For more context, sample reference materials from the NIST reference catalogue are not meant to be exceptionally high purity examples of the sample -- SRMs are usually commercially-produced products. The real value is that each sample comes with a datasheet with incredibly detailed and accurate measurements that are (as you said) used for calibration of measuring devices in industrial settings. To quote a Veritasium video about the topic from a few years ago[1], "the thing you're paying for is not the peanut butter, it's the knowledge of exactly what is in the peanut butter".
Truth. I once paid nearly $500 for a set of 17 certified, calibrated wooden pencils, with documentation showing that each one met their rated hardness levels of 9H thru 6B. And for $500, they don't even come with erasers, that is an added-cost option.
That’s $508 for two cartons for 400 cigarettes total, so $1.27 per cigarette. They’re not characterized like the peanut butter, they’re just manufactured to a specific QC standard.
California/New York prices are easily $0.75/cigg so it’s not that much more expensive.
Yes, it’s a bit frightening to know what a hidden dependency NIST is on many things. In my industry, all calibrations ultimately derive from physical NIST standards for resistance, frequency, etc.
This is why we should ditch Celsius for Kelvin. I never liked how we coupled 0 and 100 to where water freezes and boils under certain conditions. Kelvin doesn't rely on non-invariant reference points.
this is actually why i like fahrenheit the most... 0F is "really cold" and 100F is "really hot", whereas in celsius 0C is "kind of cold" and 100C is "uninhabitable". it makes sense for the 0-100 range to be reserved for temperatures frequently encountered in our habitat as humans i think...
That’s a biased view that leads to a scale only convenient for people living in particular latitudes. Now talk to someone living near the equator like Singapore, and 0F is unimaginably cold (even 40F is basically unheard of), and 100F is very hot but not extraordinarily so. Then you talk to someone living close to the arctic and the converse is true.
All this talk about 0F to 100F being frequently encountered probably reveals your local climate conditions more than it helps define a useful global standard for all humans. We should accept that humans live in more diverse conditions than we intuit and such a scale probably won’t be useful for all.
Now I'm imagining some kind of timezone-like temperature unit that varies around the world, and a "100 things programmers should know about ___" post about all the intricacies and special cases.
Celsius maps nicely to the temperatures observed in a typical Finnish sauna. 100C is among the highest typical sauna temperatures (very relaxing in dry air), and afterwards you cool off with water around 0C
Also, having 100C as water temp for black tee and 80 for green tee is nice to remember
This has always been my primary argument in favor of Fahrenheit. It makes sense for humans, which is all a majority of the population really cares about. In addition, it's more granular. Sometimes turning the thermostat up or down by 1C could be too much, where 1F is just right.
In science and math, I can see where having round numbers for commonly used temperatures would be useful, but if 0C and 100C are the only places that happens, it's not that hard to memorize 2 numbers.
You only think it makes sense for humans because it makes sense for you, where you live the same way you feel changing a thermostat by 1F is perfect but 1C is too much: that’s true for you, not universally.
Fahrenheit functionally requires you to be biased in order for it to seem like a good system.
It provides greater granularity in the range of human comfort, without descending into decimals. How is that not more useful for day-to-day human usage?
0F is cold, 100F is hot. I experience both of these temperatures outside where I live every year.
-20C to -40C seems much more arbitrary as a way to measure the zone humans mostly live in.
I would wager that the outdoor and room temperature are an overwhelming majority of human use for temperature. It seems like it would make sense to optimize for how temperature is more commonly used at a population level.
I would say those who see Celsius as a good system are also biased, because they grew up with it and it’s what they’re used to.
Celcius is pretty intuitive actually. It takes atmospheric pressure into account, Water is pretty abundant on the surface of the planet, and life kinda depends on it. It's a remarkable substance and quite stable. I dont think theres another substance where we can see all three phases with kitchen appliances and still consume it. Equally good for bathing and drinking. I need steam to clean clothes and cook my food. Using water as a benchmark for how we measure temperature is remarkably useful.
With all of those examples, the actual temperature doesn’t really matter.
For drinking and bathing, people go by feel. In the kitchen, the freezer makes water freeze, and the stovetop will make it boil into steam, but exact temperatures aren’t really relevant here either. The same goes for ironing clothes, the iron just does it.
When I think about where explicit temperatures are needed in the home, the oven and thermostat are the two things that come to mind. Water is pretty irrelevant for both of those. For cooking or frying someone may monitor meat or oil temps, but those aren’t water. A hot tub might have an actual temp on it, but those tend to be around 100F, not 100C, as they are meant for humans, not boiling.
Water is an amazing liquid, but I still struggle to see how measuring all our temperatures based on it makes sense.
OF will still kill you faster than 100F, and most folks aren't comfortable at 50F.
It inspires the following idea: A nonlinear scale where ~72F is 0, and +X represents the same level of heat discomfort as -X represents cold discomfort.
Aren’t you just remembering specific numbers though? Unless you’re trying to boil water unseen so that you can only see a thermometer, the specific 100°C mark isn't that useful.
I actually came around to Fahrenheit for the exact same reason. 0°F and 100°F are kind of the top and bottom of the temperature range I’m going to be experiencing and anything outside that is waaay too hot or cold. In Celsius a large portion of people have to flip to negative numbers regularly during the normal year, and everything between 45°C and 100°C is completely unused.
> 0°F and 100°F are kind of the top and bottom of the temperature range I’m going to be experiencing and anything outside that is waaay too hot or cold.
Food, saunas, the weather, a fever, and much more are often over 100F.
Except 0F being “really cold” and 100F being “really hot” is a personal bias of yours based on where you’ve lived. It’s not true for everyone or everywhere.
And even if it were true, it still wouldn’t work. 100F in a dry environment is nice while, in a humid environment, 100F is awful. Same for 0F, a dry and static 0F is fine while a wet and windy 0F is killer.
There's nothing stopping the use of decimal points for Celsius. Though there's also nothing stopping decimal points in Fahrenheit, the rare instance where an imperial sub-unit doesn't make things go fucky.
I think that both the Celsius and Fahrenheit scales are now defined in terms of Kelvin and a formula. Most if not all of the units of measure are no longer related to their history except for an agreement to make the new units match the old ones at some point.
most tv makers and electronics stores in europe give you the screen size in inches first and cm second. they only come in a few common sizes so the specific units dont really matter.
Speaking of tires, its actually a crazy mixture of units here in the US. You might have a size like P215/65R15 which translates to a radial passenger tire of a 15" rim, 215mm wide tread (sidewall to sidewall), and an aspect ratio of 65 which is a ratio of the sidewall to tread width.
So a tire size is expressed in inches, millimeters, and ratios.
Sure technically yes, but also celsius is now technically defined as just Kelvin plus 273.15. Degrees have technically been redefined as exactly 1.380649×10^−23 joules
You can centrifuge it. Heavier molecules move closer to the bottom. Repeat until you reach the desired purity.
You need shitloads of ordinary water to extract a liter of heavy water. And I'd be surprised if there's any superheavy water in that, because tritium radioactively decays.
Fun fact: Heavy water is slightly toxic to humans. If you only drank heavy water, you'd die as it shifts the chemical reactions in your body out of balance. A moderate amount is fine though.
Interesting article but left me wondering why they don't use ¹H₂¹⁶O as the "standard water"? Surely you can make that (relatively) easily in a centrifuge by spinning out the lightest fraction of water.
Reminds of me of the "most expensive peanut butter" at ~$2.44/g - sold by NIST, I assume for some sort of peanut butter factory process calibration: https://shop.nist.gov/ccrz__ProductDetails?sku=2387
For more context, sample reference materials from the NIST reference catalogue are not meant to be exceptionally high purity examples of the sample -- SRMs are usually commercially-produced products. The real value is that each sample comes with a datasheet with incredibly detailed and accurate measurements that are (as you said) used for calibration of measuring devices in industrial settings. To quote a Veritasium video about the topic from a few years ago[1], "the thing you're paying for is not the peanut butter, it's the knowledge of exactly what is in the peanut butter".
[1]: https://youtu.be/esQyYGezS7c?t=226
Truth. I once paid nearly $500 for a set of 17 certified, calibrated wooden pencils, with documentation showing that each one met their rated hardness levels of 9H thru 6B. And for $500, they don't even come with erasers, that is an added-cost option.
https://www.gardco.com/product/pencil-hardness-test-kit-astm...
SRM 2387 pairs well with a cup of ISO 3103.
https://en.wikipedia.org/wiki/ISO_3103
Cap your meal off with a NIST 1196a standard issue cigarette? I am told they taste like drywall and taxes.
https://www.nist.gov/news-events/news/2020/02/nist-releases-...
Perhaps you would prefer some standard issue meat.
https://www.sigmaaldrich.com/US/en/product/sial/nist1546a
A suggested product is "domestic sludge", lol. This is such a fun rabbit hole:
https://www.sigmaaldrich.com/US/en/product/sial/nist2781
Pretty much every process conducted at scale will have such reference samples.
It's only $508 per cigarette:
https://shop.nist.gov/ccrz__ProductDetails?sku=1196a&cclcl=e...
That’s $508 for two cartons for 400 cigarettes total, so $1.27 per cigarette. They’re not characterized like the peanut butter, they’re just manufactured to a specific QC standard.
California/New York prices are easily $0.75/cigg so it’s not that much more expensive.
400 cigarettes actually! Not too bad.
Yes, it’s a bit frightening to know what a hidden dependency NIST is on many things. In my industry, all calibrations ultimately derive from physical NIST standards for resistance, frequency, etc.
It’s frightening that… NIST is doing the very thing it was established to do back when it was called the Bureau of Standards?
I wouldn’t put it past people in my country to buy it just to show off wealth :/
I have recently realized we imported the western world in the most superficial ways.
This is why we should ditch Celsius for Kelvin. I never liked how we coupled 0 and 100 to where water freezes and boils under certain conditions. Kelvin doesn't rely on non-invariant reference points.
You want numbers to be understandable and K produces illogical numbers.
(There is a reason it is common to drop the century when talking about years, humans are less skilled at numbers over 100)
And I know temperatures in Fahrenheit go higher but it is common to refer to Phoenix as "over 100" even when it's high can peak over 120.
That’s 02025 thinking.
> (There is a reason it is common to drop the century when talking about years,
Yep.
> humans are less skilled at numbers over 100)
No, the reason is that the century is constant and therefore carries no information. Where the century isn't constant, it never gets dropped.
You can say the same thing about decades
No one says 6 even though it has a very reasonable information density for conversations. It is either a relative year or a two digit one.
this is actually why i like fahrenheit the most... 0F is "really cold" and 100F is "really hot", whereas in celsius 0C is "kind of cold" and 100C is "uninhabitable". it makes sense for the 0-100 range to be reserved for temperatures frequently encountered in our habitat as humans i think...
That’s a biased view that leads to a scale only convenient for people living in particular latitudes. Now talk to someone living near the equator like Singapore, and 0F is unimaginably cold (even 40F is basically unheard of), and 100F is very hot but not extraordinarily so. Then you talk to someone living close to the arctic and the converse is true.
All this talk about 0F to 100F being frequently encountered probably reveals your local climate conditions more than it helps define a useful global standard for all humans. We should accept that humans live in more diverse conditions than we intuit and such a scale probably won’t be useful for all.
Now I'm imagining some kind of timezone-like temperature unit that varies around the world, and a "100 things programmers should know about ___" post about all the intricacies and special cases.
Celsius maps nicely to the temperatures observed in a typical Finnish sauna. 100C is among the highest typical sauna temperatures (very relaxing in dry air), and afterwards you cool off with water around 0C
Also, having 100C as water temp for black tee and 80 for green tee is nice to remember
This has always been my primary argument in favor of Fahrenheit. It makes sense for humans, which is all a majority of the population really cares about. In addition, it's more granular. Sometimes turning the thermostat up or down by 1C could be too much, where 1F is just right.
In science and math, I can see where having round numbers for commonly used temperatures would be useful, but if 0C and 100C are the only places that happens, it's not that hard to memorize 2 numbers.
You only think it makes sense for humans because it makes sense for you, where you live the same way you feel changing a thermostat by 1F is perfect but 1C is too much: that’s true for you, not universally.
Fahrenheit functionally requires you to be biased in order for it to seem like a good system.
It provides greater granularity in the range of human comfort, without descending into decimals. How is that not more useful for day-to-day human usage?
0F is cold, 100F is hot. I experience both of these temperatures outside where I live every year.
-20C to -40C seems much more arbitrary as a way to measure the zone humans mostly live in.
I would wager that the outdoor and room temperature are an overwhelming majority of human use for temperature. It seems like it would make sense to optimize for how temperature is more commonly used at a population level.
I would say those who see Celsius as a good system are also biased, because they grew up with it and it’s what they’re used to.
> 0F is cold, 100F is hot. I experience both of these temperatures outside where I live every year.
Most people never experience anywhere near 0F in their lives.
> -20C to -40C seems much more arbitrary as a way to measure the zone humans mostly live in.
The zone most humans live in does not routinely (if ever) get anywhere near -20, to say nothing of -40.
Celcius is pretty intuitive actually. It takes atmospheric pressure into account, Water is pretty abundant on the surface of the planet, and life kinda depends on it. It's a remarkable substance and quite stable. I dont think theres another substance where we can see all three phases with kitchen appliances and still consume it. Equally good for bathing and drinking. I need steam to clean clothes and cook my food. Using water as a benchmark for how we measure temperature is remarkably useful.
With all of those examples, the actual temperature doesn’t really matter.
For drinking and bathing, people go by feel. In the kitchen, the freezer makes water freeze, and the stovetop will make it boil into steam, but exact temperatures aren’t really relevant here either. The same goes for ironing clothes, the iron just does it.
When I think about where explicit temperatures are needed in the home, the oven and thermostat are the two things that come to mind. Water is pretty irrelevant for both of those. For cooking or frying someone may monitor meat or oil temps, but those aren’t water. A hot tub might have an actual temp on it, but those tend to be around 100F, not 100C, as they are meant for humans, not boiling.
Water is an amazing liquid, but I still struggle to see how measuring all our temperatures based on it makes sense.
OF will still kill you faster than 100F, and most folks aren't comfortable at 50F.
It inspires the following idea: A nonlinear scale where ~72F is 0, and +X represents the same level of heat discomfort as -X represents cold discomfort.
This makes sense if weather is the main focus. For things like cooking Celsius is great though.
I am failing to understand why there would be a distinction. They all seem arbitrary set points you learn.
450F vs 230C does not feel like anything gets easier there. Maybe in some highly technical baking there are some nice deltas that appear?
Aren’t you just remembering specific numbers though? Unless you’re trying to boil water unseen so that you can only see a thermometer, the specific 100°C mark isn't that useful.
I actually came around to Fahrenheit for the exact same reason. 0°F and 100°F are kind of the top and bottom of the temperature range I’m going to be experiencing and anything outside that is waaay too hot or cold. In Celsius a large portion of people have to flip to negative numbers regularly during the normal year, and everything between 45°C and 100°C is completely unused.
> 0°F and 100°F are kind of the top and bottom of the temperature range I’m going to be experiencing and anything outside that is waaay too hot or cold.
Food, saunas, the weather, a fever, and much more are often over 100F.
Except 0F being “really cold” and 100F being “really hot” is a personal bias of yours based on where you’ve lived. It’s not true for everyone or everywhere.
And even if it were true, it still wouldn’t work. 100F in a dry environment is nice while, in a humid environment, 100F is awful. Same for 0F, a dry and static 0F is fine while a wet and windy 0F is killer.
Also for things like hvac temps Celsius is rather too coarse where as 1F is right around the perceivable difference.
There's nothing stopping the use of decimal points for Celsius. Though there's also nothing stopping decimal points in Fahrenheit, the rare instance where an imperial sub-unit doesn't make things go fucky.
I think that both the Celsius and Fahrenheit scales are now defined in terms of Kelvin and a formula. Most if not all of the units of measure are no longer related to their history except for an agreement to make the new units match the old ones at some point.
The US is technically metric in that the units are defined in terms of the metric equivalent. A US inch is exactly 2.54cm.
Now do metric vs imperial.
Nothing screams freedom more than using ancient terms like horsepower and measuring tires in inches and percentages.
The other day I heard an American talk about a TV being 65 inches. Took me a while to understand they had a 165 centimeter diagonal television.
most tv makers and electronics stores in europe give you the screen size in inches first and cm second. they only come in a few common sizes so the specific units dont really matter.
I bet you’d use lightyears over petametres or exametres though.
It's more typical to use parsecs in practice for astronomy.
Speaking of tires, its actually a crazy mixture of units here in the US. You might have a size like P215/65R15 which translates to a radial passenger tire of a 15" rim, 215mm wide tread (sidewall to sidewall), and an aspect ratio of 65 which is a ratio of the sidewall to tread width.
So a tire size is expressed in inches, millimeters, and ratios.
Surely you mean tyres? It’s all arbitrary. Get over yourself.
ahem U.S. Customary
https://en.wikipedia.org/wiki/United_States_customary_units
Kelvin does still rely on each degree being 1% of the way between freezing and boiling so it kinda does rely on a variant reference point
The reference point is the triple point, and invariant.
Sure technically yes, but also celsius is now technically defined as just Kelvin plus 273.15. Degrees have technically been redefined as exactly 1.380649×10^−23 joules
Why is that only "technically yes"? It's precisely how the thermodynamic and empirical temperature scales are matched.
A US gallon of deuterium water costs roughly $2,600 to $3,800, while a gallon of pure tritium water (super-heavy water) would cost about $44 million.
Makes me want to go to the ocean and start filtering out water somehow.
You can centrifuge it. Heavier molecules move closer to the bottom. Repeat until you reach the desired purity.
You need shitloads of ordinary water to extract a liter of heavy water. And I'd be surprised if there's any superheavy water in that, because tritium radioactively decays.
Fun fact: Heavy water is slightly toxic to humans. If you only drank heavy water, you'd die as it shifts the chemical reactions in your body out of balance. A moderate amount is fine though.
Another fun fact - heavy water has a sweet taste.
There are a variety of ways to do it, but you'd want to be careful that you don't attract attention from your local nuclear regulatory agency.
https://en.wikipedia.org/wiki/Heavy_water#Production_methods
Interesting article but left me wondering why they don't use ¹H₂¹⁶O as the "standard water"? Surely you can make that (relatively) easily in a centrifuge by spinning out the lightest fraction of water.
At the very least that has the advantage that it's possible to make more of it, so you aren't dependent on a limited sample of standard water.
How does the water taste?
Loved reading this. Thank you!
gallon is not a si unit ;)
As specialized world-class lab supplies and reagents go, that is not very expensive. Stuff like pure Xe (for space thrusters) can be more by mass.