Incorrect calculations with units
see example
Maybe I was wrong somewh

bugs Units.sm (6,4 KiB) scaricato 1527 volta(e).ere

It turns out that the program before multiplication translates into Kelvin, and then in Celsius.
It turns out that it is necessary to increase the temperature somehow in a special way. help is needed

The question is: what do you mean by "a*0.5"? What is the physical meaning of this`?

The trouble comes from the thermodynamic temperature as a physical quantity (regardless of what units are used) may be not proportional to the numerical value if non-absolute units are used.

If you have in mind to directly manipulate the numerical value, then the result obviously must depend on what units you use. Then it is better to first extract the numerical value, then do the manipulation and then re-apply a unit.

There is indeed a bug, the error message below is pointless and un-useful.

A safe alternative is to use offset-free difference units tagged with Delta.


tunits.sm (4,76 KiB) scaricato 1530 volta(e).

Wrote

The question is: what do you mean by "a*0.5"? What is the physical meaning of this`?

The trouble comes from the thermodynamic temperature as a physical quantity (regardless of what units are used) may be not proportional to the numerical value if non-absolute units are used.

If you have in mind to directly manipulate the numerical value, then the result obviously must depend on what units you use. Then it is better to first extract the numerical value, then do the manipulation and then re-apply a unit.

There is indeed a bug, the error message below is pointless and un-useful.

A safe alternative is to use offset-free difference units tagged with Delta.

I think this is a mistake. You just need to take half the temperature from 20 C. Initially, in reference books and in other literature, half from 20 C, this is 10 C, and not 146K (-126C). Using the program on autopilot, with large formulas. The user may simply not notice that the manipulations with the temperature values ​​are not the same as with other elements. There is a high probability of errors! I know that it is possible to remove the unit of measurement to count and return. You and I know this, while others can make mistakes. It would be possible to make a custom SI table for calculations, then the user himself would decide what half C is and what half F