Summary

In this chapter, we examined a statement: C++ is very difficult to learn. So, is it?

We looked at the history of C++ and how initially it was indeed a challenge to write even the simplest of programs. We saw how Java, C#, and Python deal with some of the problems programmers face with C++ and how the C++ standard has evolved unexpectedly fast in the past 15 years to remove its impediments.

While you can write C++ code that resembles Java or C# today, you will likely still need to understand memory management, a fact we exemplified using move semantics. We also saw that the methods for learning C++ have evolved with the language and with the times, with Stroustrup introducing pointers only in passing and quickly switching to higher-level structures available in STL. We saw that a modified TDD cycle can help people learn C++ in an exploratory manner and without getting overwhelmed by the complexity of error messages and the language.

We also pointed out that C++ has a disadvantage when it comes to tooling and portability. Installing a new dependency is a whole thing in C++, unlike Java, Python, or C#, which provide one de facto standard command to manage packages. This can turn off wannabe C++ programmers who make a deeper analysis.

Finally, despite the progress in the standard, we cannot forget the sheer size of C++ code that is in the world and not up to the latest standard. Chances are, even if you learn modern C++, your work will involve dealing with older code sooner or later.

We conclude, therefore, that C++ is still more difficult to learn than Java, C#, or Python, but that it’s closer than it’s ever been and that the power of the language continues to be attractive for a subset of programmers.

In the next chapter, Ferenc will examine the question: Is every C++ program standard? Or, maybe programmers are driven by solving problems and picking the solutions that work best in their environment, ignoring the standard or even creating idioms that end up in the standard after a while.