How many errors are there in source code when writing a program?

Answer 1

Ideally, none. In practice, however, a program may contain many errors. The language editor should help prevent common syntax errors through "intelisense" technology (similar to spell checking and grammar checking tools in word-processing software), as well as colour coding and auto-formatting. The compiler deals with compile-time errors while the linker deals with linkage errors. However, runtime errors are the most difficult to track down as these require extensive runtime testing to detect before they can be eliminated, which requires additional testing to ensure there are no breaking changes. Test-driven development can help minimise the amount of runtime testing required. This works on the premise that every function in a program can be tested, programmatically, to ensure the expected results prevail over the full range of possible inputs (including errant input). Static and non-static assertions can also be used to test our assumptions always hold true during debugging. Even with these safe-guards, we cannot guarantee that every error is eliminated. The larger and more complex the program code, the greater the likelihood that undetected errors will slip through the net. However, code that easy to read is also easy to maintain; sticking to best programming practices and making good use of RAII and exception handling mechanisms makes it much easier to trace runtime errors.