C++ is one of the most widely used programming languages. It runs the game engines behind major titles, the browsers people use to read this page, the databases that store shopping carts, and the trading systems that move trillions of dollars a day. This chapter gives a high-level picture of what C++ is and why companies still choose it decades after newer, friendlier languages appeared.

C++ is a general-purpose, compiled, statically typed, multi-paradigm programming language that provides direct control over memory and hardware. Each piece of that description in plain language:

General-purpose means C++ is not tied to one specific domain. It can build a video game, a web browser, a database, a tool that controls a satellite, or the firmware inside a dishwasher. It is not built for one job the way SQL is built for database queries. C++ is the language people pick when building something serious that must run fast.

Compiled means C++ source code is translated into machine code (the instructions a CPU understands) before the program runs. The compiler takes a .cpp file and produces an executable file, which runs directly on the operating system without an interpreter in between. The compilation step takes time, but the result runs as fast as the hardware allows. The compilation pipeline is covered in detail in a later chapter.

Statically typed means every variable and function has a type fixed at compile time, and the compiler checks types before the program ever runs. Declaring a variable as int and trying to put text into it produces a compile error. This catches a large class of bugs before they reach users.

Multi-paradigm means C++ does not force one style of programming. Procedural code (functions calling functions), object-oriented code (classes and inheritance), generic code (templates that work for any type), and functional-style code (lambdas, immutable data) all coexist in the same program. Most real C++ projects mix all of these.

Direct control over memory and hardware is the part that sets C++ apart from almost every modern language. The programmer decides when memory is allocated, when it is released, and how data is laid out in memory. Hardware can be addressed directly when needed. With control comes responsibility: mistakes are easier to make in C++ than in a language that manages memory automatically. The payoff is performance and predictability that are hard to get anywhere else.

C++ shows up in places where performance, memory control, or both matter. The main ones:

Game engines and games. Unreal Engine, the engine behind games like Fortnite, Final Fantasy VII Remake, and many others, is written in C++. So is the core of Unity (though game logic on top of it is usually C#). CryEngine, the engine behind Far Cry and Crysis, is C++. Studios like Rockstar (Grand Theft Auto), Naughty Dog (The Last of Us), and id Software (Doom) write their engines in C++ because a game has to render a complex 3D scene 60 or 120 times a second on hardware that varies widely between machines. There is no room for unpredictable pauses.

Web browsers. Chrome, Firefox, Safari, and Edge are all built primarily in C++. On a link click, the code that parses the HTML, runs the page's JavaScript, manages the network connection, and paints pixels is C++. Browsers handle untrusted code from across the internet at very high speed, and C++ provides the performance and control to do that securely.

Operating systems and drivers. Big chunks of Windows are written in C++. macOS uses a mix that includes a lot of C and C++. Many Linux device drivers (the code that talks to a graphics card, network adapter, or storage drive) are C or C++. When code has to run as close to the hardware as possible, C++ is one of the few realistic options.

High-frequency trading. Banks and trading firms use C++ for systems that buy and sell stocks in microseconds. When competing for a price quote and the competition is 50 microseconds faster, the slower side loses. Firms like Jane Street, Citadel, and Jump Trading write their core trading engines in C++ to squeeze every microsecond out of the hardware.

Embedded systems. The software inside cars, planes, medical devices, washing machines, and IoT sensors is often C++. These systems have small amounts of memory, slow CPUs, and strict performance requirements. C++ lets engineers fit fast, predictable code into tight constraints.

Databases. MySQL, MongoDB, PostgreSQL, ClickHouse, and many others are written in C++. A database has to handle gigabytes or terabytes of data, respond to thousands of queries per second, and never lose anything. Memory layout, cache behavior, and avoiding hidden allocations matter at that scale.

Graphics and VFX. Adobe Photoshop, Premiere Pro, and After Effects are mostly C++. So are the rendering pipelines used by Pixar, Industrial Light and Magic, and DreamWorks. Image processing on large files needs the performance only a language like C++ can deliver.

Search engines and big infrastructure. Google's core search infrastructure, parts of Facebook's backend, and large chunks of the systems that run modern social media and streaming services are C++. When serving billions of requests per day, even small performance wins translate into massive savings on hardware.

The diagram shows the common thread. These products look nothing alike on the surface, yet they all share the same need: run fast on real hardware, with predictable behavior, often under tight memory or latency constraints. That is the kind of work C++ is built for.

C++ has been around since the 1980s. Newer languages have promised to replace it, and yet it keeps showing up in the most demanding systems built today. Here is why.

Performance. C++ compiles directly to machine code with no runtime interpreter, no virtual machine, and no garbage collector deciding when to pause the program. The compiler can perform aggressive optimizations: inlining functions, vectorizing loops, eliminating dead code, reordering instructions to fit the CPU pipeline. For workloads where every microsecond counts, this is decisive. A well-written C++ program is usually within 10 to 20 percent of hand-tuned assembly, and often matches it.

Control. In C++ the programmer decides everything: when memory is allocated, when it is freed, how objects are laid out, whether something lives on the stack or the heap, whether a function call is inlined. Most languages take these decisions out of the programmer's hands, which is convenient most of the time but limiting for specific needs. Writing a database storage engine or a game's rendering pipeline makes that control non-negotiable.

No runtime overhead. C++ follows a design principle called "you do not pay for what you do not use." Unused exceptions cost nothing. Unused virtual functions cause no extra indirection. A program that does not allocate memory dynamically does not include allocator code. The output is lean, and the only cost is the features actually chosen. This principle (often called "zero-overhead abstraction") returns in a later chapter.

Mature ecosystem. C++ has been refined over four decades. The standard library has solid building blocks for containers, algorithms, threading, file I/O, and more. On top of that, there are battle-tested third-party libraries for nearly anything: networking (Boost.Asio), graphics (OpenGL, Vulkan, DirectX), GUI (Qt), serialization (Protocol Buffers), testing (GoogleTest, Catch2), and machine learning (parts of TensorFlow and PyTorch). Compilers like GCC, Clang, and MSVC are highly optimized and still actively improved.

Interoperability. C++ talks to almost everything. C libraries can be called directly. A C++ library can be exposed to Python, Java, Rust, or any other language through a C-style interface. When a team needs to integrate with hardware drivers, legacy systems, or libraries written in other languages, C++ is often the glue.

Continued evolution. New versions of C++ ship every three years (C++11, C++14, C++17, C++20, C++23). The language today is far more modern and pleasant than it was in 2000. Features like smart pointers, range-based loops, lambdas, move semantics, and concepts have changed how C++ is written. The old way is still available, but modern C++ feels like a different language.

Picking C++ over something higher-level is a trade-off, not a free win. The benefits of that choice:

Speed and predictability. A C++ program does what the programmer wrote, when they wrote it, with no hidden runtime deciding to garbage-collect or recompile in the middle. A program that must respond in under a millisecond, every millisecond, can be built. A program that must start in 5 milliseconds instead of 500 can be built too.

Tight control over resources. A thread can be pinned to a specific CPU core, a struct can be laid out so that hot fields share a cache line, a pool of objects can be allocated upfront to avoid runtime allocations, or hardware can be addressed over memory-mapped I/O. None of this is exotic in C++. It is how serious systems are built.

Reach. A C++ binary runs on Windows, Linux, macOS, FreeBSD, Android, iOS, game consoles, microcontrollers, mainframes, and almost anything else with a CPU. There is a C++ compiler for almost every platform that exists, and the same code, with minor adjustments, often runs everywhere.

The cost side. C++ asks more than a managed language does. The programmer has to think about ownership of memory, lifetimes of objects, undefined behavior, and a much larger surface of language features. Build systems are clunkier. Error messages, especially around templates, can be hard to parse. The language has sharp edges, and the first few months involve a fair amount of "why does this not compile" and "why did this crash." That is normal. The course walks through those edges deliberately, not around them.

A partial list of teams that pick C++:

The pattern is consistent. When a team is building software that has to be fast, run close to the hardware, scale to enormous workloads, or operate in tight resource constraints, C++ is on the short list. It is not the only choice anymore (Rust, Go, and modern Java have all eaten into specific niches), but in the categories above, C++ is still the default for a lot of work.

No C++ syntax has been introduced yet. The goal here is to see what a tiny C++ program looks like so the language feels concrete instead of abstract.

A few details, without worrying about specifics. The lines starting with #include pull in pieces of the standard library. Execution begins in a function called main. Three variables hold a store name, a product name, and a price. The std::cout lines print text to the screen, and the << operator chains the pieces together. The return 0 at the end signals that the program finished successfully.