Hello,
Why this discrepancy?


Thank you,
Basile

Wrote

Why this discrepancy?

Maybe user style ?

°C can't be cancelled against K. Changes or differences of temperatures have to be tagged by Delta in the non-absolute temperature units.

See this post

Wrote

°C can't be cancelled against K.

Wolfram alpha says K/°C=1.
Do we need the same behavior?

Wrote

Wolfram alpha says K/°C=1.
Do we need the same behavior?

It doesn't show that here: https://www.wolframalpha.com/input/?i=1kelvin%2F1degC

It converts °C to Kelvin before performing the division.

Edit: Here shows the result 1 K/K, but it specifies it's interpreting °C as delta degree C.

Personally, I prefer using the explicit Δ°C when performing these types of computations. I understand it's may be more confusing to some, but I also believe it's important to be explicit in syntax.

Wrote

Wrote

°C can't be cancelled against K.

Wolfram alpha says K/°C=1.
Do we need the same behavior?

Hi Andrey. Wolfram says what the author of the units package thinks that it's correct. Actually, here it's my own Mathematica units package: http://library.wolfram.com/infocenter/MathSource/5511/

I think that it's better check what says the SI norm:

https://en.smath.com/forum/yaf_postsm51217_Temperature-Units-Discrepancy.aspx#post51217

Best regards.
Alvaro.

Wrote

It doesn't show that here: https://www.wolframalpha.com/input/?i=1kelvin%2F1degC

It converts °C to Kelvin before performing the division.

I've tried this one: https://www.wolframalpha.com/input/?i=1*K%2FCelsius

Wrote

I've tried this one: https://www.wolframalpha.com/input/?i=1*K%2FCelsius

Interesting. I would think these both should compute to the same value since WA says they are assuming temperature in degrees Celsius and not delta temperature.

For what it's worth, I think the image on the right is incorrect IF assuming degrees Celsius as stated.



The following explicitly states the assuming of degrees difference and computes to 1 appropriately.

Did you attach the wrong image? I don't see an appropriate computation to 1. Instead, I see 1*K/K, K is not cancelled, perhaps because each stands for different quantities. Yet I don't see why any tagging or assumption making should be required for offset-free temperature units. Is there any example where such distinguishing would be required?

Wrote

The following explicitly states the assuming of degrees difference and computes to 1 appropriately.

:d

All the elaborate assumption making doesn't save WA from adding two absolute temperatures and taking the result as a difference.
I still vote in favour of the Delta-tagging of non-absolute units. Perhaps we should add a hint on that to the help strings of °C and °F (to use Deltas if changes or rates are expected).
As long as temperature results strictly default to K units, we are safe. If someone wants to write °C, he will see the hint and still can judge whether to tag or not to tag with Delta. This might reduce the frequency of this issue popping up in the forum.

SI units system is largely not compatible with Engineering.