int find_largest (int data[], int n) {

  int i, largest;

  // start of by assuming that the first value is our largest value.
  // this is cleaner than using -INT_MAX (which does work) and much
  // better than using an arbitrary value like zero (which doesn't work
  // if all of the values are less than the value chosen)
  largest = data[0];

  // now consider all of the remaining values (note: i starts at 1)
  for (i = 1; i < n; i++) {
    if (data[i] > largest) {
      largest = data[i]; // we have a new champion
    }
  }

  return largest;

}


void analyze_temperatures (double temperatures[], int years,
                           int &big_ups, int &big_downs) {

  int i;
  double increase;

  big_ups = big_downs = 0;  // nothing seen so far

  for (i = 0; i < years - 1; i++) {
    // the following will cause a problem if temperatures[i] is zero,
    // and will produce a nonsense value if temperture[i] is negative.
    // this was a flaw in the question (in retrospect I should have
    // specfied Kelvin temperatures) and the issue is ignored here.  instead
    // it is simply assumed that all temperatures are greater than zero.
    // note the order of operations and the use of 100.0 instead of 100
    // (there is a danger here of getting integer division and this isn't
    // what is called for).
    increase =
        ((temperatures[i + 1] - temperatures[i]) * 100.0) / temperatures[i];
    if (increase >= 2.0) {
      big_ups++;
    } else if (increase <= -2.0) {
      big_downs--;
    }
  }

}