Why Google’s New Carbon Language Could Be the Future Successor to C++

Carbon Language Overview

At the C++ North conference in Toronto, Google software engineer Chandler Caruth presented Carbon, described as an experimental successor to C++. He emphasized the community’s strong interest and promised accelerated release of conference recordings.

Caruth, Google’s technical director for core programming languages and a representative on the C++ standards committee, also contributes to LLVM and Clang.

Carbon aims to address C++’s large legacy and technical debt, which make incremental improvements difficult. While many developers consider migrating to languages such as Rust, Kotlin, Swift, or Go, such migrations often incur performance overhead and steep learning curves.

Carbon’s goal is to match C++ performance, provide seamless bidirectional interoperability, and offer a smooth learning curve for C++ developers. The team envisions “source‑code translation” capabilities similar to TypeScript for JavaScript or Kotlin for Java.

The language promises full interoperability with existing C++ code and a toolchain that can compile C++ sources.

Reasons C++ is hard to evolve include its long history, accumulated features, binary‑compatibility constraints, and a standards process focused on standardization rather than design.

Carbon proposes an open‑source approach with a built‑in package manager to bridge gaps in the C++ ecosystem. Its roadmap targets a 0.1 release this year, 0.2 in 2023, and a 1.0 version in 2024 or 2025.

Key Language Features

Functions are declared with fn, variables with var, and the language includes strongly‑typed tuples.

The auto keyword supports type inference; pointers are allowed but pointer arithmetic is not.

Pointer operations are limited to addressing and dereferencing.

Classes support single inheritance only.

Memory safety is an important consideration, though not the primary initial goal.

Compared to Rust, which is designed for safety from the ground up, Carbon focuses on migration from C++. Developers note that Rust is becoming the standard low‑level language, but Carbon’s 100% C++ interoperability could make it a compelling alternative.

Developer opinions highlight that many C++ projects start daily, and a solution that surpasses C++ while remaining fully interoperable would be highly valuable.