Source code
Human-readable text that describes computation in a programming language, meant to be translated and executed.
Definition
Source code is the human-readable text we write in a programming language to describe what a computer should do.
Source code captures intent and structure. It’s designed for humans first (readability, maintainability) and only indirectly for machines.
Synonyms and related terms
- Synonyms: codebase (when referring to a whole project), source files
- Related: compiler, interpreter, build system, artifact
How source code becomes runnable
Most machines can’t execute source code directly. Source code typically needs translation via:
- compilation (ahead of time), or
- an interpreter (as it runs), often with JIT compilation inside it.
The end result is an artifact you can ship and run, such as a binary or executable (often after linking).
Source code vs artifacts
Source code is an input. Artifacts are outputs.
Operationally, what’s deployed and executed is almost always an artifact produced from source code. When diagnosing production issues, it matters to know:
- which source revision produced the deployed artifact
- which dependencies and toolchain were used
- whether the runtime environment matches expectations
Common confusions
- “The code is deployed” usually means an artifact is deployed, not source code.
- “Same source code” does not guarantee “same behavior” if dependencies, build settings, or the runtime environment differ.
Mini-scenario
Two teams deploy “the same code” from the same Git commit. One uses a newer compiler and a different dependency lockfile. The resulting artifacts differ, and a performance regression appears only in one environment. The root cause is not the source code text - it’s the build inputs that sit beside it.