Principles of Quantitative Development

5.2 BASICS OF COMPUTING

5.2.1 Development cycle

It is instructional for a developer to review the basic process of building a program from its source code. Once the developer has created the source code file, he will go through the cycle of compile–link–run–debug until he can call the program relatively bug-free. No nontrivial program can be guaranteed to be bug-free, because the elimination of bugs is rarely an absolute concept; the developer can only minimize the number and probability of bugs.

Modern integrated development environments (IDEs) hide these distinct steps of compilation, linking, etc., from the developer's view. At times, however, we have to tweak the compiler or linker options, which requires us to know what is involved at each step.

A developer writes his code almost always in a high- level language, such as C, C++, etc. These languages are human-readable, however cryptic they may look to the uninitiated. For instance, a simple declaration such as int n = 0; (in C or C++) indicates to the human developer that he has access to a memory location that he has labelled n, and that it can take integer values. Hidden from his view are the steps necessary to allocate this memory location and keep track of it, and to move a zero integer value into it. These steps are machine-dependent operation codes (opcodes), which are mere series of binary numbers, completely unintelligible to a human being.

Compilation is the process by which the source code written in a high-level ...