Is ± read from left to right or right to left?
So if I write 1m ± 1% ± 1% does that mean 1m ± 2.01% (1% applied twice) or 1m ± 1.01% (1% applied to 1%)?

You hit the one who writes like that 😠

It would depend on the precedence rules for the language, if there even are any for that sign.

Since it symbolizes deviance I cannot think of any real world scenario where it would be a valid syntax and that could mean its just wrong either way :)

I would interpret that as widening the interval twice.

So 1m ±1% ±1% = 1m ± 2.01%

@Voxera But what if the accuracy of my measurement can't be accurately measured?

I would expect parenthesis in the latter meaning.

Well, except that I wasn't expecting two inaccuracies to begin with.

What the ever loving hell? Agree with @iiii don't write like that. Should just be ± whatever your inaccuracy is. If you have a range for your accuracy, you give the error in your measurement as worst possible accuracy.

@lbfalvy "But what if the accuracy of my measurement can't be accurately measured?"

What the actual fuck do you mean by that?

If you have multiple measurements of the same thing, you can calculate the expected value, confidence interval, and standard error of your measurement process.

If you have multiple things but only one measurement per thing, then you can only calculate the mean and standard deviation across the things, but you know nothing about how accurate each measurement is. So there's still only one error you can have.

If you have multiple things and multiple measurements of the same thing, then you can first calculate the error of your measurement process, and then use posterior probability methods to combine the measure error and deviation of what you're trying to measure. In that case, the error is much more complex than a simple product of some relative values, so reporting it as two values instead of the actual error is nothing short of a heresy.

@lbfalvy Also suppose you have a stick which is 100 ± 2 units long, and a measuring tool with an error of ± 1 unit. Your measurements can be anywhere from 97 units to 103 units, which is an error of ± (1+2 = 3) units. Which is why you should only ever use relative error to illustrate the total (combined) error value, and an absolute error everywhere else.