Compiled vs Interpreted
How Compilers and Interpreters Run Code
Compiled vs interpreted languages describes two ways of turning source code into instructions a computer can execute. A compiled language is translated into machine code, or an intermediate form, before the program runs. An interpreted language is read and executed by a program called an interpreter while the program runs.
The difference is mainly about timing. A compiler does its work ahead of time and produces a file that is ready to run. An interpreter starts immediately but repeats part of the translation work as the program runs, which makes simple interpreters slower at repetitive tasks. For the software that executes code, see what a runtime environment is.
Bytecode and Just-in-Time Compilation
Many languages sit between the two extremes. Java and C# compile source code to bytecode, a compact intermediate format that a virtual machine runs. Python also compiles its source to bytecode internally before executing it.
Just-in-time (JIT) compilation is a technique in which the runtime translates frequently used parts of a program into machine code while it runs. JIT tries to combine fast execution with the flexibility of interpretation. Implementations differ, so the labels are rarely pure.
Compiled code is not limited to servers. Compilers can also target browsers, as with WebAssembly.
What the Difference Means for Websites and Hosting
For web hosting, the distinction shows up in deployment. A compiled application is usually built once and uploaded as a single executable or package. Go, for example, produces a binary that runs on a server without a separate language runtime, which simplifies deployment.
Interpreted and bytecode languages need a matching runtime on the server, such as the PHP engine or the Python interpreter, and the source files are often uploaded directly. This makes changes quick to deploy, but the server must have a compatible version installed.
PHP is usually described as interpreted, although the PHP engine compiles each script to an internal form before running it. Modern PHP installations typically include a bytecode cache that keeps that form in memory between requests, so most sites do not notice a separate build step.
A Typo Fix in Each Model
Suppose a developer corrects a spelling mistake in a message shown on a site. In an interpreted language, the change is a single edited file. The developer uploads it, and the next request uses the new text, provided the runtime is the same version.
In a compiled service, the same fix requires rebuilding the program and deploying the new binary. The extra step takes longer, but the result is a single artefact whose behaviour is fixed at build time. Both approaches are valid. The trade-off is between speed of change and the predictability of a built artefact.
Frequently asked questions
Is PHP a compiled or interpreted language?
PHP is usually described as interpreted, because the PHP engine executes scripts at run time rather than producing a standalone program. Internally it compiles scripts to an intermediate form first, so the strict split between the two labels does not fit it perfectly.
Is an interpreted language always slower?
Not necessarily. Interpreted languages are often slower than compiled ones for heavy computation, but the gap depends on the engine, the workload and optimisations such as JIT compilation. For many websites, time spent waiting on the database or network matters more than the execution model.