that goes out the window.
I think a degree F was 1/10,000 of the volume of mercury he happened to use in his first thermometer. The 180 was probably a coincidence because bimetal spring thermometers came along later.