A lot of tasks are not really worthy of being implemented in completely custom code. Let's, for example, have a look on how a programmer might solve the task of finding out how many unique words a text contains with the programming language Haskell. The first line defines a function unique_words and the second one demonstrates its use with an example string:

Wow, that is short! Without explaining Haskell syntax too much, let's see what the code does. It defines the function called unique_words, which applies a series of functions to its input. It first maps all the characters from the input to lowercase with map toLower. This way, words like FOO and foo can be regarded as the same word. Then, the words function splits a sentence into individual words, as from "foo bar baz" to ["foo", "bar", "baz"]. Next step is sorting the new list of words. This way, a word sequence such as ["a", "b", "a"] becomes ["a", "a", "b"]. Now, the group function takes over. It groups consecutive equal words into grouped lists, so ["a", "a", "b"] becomes [ ["a", "a"], ["b"] ]. The job is now nearly done, as we now only need to count how many groups of equal words we got, which is exactly what the length function does.

This is a wonderful style of programming, as we can read what happens from right to left because we are just, kind of, describing a transformation pipeline. We don't need to care how the individual pieces are implemented (unless it turns out that they are slow or buggy).

However, we are not here to praise Haskell but to improve our C++ skills. It is possible to work like this in C++ too. We will not completely reach the elegance of the Haskell example but we still have the fastest programming language there is. This example explains how to imitate function concatenation in C++ with lambda expressions.