Thank you for this post! I did not follow the whole history on the math forge, but I believe your post summarised the current issue quite well.
tl;dr: I agree with option B.
Speaking from a mathematician’s POV, I think you’re quite right to point out that the ambiguity between a symbol and a function (Typst or mathematical) can never be completely resolved. Whether pi in pi(x) is pi x or pi: x |-> pi(x) is completely contextual.
Regardless of LaTeX, I think most users expect f_i(x) to render as f_(i) (x). I have had to repeatedly correct my equations to fix this. Even though it’s a breaking change, it is to be expected. I would rather this is changed now than later.
IMO, this is the right way to go especially if Typst can warn users about script ambiguity with functions.