I won't argue against this specific case. But in almost any other context, I think comprehension would be better served with zero significant digits, or at most 1. This is one reason that I propose magnitude notation: https://saul.pw/mag.
Also, thinking in terms of log quantities makes me think of Fermi estimation, which I love.
But dropping units seems mad to me! From the page:
For instance, “10 million kWh” becomes “↑13.6”
But then I'm a proponent that people should use WAY MORE units every day. Almost every number should have a unit after it. If I'm estimating how many people live on a street it should always be: "10 [houses] x 2.5 [people/house]" not "10 x 2.5".
Anytime you actually need to do anything with a number, you usually need more precision, and any new notation breaks the ability to copy and paste to a calculator, until all the calculators get support.
And actually, anytime I need to do anything with a number, I either use a computer (in which case I'm not entering in the numbers manually), or I don't need more precision. In fact more precision completely gets in the way.
Also, we have an existing "scientific notation", which is completely incapable of being copy-pasted into any calculator I've ever used.
Is this something people do? I only saw this in English (American I guess) online articles.
Once you bother to get past the first paragraph, the entire rest of the article explains exactly why the extra digits are useful to the author in some specific circumstances. (Hint: it has exactly nothing whatsoever to do with propagation of significant digits, which they also clearly demonstrate their understanding of.)
Since the data in the mentioned experiment are "exact" measurements without uncertainty, the data is being treated as infinite (or rather, machine) precision, where the percentage difference would also have to be reported at machine precision, which is what the author is getting at:
>>> 100 * ((37/45) / (312/401) - 1)
5.676638176638171
However, assume that the same kind of data came from a sensor with some uncertainty (say, 2 digits precision), then you could have, within that uncertainty bound, L = 46 (instead of L = 45, etc.), R = 400 (and still L+R = 446), which would give
>>> 100 * ((37/46) / (312/400) - 1)
3.121516164994431
Obviously, this is a huge difference, which is why propagation of significant digits needs to be considered if there is any uncertainty in the data. And in that case adding more digits will not buy you anything.
The author's case is a special case where the values are "exact" and therefore you need more precision in the reported percentage value. But it is often not applicable in science when measured data has uncertainty.