Is defining the length of an array with a macro "excellent practice"?

Question

I'm working through a book called C Programming: A Modern Approach and in the first section discussing arrays, the author states:

using a macro to define the length of an array is excellent practice

Then uses the brief example:

#define N 10
...
int a[N];

I understand that it has something to do with being able to go back into the source code of the program and change the value, and making it a macro maybe makes it easier for the programmer, but I'm not certain. Why is this an excellent practice, or is it objective?

Answer 1

It's a good practice because

• the array size obviously has to be hard-coded
• but it shouldn't have a magic number injected directly into the source code
• therefore, a macro is a good way to give it a readable name and remove it from the source

That being said I'm not sure I agree this is the best way. An enum also works and avoids some of the problems with macros (e.g. harder to overwrite and silently compile). And IIRC a const int works as well.

For reference this compiles with cc:

const int s = 1;
int a[s];


int main() {
return 0;
}

Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn)
Target: x86_64-apple-darwin12.4.0
Thread model: posix

Answer 2

It is a very good practice, the C language specification itself says to NEVER bury constants into code, but to define them with meaningful names. There are a few ways to do it, macros (my personal favorite since they use no memory), globals (use memory and can be modified), constant globals (use memory but never change).