Google Announces Carbon: An Experimental Successor to C++

At the CPP North conference, Google announced a new programming language called Carbon , branding it as the "successor to C++" and open‑sourcing it on GitHub, where it quickly rose to the top of the trending list with over 21.3k stars.

These years there have always been new languages trying to become the successor of C++; I welcome experiments in programming languages and styles. But Carbon is too new and under‑specified , so I cannot make a meaningful technical comment. Without developing entirely new language rules, libraries, and management schemes, it is hard to provide a viable alternative to C++.

Google engineers, led by Chandler Carruth, argue that C++ is now 40 years old, burdened by technical debt, heavy backward‑compatibility requirements, and a bureaucratic standardization process, making it difficult to add new features.

They propose building a language from the ground up with modern generic systems, modular code organization, and a simple, consistent syntax, rather than inheriting C++'s legacy.

Why Not Just Evolve C++?

According to Carruth, C++’s many decades of accumulated technical debt and its focus on backward compatibility lead to massive codebases, while its open‑source and proprietary compilers evolve slowly, and the standards committee’s bureaucratic approach hinders adding new capabilities.

"Exploring What C++ Can’t Do"

Similar to how Microsoft created TypeScript to evolve JavaScript, Carbon aims to address memory safety and generics shortcomings in C++ by offering features such as the Introducer keyword, read‑only function parameters, pointer indirection, expression‑named types, API imports by package name, powerful generic definitions, and more.

Introducer keyword and simple syntax;

Read‑only function parameters;

Pointer indirection and variants;

Expression‑named types;

API imports by package name;

Powerful, definition‑checked generics;

......

Future development focuses on performance‑critical software, language evolution, readable and writable code, practical safety and testing mechanisms, fast and scalable development, modern OS platforms and hardware, and seamless interoperability and migration with existing C++ code.

The team also plans to add a built‑in package manager, a feature long missing from C++.

Example: calculating the sum of areas of two circles (radii 1 and 2) in both C++ and Carbon, illustrating syntax differences.

Carbon and C++ interop is demonstrated, showing that programmers can call C++ libraries from Carbon without extra overhead and vice versa.

Carbon does not initially provide a memory‑safety system; the project prioritizes migration over safety, hoping future versions may achieve safety comparable to Go or Swift.

Before Carbon, Mozilla released Rust in 2015, a language also positioned as a C++ successor with a strong focus on memory safety.

Critics note that C++ and Rust have poor interoperability, making migration to Rust difficult for large C++ codebases, whereas Carbon is designed for seamless two‑way interoperation with existing C++ libraries.

Community Reactions

Some developers praise Carbon for simplifying language syntax and improving safety, while others doubt its ability to overcome C++’s entrenched features and worry about Google’s track record with experimental projects.

Comparisons are drawn to Rust and the D language, which also aim for C++ compatibility and share many of Carbon’s features.

Overall, the article invites readers to share their opinions on Carbon’s potential as a C++ successor.