You're right, sort of. Because C doesn't have proper arrays, you make up for it by passing a pointer to the first element and the range. So for C to do as well

You're right, sort of. Because C doesn't have proper arrays, you make up for it by passing a pointer to the first element and the range. So for C to do as well as Fortran on matrix operations the compiler developers need help. One source of help is that if you have a pointer to double and a pointer to int, you can assume that writing pdouble[k] doesn't alter anything reachable as pint[m].