Java for computer science students in Peru

Peru has many universities to choose from: SUNEDU, the national university regulator, reported 94 licensed universities in its press note of 15 September 2021. We are not connected with SUNEDU, the Universidad de Lima or any other university, and we do not award credits. Our classes run alongside the degree, explaining the ideas your courses assess. Students weighing up Python and Java together can also read cursos de programación para universitarios.

import java.util.ArrayList;
import java.util.Comparator;
import java.util.List;

public class Ranking {
    record Entry(String name, int solved, int minutes) {}

    public static void main(String[] args) {
        List<Entry> board = new ArrayList<>(List.of(
            new Entry("Valeria", 4, 210),
            new Entry("Diego", 5, 260),
            new Entry("Ana", 4, 180),
            new Entry("Bruno", 5, 260),
            new Entry("Camila", 3, 95)
        ));

        board.sort(Comparator.comparingInt(Entry::solved).reversed());
        System.out.println("By solved only:");
        board.forEach(e -> System.out.println("  " + e));

        board.sort(Comparator.comparingInt(Entry::solved).reversed()
                .thenComparingInt(Entry::minutes)
                .thenComparing(Entry::name));
        System.out.println("Solved, then minutes, then name:");
        board.forEach(e -> System.out.println("  " + e));
    }
}

The fix is a chain: thenComparingInt(Entry::minutes) settles ties on solved problems, and thenComparing(Entry::name) settles the rare full tie. Now Ana correctly passes Valeria, and Bruno and Diego appear in alphabetical order whatever order they were entered in. A stable sort keeps your input order for ties; only your comparator can replace it with a rule.

The Comparator documentation adds the contract behind all of this: comparing x with y must give the opposite sign to comparing y with x, and the relation must be transitive. Comparators that break the contract can make sorting behave unpredictably, which is why we teach students to build them from comparingInt and thenComparing rather than hand-written subtraction. The scoreboard and names are ours, invented for the exercise.

1. Predict. Swap Diego and Bruno in the input list and run the first sort again. What changes, and why does the second sort not change at all?
2. Reverse one key. Change the rule so that more minutes ranks higher. Which single call do you edit, and what happens if you put reversed() at the very end of the chain instead?
3. Prove stability. Sort first by name, then by solved problems only. Explain, in English, why the names inside each group are now alphabetical.
4. Scale it. Generate 100,000 random entries and time both sorts. Then read the complexity of List.sort in the documentation and say whether your timings agree.

Source: tz database (IANA), America/Lima and Asia/Kolkata. Conversions are ours and do not change during the year.

Mentors explain in clear, measured English and type the key terms into the chat, so a student can review them later. Contest problems, API documentation and most of the CS literature are in English too, so practising technical English in class pays off well beyond the course.