In certain contexts, you have to treat a vector as an array of its elements. For example, suppose you have a function that takes int * as its argument. How can you pass a vector of int's as an argument to this function? The simplest solution is to take the address of the value returned by front() or operator. Because both these member functions return a reference to the internal array of elements, you can use this address as the beginning of an array. For example:
void func(const int arr, size_t length );
vector < int > vi;
//.. fill vi
vi is the first element of the vector's internal array. You can treat the address of this expression as an array of int's under two conditions: func() doesn't access out-of-range array elements, and the elements inside the vector must be stored in contiguous memory. Although the C++ Standard doesn't guarantee that yet, I'm not aware of any implementation that doesn't use contiguous memory for vectors. Furthermore, this loophole in the C++ Standard is about to be fixed within weeks; the Standard will soon guarantee that vectors use contiguous memory. Therefore, this technique of transforming a vector to a bare array is safe and portable.