dynamiclink

Dynamic linking is a technique in software development in which a program references libraries at run time rather than embedding them at compile time. This allows multiple processes to share a single copy of a library in memory, reducing disk and memory usage and enabling updates to libraries without recompiling dependent executables. By contrast, static linking embeds library code into each executable, increasing size and duplication.

The dynamic linker or loader resolves references to library symbols when a program starts and may perform lazy binding, binding symbols on first use. The process relies on metadata such as dependencies and relocation information stored in the executable. At load time, the runtime linker loads required libraries, performs symbol relocation, and ensures compatibility. On many systems, the operating system loader invokes the dynamic linker automatically, with configuration options to influence library search paths.

Platform implementations vary. Unix-like systems commonly use the ELF format with a dynamic linker (for example, ld.so). Windows uses dynamically linked libraries (DLLs) and a loader in the kernel. macOS uses the dyld dynamic linker for Mach-O binaries. Paths and mechanisms to locate libraries can be controlled through environment variables, linker options, and run-time configurations (for example, LD_LIBRARY_PATH, rpath/runpath).

Advantages of dynamic linking include shared libraries across processes, lower overall memory usage, and easier updates. Disadvantages include potential startup overhead, dependency management challenges, and risk of version conflicts, sometimes referred to as dependency hell. See also shared libraries, dynamic loader, DLL, and plugin architectures.