Talk:3001: Temperature Scales

Shouldn't Rankine say "0ºR is set to absolute zero"? 172.70.230.29 (talk) 22:58, 21 October 2024 (UTC) (please sign your comments with ~~~~)

Yep. 162.158.186.253 04:38, 22 October 2024 (UTC)

Furthermore, should it be 0°R or just 0R (no °)? I've been told that Kelvin doesn't use degrees because it's an absolute scale, so a) is this true and b) should it apply to Rankine? 172.71.211.54 14:44, 24 October 2024 (UTC)

Kelvin is rather strange, for reasons never totally explained. It's "the Kelvin scale", but the unit is "kelvin" and I never got on with the official absence of the ° symbol by the "K". I was always taught to say "degrees kelvin" (for temperatures) and "kelvin degrees" (for a change/range of temperature) in order to not cause confusion and technical misunderstandings (perhaps easier to contextualise when down in writing?) but no accounting for taste, or possibly official laziness.

On the basis that Rankine is not kelvin (whatever the reason for how kelvin is what it is), I would use the degrees, as I would any other absolute scale (whether it be an adjusted form of °Rø or °Ré or whatever else might be invented), because kelvin is just inexplicably (to me, and to others) the exception to absolutely every other reasonably equivalent contemporary measure, including capitalisation. YMMV! 172.70.162.2 19:53, 24 October 2024 (UTC)

yo,i thought comic 3000 was anticlimactic so randall would make this one COOL but sadly not Same. Hope he does something cool for 3072.172.69.134.225 23:44, 21 October 2024 (UTC)

really he didn't do anything special for this either? come ON randall if you don't do something cool for comic 3072 i will come to your house personally and yell at you RadiantRainwing (talk) 23:57, 21 October 2024 (UTC)
What's random about Fahrenheit? (Answer: nothing.) 0F is the freezing point of brine, 100F (or 98.7) is the human body temperature. 172.68.54.65 00:00, 22 October 2024 (UTC)

What concentration of brine? (And which specific salt... No, not NaCl, as you might presume but NH₄Cl!)

And body temperature varies a lot ('typically' 36.5–37.5°C or 97.7–99.5°F, though even this range is thought to be too small), across genders, individuals, time of day and which orifices/surfaces you try to measure it from. (Originally, it was set so that 90°F was to be the 'best guess' of human body temperature. It gradually changed, including via various compounded misunderstandings so that the best you can say is that 100°F is arbitrarily slightly above most afebrile human body temperature measurements.)

Celsius might be a bit off (arguments about triple-point or STP freezing, etc), but it still has far more physical logic to it. 172.70.160.188 01:14, 22 October 2024 (UTC)

Sorry, Randall, for my comfort, Fahrenheit is the least cursed. It's the best scale to use for my personal use, especially when hearing the weather report and deciding what to wear outdoors: temp in the 80's - no jacket. temp in 70's - maybe a windbreaker if it's breezy. 60's - sweater weather. 50's - medium weight coat. 40's - winter coat. 30'3 - winter coat with scarf and gloves. 20's - multiple layers. teens - stay indoors. None of the other scales provide such convenient distinctions for my daily life. Kelvin is great for astro physics or super conductivity, but useless for any common uses. Celsius is great for hanging out with the Euro crowd but still not so useful to scale my home thermostat. I judge Fahrenheit as 1.0 for cursedness. Rtanenbaum (talk) 14:19, 22 October 2024 (UTC)

I conveniently use Celsius in tens, also. Negative °C: Cold; 0-10°C: Nippy; 10-20°C: Generally pleasant; 20-30°C: Too warm to exert oneself; 30°C+: Definitely too warm. 172.70.86.205 15:24, 22 October 2024 (UTC)

Personally, I'm most disappointed that Delisle scale was not represented... 172.70.160.188 01:14, 22 October 2024 (UTC)

I was so hoping for a Planck temperature quip. Like: "Water freezing point: 0; Water boiling point: 0; Notes: 1 = highest possible temperature (1.4E32K) where thermal radiation creates black holes; Cursedness: 0/0" 162.158.164.184 01:27, 22 October 2024 (UTC)

Same here. Freezing is 0.000000000000000000000000000001928 and boiling is 0.0000000000000000000000000000026338. DanielLC (talk) 03:38, 22 October 2024 (UTC)

Wow, those are even smaller than the IEEE floating point representations of 1-1.0/3*3! 162.158.90.109 03:59, 22 October 2024 (UTC)

The Planck temperature quip is definitely well deserved. Good catch! Mumingpo (talk) 17:24, 23 October 2024 (UTC)

I guess I was wrong in my comment on the last comic. sigh. -P?sych??otic?pot??at???o (talk) 01:16, 22 October 2024 (UTC)

It's actually spelled Wedgwood scale, not Wedgewood. Wilh3lm (talk) 01:17, 22 October 2024 (UTC)

I still call the modern version of the "Celsius" scale "centigrade", but if people start nitpicking, I'm happy to switch to "Carolus" to avoid ambiguity. For some reason that tends to annoy people more though. 172.68.22.191 01:32, 22 October 2024 (UTC)

Every temperature scale is equally "random" as every other scale. People always say that Celsius is so much better because it's defined by the phase changes of water. Okay, cool...why should THAT of all things be what we use as the base for a system of temperature measurement? And, who cares? I'm a Homo sapiens, not a water molecule. If anything we should use the freezing and melting points of humans as our two reference points for temperature (which, I must say, Fahrenheit approximates better than Celsius, assuming 0 and 100 are your points "A" and "B"). Pie Guy (talk) 03:42, 22 October 2024 (UTC)

Every temperature scale is arbitrary, but since boiling and freezing water is a thing humans have a lot of experience with it makes sense to use that as the reference point. At least it makes more sense than whatever the coldest recorded temperature in Fahrenheit's home town was, because he didn't like negative numbers 172.70.250.23 03:56, 22 October 2024 (UTC)

Planck temperature (as above) is probably the least arbitrary, and some would say it is to some extent free from arbitrariness. However, it's completely impractical for everyday use (as above.) 172.69.34.138 04:31, 22 October 2024 (UTC)

Do the physics of black holes or neutron stars involve Planck temperatures greater than 0.0000001? Liv2splain (talk) 07:23, 22 October 2024 (UTC)

Chat Gippity told me:

Black holes and neutron stars do not typically involve temperatures reaching the Planck scale. While both objects exhibit extreme physical conditions, their temperatures are far below the Planck temperature, even though they can be incredibly high compared to everyday phenomena.

- **Neutron stars** have surface temperatures in the range of millions of Kelvin, and the core can reach even higher, possibly up to a few billion Kelvin. These temperatures are still vastly lower than the Planck temperature.

- **Black holes**, especially the smaller ones, can emit Hawking radiation, with temperatures inversely proportional to their mass. However, the temperature of even a very small black hole is still far below the Planck temperature. Hawking radiation is not expected to reach temperatures close to the Planck scale under normal circumstances.

The Planck temperature (TP=1) represents an energy scale so extreme that no known physical models, including those describing black holes and neutron stars, operate near or above this threshold. Temperatures reaching **0.0000001 TP** (or 1.416 × 10^26 K) would still be beyond current observational and theoretical frameworks related to these cosmic objects. A quantum theory of gravity would be required to describe physics at or near the Planck temperature, which remains speculative and is far beyond the conditions found in black holes or neutron stars.

Liv2splain (talk) 08:46, 22 October 2024 (UTC)

If the °X scale is based on the temperatures of Earth from all time (for some definition of "Earth"), then the scale is very hard to define and highly impractical. The earth appears to have gotten to more than 2,300 Kelvin (hot enough to melt steel and platinum and to boil lead) and while I can't find any sources for the lowest temperature, I imagine it is lower than -100°C. The recorded minimum, maximum and average temperatures appear to be around -89.2 °C, 56.7 °C and 15 °C respectively. This would make the scale somewhat useful, but this would make typical values between 41 °X (cold winter's day) and 68 °X (hot summers day) which I think is pretty cursed. I recommend the clearly superior °Y, based around average temp at 0 °Y, low at -100 °Y and high at 100 °Y. These would be measured by the yearly high, low and mean temperatures averaged per person. Then saying "It's 2 times colder than yesterday" would have some reasonable meaning. --198.41.236.147 04:01, 22 October 2024 (UTC)

"Record ... surface temperature" implies it was recorded. 172.68.22.9 04:08, 22 October 2024 (UTC)

How do you all feel about adding an additional column for room temperature 22C/72F?

Unit	Room temperature
Celsius	22
Kelvin	295
Fahrenheit	72
Réaumur	18
Rømer	18
Rankine	531
Newton	7
Wedgwood	-7
Galen	0
Real Celsius	78
°X	59

Or 0.00000000000000000000000000000208 °Planck, lol. 108.162.245.211 05:36, 22 October 2024 (UTC)

I feel like decigalens would be the most practical unit. Who's with me? 162.158.186.5 06:20, 22 October 2024 (UTC)

It's interesting; calculating the equilibrium temperature (with 2.05 and 4.24 being used for the heat capacities of ice and boiling water) gives 67... If I use water that's about to freeze and steam, I get 31. 172.69.0.178 07:59, 22 October 2024 (UTC)

Would you please explain in more detail? Liv2splain (talk) 09:03, 22 October 2024 (UTC)

The equilibrium temperature of a mixture (?) of equal quantities of ice at 0 C and water at 100 C (with the heat capacities 2.05 and 4.24) is 67 C; if I use the data for water at 0 C and steam, I get 31 C. Additionally, if I use equal volumes, I get 68 (which isn't much different.) 172.69.0.178 17:15, 22 October 2024 (UTC)

One can obtain 0 = 22 C by setting the heat capacity of ice to be 39 and that of water to be 11. For any particular "normal temperature" R °C (that is, the temperature at 0 is R,), I find that x °C = 50R(x+4)/(x(R-50)+200). In particular, for R = 22, we get (1100+275x)/(50-7x). 198.41.236.163 05:58, 23 October 2024 (UTC)

How about Tnew=0.1694×degC+46.25; degC=(Tnew-46.25)/0.1694, where 0 is absolute zero and 50 is room temperature? (Freezing point of water: 46.25; Boiling point of water: 63.19) 162.158.186.248 05:21, 24 October 2024 (UTC)

I would absolutely pull the trigger on an additional column if I didn't think it would further screw up what are most probably extremely cursed mobile portrait renderings of the table. How about a Trivia section? Liv2splain (talk) 08:44, 24 October 2024 (UTC)

Question regarding the X scale - when it‘s defined by *three* (somewhat, implying average is real and not just calculated by (max-min)/2)) independent points, how will linearity be achieved? 162.158.155.76 05:43, 22 October 2024 (UTC)

Or click "[Expand]" in the bottom right table cell Derivation.

Please see 2701: Change in Slope. 172.70.206.179 05:50, 22 October 2024 (UTC)

Sure, "a linear scale between each point":

Here you go. Liv2splain (talk) 06:33, 22 October 2024 (UTC)

The reference for the average surface temperature, https://www.space.com/17816-earth-temperature.html, suggests it has increased above 15°C. What value should we use in late 2024? Liv2splain (talk) 07:30, 22 October 2024 (UTC)

The World Meteorological Organization, Carbon Brief, and Copernicus Climate Change Service suggest 17.16°C. Liv2splain (talk) 07:42, 22 October 2024 (UTC)

Updated water temperatures, Derivation, and graph. So we've already had more than the +2°C warming we were trying to avoid in 2019? Liv2splain (talk) 08:05, 22 October 2024 (UTC)

The +2°C (or +1.5°C that we were originally supposed to be avoiding) is over some (undefined) number of years, though, which allows us to ignore the fact that we're cooking ourselves by repeatedly saying 'Oh, but it doesn't count yet.' 172.70.91.62 11:13, 23 October 2024 (UTC)

Regarding [1], are the average surface temperatures from the sources supposed to be yearly or overall averages? Liv2splain (talk) 09:06, 22 October 2024 (UTC)

According to https://climate.copernicus.eu/climate-indicators/temperature the global average near-surface temperature in 2023 was 14.4 + 0.4 = 14.8°C. (see Figure 1 and click "Increase above: [1991–2020 reference period].") 172.68.22.8 21:06, 23 October 2024 (UTC)

where is the interactive epic 3000 comic we should've gotten? This one's cool but 1000 seemed to have more effort in it and 2000 was at least tangetially related. Does Randall just not like making these anymore and is only making more comics as a business? 108.162.238.185 12:14, 22 October 2024 (UTC)

The comic is free on the website and it doesn't have ads; although the comic is part of his "brand" there are many more profitable things he could be doing with his time, and yet he continues to update it every Monday, Wednesday, and Friday. I just don't like the idea of claiming that a creative person "should" produce any particular thing to satisfy their fans. He's a busy guy! Maybe he's working on a book, or a Scientific American article, or a TV show. He's under no obligation to give us anything, and maybe one day he'll stop making xkcd altogether; that's his choice. Sorry to single you out; I know a lot of people feel the same way as you do, but to me it doesn't make sense. He's not a content machine--he's a guy who started posting sketches on the internet. Dextrous Fred (talk) 15:23, 22 October 2024 (UTC)

Sorry if I sounded overly brash, I wasn't trying to imply "wahhh no special entry wahhh", I was just wondering if Randall still likes to make these or if he doesn't, mainly because he just didn't do anything special, which feels like he just didn't care. I wasn't trying to imply Randall should just do it for the fans108.162.238.80 17:52, 22 October 2024 (UTC)

It could be that 3000 (or even 3001) was going to be special but, as fairly frequently with April Fool 'specials', it just wasn't doable on time. (If it's still considered fixablez it might pop up sometime before 3020 or so. Or, if transferable to another occasion (rebranding the obvious "3000!"ness), held over until Haloween, Christmas, April or 4000, perhaps with additional perfections.)

Hard to know, unless Randall (or his technical collaborators) say anything. And it's probably not worth doing so right now. Maybe "Hey guys, this was going to be #3000!" might accompany its eventual emergence, but also maybe not. Does it really matter? 172.70.85.139 13:03, 23 October 2024 (UTC)

Is this the first list-style comic where every single entry is real? (Usually he has several joke entries.) 172.70.114.182 14:26, 22 October 2024 (UTC)

Where would Felsius go on this list?

One can find a smooth function for °X, namely, (477879x-17634840)/(3341x+197700), which takes °X and returns °C. The inverse is (-197700x-17634840)/(3341x-477879). Should this be included in the wiki article? Or maybe another way of fitting it (like exponential) should be used. 172.69.0.165 06:27, 23 October 2024 (UTC)

It says "a linear scale between each point". 172.70.210.130 21:09, 23 October 2024 (UTC)

Should it be noted that in the first _What If?_ book, there's a reference to units and how much Randall loathes rankine? Someone can go take the book and cite it; it's in one of the early pages 172.64.236.10 08:45, 23 October 2024 (UTC)

I remember it being drummed into us in school physics (admittedly over 50 years ago) that 0 Celsius is defined as the melting point of ice, not the freezing point of water (presumably because of supercooling). 172.70.160.189 08:49, 23 October 2024 (UTC)

It seems he wrote "Earths'" (plural possessive) instead of "Earth's". 141.101.98.151 08:58, 23 October 2024 (UTC)

What? No gas mark? It's linear for temperatures over 275°F but inverse powers of 2 below That's pretty cursed, but I still put it in my unit conversion app. It's only used in gas stoves in a few countries, so it doesn't come up very often. By the way, boiling is 1/5.7358 and freezing/melting is 1/843.3572. Interestingly, France has it's own stove temperature scale that seems to be based on °F.

Also, my understanding is that 7.5 and 32 aren't random. Both Romer and Fahrenheit put numbers on things so that freezing/melting of water and "Normal human body temperature", which was thought to be standard at the time, would be some number X (15 for Romer and 64 for Fahrenheit) and the water thing would be to be X/2 and NHBT would be X/2+X. Pretty nerdy. Sadly, the calibration was off and 212 degrees for boiling was found to be less cursed. But I could be wrong.172.68.54.138 20:39, 23 October 2024 (UTC)

Re: water/ice point 'random number', I think it's more that it wasn't considered "special", such that "On my scale, that will be zero" (or whatever choice of handily round number, including zero, some scale-setters allocated to the BP of water).

After all the other messing about ("my zero will be that of brine!", etc), obviously then the ~0°C equivalent would have a number, and perhaps there would then be a slight change to make it a whole (or easy-fraction) number for convenience's sake, but (before the concept of binary computers) there's not much special about landing on the number 32, for what is actually a temperature that is quite significant to the human experience, and less so with 7-and-a-half.

Maybe landing on 90°F (at one time) for body temperature (and 180 F° between MP and BP) was considered useful as the analogue to angular-degrees where 90 (and 180) indeed features significantly, but I don't think there'd have been too much fuss if the value would have turned out to be 60(/120), also with plenty of handy factors to divide by, 70(/140), without so much, or whatever number(s) happened to depict one realistic real-world measurement that (overall) has no reason to have a factor-based relationship with various quite separate phenomenon measurements.

And it went through several 'corrective' iterations so that even its handy relationship with 'about 100°F' can be said to be an incidental accident, at best, unless we do something like Randall's °X scale and actively triple-tie the central value of the slope(s) to be exactly something useful by using the "currently accepted mean human body temperature (given various complicated caveats)".

It's pretty much all random, in the same way that only because of anthropocentric choices of 'standard' time and distance measurements is the speed of light 'pretty much' 3×10⁸ m/s (a handily round value that works well enough for most purposes, even after back-standardising its component SI measurements to make "actually, precisely 299792458" the proper answer, and it could be far worse...). Avagadro's number never had it so good (6.022(+change)×10²³...? ...where's the handily mnemonic value in that?), and Pi (in this universe's system of fundemental mathematics) clearly never ever had a chance! And, on at least one occasion, such happenstance numeric roundedness in its exactitude (29,000 ft) was considered actually quite awkward... 172.70.91.90 21:33, 23 October 2024 (UTC)

My understanding is that the brine thing was an ad-hoc justification when the scale was presented to the Royal Society. It may have seemed less cursed. But yes, there's a BIG reason for using 32 or 64, halving a distance is trivial and as an instument maker, Fahrenheit would have found that attractive. Mind you, I'm getting this from the Straight Dope, so I could be a dope getting it straight. https://www.straightdope.com/21344240/did-cecil-err-in-explaining-the-significance-of-zero-fahrenheit 162.158.10.189 20:04, 24 October 2024 (UTC)

Ok, then, regarding this edit (and the original IP-led one(s) that even made huge and revert-necessary changes), was it really intended to get rid of whole paragraphs such as "Randall also fails to specify what happens with temperatures[...]" that had nothing to do with the numeric adjustments? When I see that, I see mistakes (especially in light of the "clobber" that happened, where typos reappeared and other things became unexplained/worse-explained once more). — Basically, if your edit summary is nust about updating baseline data, and the resulting maths, I don't expect (maybe good, maybe bad) edits to unrelated bits. Or I may (and have) presumed accidental (or deliberate?) carelessness that I'd rather not try to go back to first principles to re-re-check for the editor concerned. That is all. At least try to justify enough of your edit in its own way, even if it means diving in several times to get enough space to summarise your whole "why" to each tweak. 172.68.186.104 22:44, 23 October 2024 (UTC)

Yes. I have re-removed the removals piecemeal with individual edit summaries for clarity. Many of them involved detailed obscure technical misunderstandings, such as whether the Vostok and Death Valley measurements were surface temperatures (the WMO says they are, and there are the WMO's photos of the observation stations in the linked references now) which combined with the incorrect yearly average global mean temperature, added five paragraphs unnecessarily. 172.68.23.152 01:46, 24 October 2024 (UTC)

Definitely some points made (some incorrect, "ne er" was obviously more just a basic typo of "never", not my attempt to use "ne'er" for no good reason, and my attempt to fix that and some other bits ran into a set of Edit Conflicts ...hope I caught all the remaining ones when I finally could try again on the settled-down page) and I've blended answers to your objections in while giving back what useful nuances (from a number of past editors, only a couple of bits even having had my own hand primarilly behind them as they were) really needn't have been removed. I dispute the terms of your objections (as summarised) behind some changes, but have rephrased based upon what I think you mean, giving you should prefer and wouldn't feel the need to be as randomly censorious about. 172.69.195.173 02:54, 24 October 2024 (UTC)

Figure 1 in https://climate.copernicus.eu/climate-indicators/temperature does not seem like a random walk to me. 172.69.33.118 05:02, 24 October 2024 (UTC)

Added the "Random Walk" because, ignoring long-term trends, year-on-year the measured average is going to blip up and down for all kinds of reasons (physical and measuring issues, both), so it will be lower than expected or higher than expected compared to the smoother track it actually takes on a rolling average. I think one of the versions I replaced had partial suggestion that the average was effectively constant (in °C, not just °X), and while records adjusted every now and then (or every year!), it all just rather settled down at the °X midpoint. Which it doesn't. (And also that it's possible that Average and Minimum change such that the projected Absolute Zero doesn't move so much, one year, although mostly the fulcrum will be closer to 0°X rather than 0K.)

But I've yet to see what's been changed (maybe improved) since I was last looking at it. Maybe this is an out-of-date explanation. 172.70.90.202 12:04, 24 October 2024 (UTC)

Division by zero

I saw this in an edit summary: "10/0 is not ∞, it's also an error, not NaN according to the IEEE. It's closer to {+∞, -∞} than NaN but it's still neither because you can't make limits work"

Actually, IEEE floating point 10/0 can be an error, +∞, or NaN depending on the rounding mode. This is one of the reasons why mathemeticians don't appreciate the IEEE as much as they might. Division by zero is strictly undefined because of the problems with limits alluded to in the summary. https://www.youtube.com/watch?v=dHdg1yn1SgE 108.162.245.66 03:48, 24 October 2024 (UTC)

"When considering division by zero through limits, assigning {+∞, -∞} as potential results is insufficient because limits require consistency and well-defined behavior. In the case of dividing a number by values approaching zero, the results differ depending on whether zero is approached from the positive or negative direction. As a divisor approaches zero from the positive side, the quotient grows towards +∞, and from the negative side, it tends towards -∞. Since limits must converge to a single value for consistency, this disparity leads to an undefined result. Moreover, in many mathematical contexts, infinity is not a number but rather a concept describing unbounded growth, meaning operations involving infinity, like addition or multiplication, are not well-defined in the same way as with finite numbers. This inconsistency in approaching zero prevents {+∞, -∞} from being an adequate solution set for division by zero. Defining division by zero as infinity would create contradictions in both arithmetic and algebraic contexts, as it disrupts fundamental properties like continuity and field structures in mathematics. Hence, division by zero remains undefined to preserve mathematical rigor and coherence."

[Chat Gippity 4o] Liv2splain (talk) 08:53, 24 October 2024 (UTC)

https://imgflip.com/i/7yd7gz 172.71.150.131 09:06, 24 October 2024 (UTC)

Trivia section table values?

Someone please double-check the Trivia section temperatures. I am not convinced they are entirely correct or consistent. I'm least sure about the Galen row. And Wedgwood obviously needs more digits of precision. 162.158.41.28 13:10, 24 October 2024 (UTC)