I never liked Mathias’s description there. A better description is that JavaScript uses unvalidated UTF-16, with UTF-16 code unit indexing and access in most pl

I never liked Mathias’s description there. A better description is that JavaScript uses unvalidated UTF-16, with UTF-16 code unit indexing and access in most places. (Notwithstanding this, new functionality in the language tends to use Unicode scalar values where possible, e.g. string iterator decodes valid surrogate pairs.)