This separation was never a hard requirement. For example `std::map` implements `lower_bound` and `upper_bound` as member functions because they can be implemen

This separation was never a hard requirement. For example `std::map` implements `lower_bound` and `upper_bound` as member functions because they can be implemented more efficiently than the general `std::lower_bound` and `std::upper_bound`, at least at the time. The paper itself discusses the member function vs. free function problem: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2020/p1679r3.html#42-member-function-vs-free-function http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2020/p167...