A difference in temperature has the same numerical value in both degrees C and K. Often differences in temperature are all we're interested in so C is often interchangeable with K.
If you use Fahrenheit, then you lose that easy relationship unless you also use Rankine for absolute temperature. But nobody, not even Americans, has even heard of Rankine so it would be a far bigger challenge to get anyone to use that than just Fahrenheit alone.
Perhaps an even more elegant unit for temperature would be thermal energy per degree of freedom per particle, which would be joules in SI. I'm not sure if that's really correct tho.
> A difference in temperature has the same numerical value in both degrees C and K.
I know, I used them basically interchangeably in my post.
> unless you also use Rankine
If people used Fahrenheit more often they would use Rankine more often as well (and yes, I was aware of it before this discussion).
> which would be joules in SI
Joules is a unit of energy, not temperature. You need to multiply by entropy to get temperature.
It would be one thing if the unit for temperature was based on the SI units, fine. But it's not - it's completely arbitrary, and since that is the case you might as well use the more usable Fahrenheit.
If I were doing it I would redefine it such that 1 joule of energy applied to 1 gram of water = 1 new-degree. (Right now it's 4.2 joules.) That would make the temperature scale less coarse, which is good. Then make absolute zero -1000 new-degrees.
Isn't temperature an energy scaled by the Boltzmann constant? That means it's proportional to the energy of something. Why not just use that energy directly instead of converting it to another unit? That's what I means about using Joules for temperature - divide kelvins by Boltzmann's constant
Why on earth would you make absolute zero anything other than 0? That's the major problem of Fahrenheit and the reason Celsius is not the SI unit. Instead, you could put a human-level temperature at a round number like 1000 new degrees so everyone always has the leading 1 in there for everyday values.
If you use Fahrenheit, then you lose that easy relationship unless you also use Rankine for absolute temperature. But nobody, not even Americans, has even heard of Rankine so it would be a far bigger challenge to get anyone to use that than just Fahrenheit alone.
Perhaps an even more elegant unit for temperature would be thermal energy per degree of freedom per particle, which would be joules in SI. I'm not sure if that's really correct tho.