No you're quite right, not least since my insertion routine traverses the whole list on each call. I didn't write this for speed of execution, merely speed of writing it :)
For speed of writing, here's the whole thing in C++ using STL:
#include <vector>
#include <algorithm>
#include <iostream>
#include <cstdlib>
#include <ctime>
using namespace std;
int main()
{
srand(time(NULL));
vector<int> array; // original array
vector<int> unique; // each unique element from
original included only once
enum {SIZE=100};
for (int i=0; i < SIZE; ++i) // populate with random ints
array.push_back(rand()%50);
// remove duplicates:
sort(array.begin(),array.end());
unique_copy(array.begin(), array.end(), back_inserter(unique));
cout <<"Unique size: " <<unique.size() <<endl;
return 0;
}