CppCon 2019

The const qualifier has various uses in C++. One of the most valuable uses is in declaring function headings that constrain the effects of function calls. Using const appropriately can reduce bugs and development time by turning potential run-time errors into compile-time errors that are much easier to find and correct. Using const can even reduce your program’s code size and execution time.

Despite these benefits, too many C++ programmers still use const reactively rather than proactively. That is, they tend to add const as needed to quell compiler error messages, rather than design const in as they code. To get the most out of const, programmers really need to understand (1) when and where to place const in declarations, (2) when to leave it out entirely, and (3) how type conversions involving const behave.

The key insight about const is to understand const as a promise—a promise not to modify something. This session explains the real meaning of that promise and how that insight can guide you in declaring function parameters and return types. It also explains why you shouldn’t declare by-value parameters and return types as const, why overloading on const is such a useful and ubiquitous idiom, and why it is meaningful to declare constexpr member functions as const.