Calculation of number of sides based on $fs is slightly incorrect about openscad HOT 2 OPEN

Well, it's neither the minimum chord length nor the minimum arc length when we factor in the more nuanced behavior of $fs in sphere, which ends up with edges well below $fs, as it is the like the equitorial arc length of a fragment for which equitorial fragments don't always even exist in a sphere. There has been a long stream of debate about how to describe these in a combination of clearly and correctly, which is why that section goes as far as including the code that calculates them (although weirdly including a C version of what is actually a C++ function in the code with a few extra cases of edge case handling). I don't see an issue with the current method of calculating them. But I'll note that the language about "minimum" is really trying to coarsely explain that fmin on the $fa and $fs computations. And there is definitely an opportunity to yet again try to explain that more clearly.

from openscad.

I suspect (without looking at the implementation) that for sphere they control the equator and the lines of longitude.

from openscad.