Hacker News Reader

Comment by ars | Hacker News Reader

That has nothing to do with temperature.

You can keep metric and use Fahrenheit with very little change.

And you got your units wrong anyway: "To heat one cubic centimetre of water up 1°C takes 1 Joule" - you mean 1 calorie, not 1 Joule.

Change Celsius to Fahrenheit and only calorie changes, and since it's anyway not part of SI it makes little difference.

Joule is the main energy unit, and it's not tied to Celsius at all.

I totally get the arguments for metric - except for temperature. For temperature Fahrenheit is the better scale.

Look here: https://en.wikipedia.org/wiki/International_System_of_Units#...

And notice how Kelvin stands alone - it's not tied to any other unit. It can be replaced with almost no impact.

View on HN

Ah, you're right about calorie vs. Joule. That is a bit rough around the edges.

> For temperature Fahrenheit is the better scale.

Yeah, no. It's a stupid system. It's not even calibrated properly. 100° was supposed to be body temperature. They got it wrong. 0° was the lowest temperature they could measure at the time because that's when mercury froze. That's so useful! I really want to know if the mercury I carry around all the time is going to freeze outside.

Meanwhile °C is simple: 0°C is "watch for ice" and 100°C is "water boils" (minus altitude factor).

I don't know how you can praise Kelvin and slam Celsius since they're the same "size", it's just that the zero point of Kelvin was set to absolute zero instead of the freezing point of water. °C = °K + 273.15. That's it.

> That is a bit rough around the edges.

Correct. It would be one thing if Celsius is somehow tied into the rest of the SI units, but it is not, which means we can replace it.

> It's not even calibrated properly.

Who cares if they got the calibration of Fahrenheit wrong, at the end of the day it's more usable than Celsius for humans.

0 Fahrenheit is the temperature when it's really cold outside, 100 is when it's hot. It doesn't get simpler than that. If you are outside those ranges you need to take special precautions. No, I don't care about mercury either, but I do care to know when it's cold.

It's not that hard to remember 32 for ice. With Celsius you need to deal with negative temperatures all the time, and worse, the size of the degrees is too coarse.

> 100°C is "water boils"

And exactly how often do you need to measure the temperature when water boils? Probably about as often as I freeze mercury. So 100 is a very useless range on Celsius - you are using your "best" numerical range on useless temperatures.

On the other hand it's quite common to want to know when it's hot enough that you will sweat. Above 100 (body temperature) you will sweat.

Labs and Ovens don't really care about which exact unit, so let them deal with the higher (or lower) numbers - the only time it matters is for ordinary humans checking the weather.

> I don't know how you can praise Kelvin and slam Celsius

You misread, I did not praise Kelvin, I said we could replace it.

> It's not that hard to remember 32 for ice. With Celsius you need to deal with negative temperatures all the time, and worse, the size of the degrees is too coarse.

Well, the whole world manages to deal with °C and only America and a handful of backwater countries insist on using °F.

"Too coarse" is absurd. People just measure to the nearest 0.5°C if you want to get all fussy, but that's not normally the case. It's 12° out? Wear a jacket. 12.0 vs. 12.5° isn't going to change that.

It keeps the numbers smaller. Temperatures here flop between -40°C and +40°C, so the numbers never get large. Meanwhile in the US you see temperatures of 100°F all the time as well as negative ones like -30°F. It's a much wider spread for no real reason.

I hope you know the only reason America didn't get with the program and metricize like it was fully intending to do in the early 1980s is because Reagan was an asshole.

I'm onboard for metric - just not for temperature.

Just because you manage to deal with C doesn't make it better.

Do you see the parallels? You tell Americans all the benefits of metric, they don't care. You tell Europeans all the benefits of Fahrenheit, they don't care.

People don't want to change.

> It keeps the numbers smaller.

No, it doesn't. You need 3 digits since you need the decimal (-20.5).

I can't think of a weather station that uses decimals, the forecasts and current temperature reports are almost always whole numbers. My Nest thermostat goes to the closest 0.5°C because I guess why not, but even that seems overkill. Most systems go to the nearest whole degree.

Decimals are only for those who are making a fuss over nothing unless they're doing something that requires precision, like chemistry or in rare cases cooking.

There's no benefits to Fahrenheit. If you want to argue they're both arbitrary, fine, but the entire world uses °C, so get with the program.

Good luck getting the US to ever change. I'm personally on the side that Fahrenheit is more useful for day-to-day usage, because it has a range calibrated to the temperatures experienced in the part of the world I happen to live in. But it would take a whole generation to get people to change their frame of reference into Celsius, which isn't likely to happen, because the biggest interaction people have with temperature is the daily weather report, and it's primarily older people that watch the news.

I could get on board with kilometers, if I had to, because the conversion factor is simple enough to do in your head (roughly 3/5), but trying to figure out what anything is in Celsius is just a little too complicated.

So, 40 or 50 Degree for when it's dangerously hot is too hard to remember, just like 32 for freezing is somewhat easier than 37 for body temperature?

Why let's just change the numeral base to 5, or use roman numerals for that matter. Then scientists can finally stop to convert some arcane unit to get to the Standard International Base Unit, ie. Kelvin. At least everyone gets to convert their numbers then and it's only a division by two, too.

</rant>

It's more that you can't tell the difference between 30C and 30.5C. No one ever forecasts temperatures using decimal degrees here in Australia.

You only need decimals for highly accurate measurements, but at that point you'd need decimals for Farenheit as well.

> You tell Europeans

Erm, how about the rest of the world?

Judging by the title of this thread, I think the GP meant to type Europans, not Europeans.

To me "hot" starts at 80F and "cold" around 40F. When cooking I frequently care about temperature relative to the boiling point when deciding what temp to set the oven or water bath to.

"""And exactly how often do you need to measure the temperature when water boils?"""

I drink green tea a lot. You pour the water at 70-75 degrees C. It's a lot easier to do the math of how much water you need to add to instantly cool the water from the water boiler down to that range.

"""Labs and Ovens don't really care about which exact unit, so let them deal with the higher (or lower) numbers - the only time it matters is for ordinary humans checking the weather."""

What? I use temperature more when cooking/BBQing than when checking the weather.

"""With Celsius you need to deal with negative temperatures all the time"""

This is a feature, not a bug. Negative number -> below freezing.

"""0 Fahrenheit is the temperature when it's really cold outside, 100 is when it's hot. It doesn't get simpler than that. If you are outside those ranges you need to take special precautions."""

I also need to take special precautions within these ranges. Most notably switching between summer/winter tires.

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.

> Isn't temperature an energy scaled by the Boltzmann constant?

Only for an ideal gas. For all other materials the scale number is different. That's why temperature is a unit of its own, and it not just energy.

> Why on earth would you make absolute zero anything other than 0?

I want a "human" oriented scale, that is also tied into SI.

Reply on news.ycombinator.com