What are the difference between interpreter and compilers?

Answer 1

Compilers translate the entire source code into machine code. Each translation unit (source file) is independently converted into an object file. Once all translation units have been compiled, the object files are linked together to produce an executable file. The executable file does not require any further translation and can be executed independently of the compiler. The executable is non-portable, but will execute as fast as the architecture will allow. In order to port the program to other architectures, the source code must be recompiled from scratch by a compiler intended specifically for those machines.

Interpreters translate one statement at a time and immediately execute them before moving onto the next statement. No independent executable is produced by an interpreter; the source code must be re-interpreted every time it is run. Thus the code is dependant upon the interpreter but is not machine-specific. While interpreted languages are more portable than compiled languages, execution is greatly impaired by the need to interpret the code every time it is run.

Some languages are both compiled and interpreted. Rather than producing a machine code executable, the compiler produces an intermediate byte code which can then be interpreted. Java is a typical example. The Java source code is compiled against the Java Virtual Machine (JVM) to produce portable Java byte code that can then be interpreted by any JVM implementation. Although execution is still much slower than that of machine code, the byte code is much more compact than the original source code and gives greater opportunities for optimisations that would not be possible with more traditional interpretation.

Because interpreted languages are dependant upon an interpreter, they are only suitable for writing applications software. This includes Java. Compiled languages are much less restricted. C++, for instance, can be used to write general purpose programs, including operating system kernels, device drivers, subsystems programming as well as applications software.

Answer 2

The principle difference is that an interpreter takes a computer program written in a high-level language and converts it to machine code "on the fly"; converting and then executing it as it goes through the program.

A compiler takes a computer program and converts the entire program into machine code file, this file can then be executed at a later date.

Modern computer systems often combine the two techniques; a compiler will take a computer program and convert it into a "pseudo" machine code file. This file is then taken by an interpreter and converted to actual machine instructions to execute. The advantage of this is that the compiler doesn't have to "know" about the final target machine, providing that the target machine has an interpreter it can execute the program. Java (and hence Android) uses this system as does much of Windows (using the .NET system).