An interesting alternative to implementing our own algorithm for splitting strings would be implementing an iterator that does the same. We are not going to implement such an iterator at this point, but let's have a brief look at such a scenario.

The iterator would need to jump between delimiters on every increment. Whenever it is dereferenced, it needs to create a string object from the iterator positions it currently points to, which it could do using a binary function such as binfunc, which we used before.

If we had an iterator class called split_iterator, instead of an algorithm split, the user code would look as follows:

string s {"a-b-c-d-e-f-g"};
list<string> l;

auto binfunc ([](auto it_a, auto it_b) {
    return string(it_a, it_b);
});

copy(split_iterator{begin(s), end(s), ‘-‘, binfunc},{}, back_inserter(l));

The downside of this approach is that implementing an iterator is usually more complicated than a single function. Also, there are many subtle edges in iterator code that can lead to bugs, so an iterator solution needs more tedious testing. On the other hand, it is very simple to combine such an iterator with the other STL algorithms.