Why does Python run your code as it goes while C translates it first?

C is translated ahead of time because compiled code tends to run faster; Python runs through an interpreter because that makes code quicker to change and easier to carry between computers. Still, the two are less different than they look. A compiler translates your whole program into machine code ahead of time, and the processor then runs that. An interpreter is software that executes your code without first compiling it to machine code: it carries out the instructions itself as it goes. "Line by line" is a handy picture, but a simplification. In practice an interpreter may first translate your code into an intermediate format called bytecode. That is what CPython, the standard Python, does: it compiles your program to bytecode, which a virtual machine then executes.

Why have both? Interpreting is slower, because the interpreter runs many machine instructions for every step of your program. A program compiled to native code tends to run faster, and bytecode setups land in the middle. But faster development can outweigh faster execution. With an interpreter the edit-build-run cycle is short, which matters most when you are prototyping and testing. Python's developers typically prioritize readability over performance, and when speed really matters, hot spots can move into extension modules written in C.

Interpreters have another gift: portability. Code that runs in an interpreter can run on any platform that has a compatible interpreter, so you do not need to build an executable for each one. Most Python setups also offer a read-eval-print loop where you type one statement and get the answer immediately.

The idea has a charming origin. Lisp was the first interpreted high-level language. Steve Russell read John McCarthy's paper and realized, to McCarthy's own surprise, that the Lisp eval function could be written in machine code.