Although they are identical in arrays, I prefer a[i] for the following reason. Suppose you have code like this:
double gradient[3];
// 200 lines later
double dx = gradient[0]; // Works fine
double dy = *(gradient + 1); // Works just as well
Then somebody decided that std::vector looks better than double[]:
vector<double> gradient;
// 200 lines later
double dx = gradient[0]; // Still works
double dy = *(gradient + 1); // Ouch!
Then somebody else decided to go object-oriented all the way:
class Gradient
{
public:
double operator[](int index) const;
};
// 1000 lines later
Gradient gradient;
// 200 lines later
double dx = gradient[0]; // Still works
double dy = *(gradient + 1); // Ouch!!!
Thus, from the viewpoint of maintainability, I prefer to say exactly what I mean, without obscuring the intent with context-dependent synonyms. If I mean "take i-th element of a" then I say exactly that in C++ lingo:
a[i];
rather than relying that for the particular implementation of a one can use a fancier expression for the same.
