Hi Yuan,   isn't the aim (or one of it) of generic programming to use function templates as a function and especially class templates as a class?   C++ supports this with deeply. For example template arguments are deduced not only directly from a parameter, but also indirectly from a deeper level:   Not only   template<class C> function void f(C)   is possible, but also   template<class C> function void f(vector<C>)   You can have template functions as types/parameters and as member functions of classes.   You cannot take the single address of a template function, because at that level the template function is separated/resolved into actual functions.   You say, a "protocol" is a function in a truer sense than a template function. In what ways and where is this useful?   Best, Sebastian   On Saturday, August 20th, 2022 at 6:27 AM, Jason McKesson via Std-Proposals <std-proposals_at_[hidden]> wrote: I don't know why "concept model" is brought up here. A generic function is a form of function, in the way "member function" being a form of a function. A function template is a form of a template, not even a function. If you want to constrain a template, use a concept. To declare a generic function, you need a protocol, just like to use subtype polymorphism, you need a type with virtual functions.