Each assignment will have an FAQ linked at the top. You can also access it by adding “/faq” to the end of the URL. The FAQ for Project 1B is located here.

Introduction #

In Project 1A, we built LinkedListDeque. Now we’ll see a different implementation of the Deque interface that uses a backing array, rather than linked nodes.

By the end of Project 1B, you will…

  • Gain an understanding of the implementation of a backing array in data structures.
  • Have more experience using testing and test-driven developoment to verify the correctness of these data structures.

Check out the Project 1B slides for some additional visually oriented tips.

We will provide relatively little scaffolding. In other words, we’ll say what you should do, but not how.

It should (still) go without saying that you may not use any of the built-in java.util data structures in your implementation! The whole point is to build your own versions! There are a few places where you may use specific data structures outside of tests, and we will clearly say where.

Style #

As in Project 1A, we will be enforcing style. You must follow the style guide, or you will be penalized on the autograder.

You can and should check your style locally with the CS 61B plugin. We will not remove the velocity limit for failing to check style.

Getting the Skeleton Files #

Follow the instructions in the Assignment Workflow guide to get the skeleton code and open it in IntelliJ. For this project, we will be working in the proj1b directory.

You see a proj1b directory appear in your repo with the following structure:

├── src
│   └── Deque.java
└── tests
    └── ArrayDequeTest.java

If you get some sort of error, STOP and either figure it out by carefully reading the git WTFs or seek help at OH or Ed. You’ll potentially save yourself a lot of trouble vs. guess-and-check with git commands. If you find yourself trying to use commands recommended by Google like force push, don’t. Don’t use force push, even if a post you found on Stack Overflow says to do it!

You can also watch Professor Hug’s demo about how to get started and this video if you encounter some git issues.

ArrayDeque #

As your second deque implementation, you’ll build the ArrayDeque class. This deque must use a Java array as the backing data structure.

You may add any private helper classes or methods in ArrayDeque.java if you deem it necessary.

Creating the File #

Start by creating a file called ArrayDeque. This file should be created in the proj1b/src directory. To do this, right-click on the src directory, navigate to “New -> Java Class”, and give it the name ArrayDeque.

Make this class implement the Deque interface following a similar process to one described in the Project 1a spec.

Now you’re ready to get started!

Writing Tests #

Refer to the Project 1A spec for a review of how to write tests. Similar to Project 1A, you will be scored on the coverage of your unit tests for Project 1B. You might find some of your tests from Project 1A to be reusable in this project; don’t be afraid to copy them over!

Constructor #

You will need to somehow keep track of what array indices hold the deque’s front and back elements. We strongly recommend that you treat your array as circular for this exercise. In other words, if your front item is at position 0, and you addFirst, the new front should loop back around to the end of the array (so the new front item in the deque will be the last item in the underlying array). This will result in far fewer headaches than non-circular approaches.

See the Project 1B demo slides for more details. In particular, note that while the conceptual deque and the array contain the same elements, they do not contain them in the same order.

Task: Declare the necessary instance variables, and implement the constructor.

The starting size of your backing array must be 8.

addFirst and addLast #

As before, implement addFirst and addLast. These two methods must not use looping or recursion. A single add operation must take “constant time,” that is, adding an element should take approximately the same amount of time. This means that you cannot use loops that iterate through all / most elements of the deque.

Don’t worry about resizing your array now, we’ll get to that later.

Task: Implement addFirst and addLast for deques with fixed size backing arrays.

get #

Unlike in LinkedListDeque, this method must take constant time.

As before, get should return null when the index is invalid (too large or negative). You should disregard the skeleton code comments for Deque.java for this case.

Task: Implement get.

toList #

toList will continue to be useful to test your Deque.

Write the toList method. The first line of the method should be something like List<T> returnList = new ArrayList<>(). This is one location where you are allowed to use a Java data structure.

Some later methods might seem easy if you use toList. You may not call toList inside ArrayDeque; there is a test that checks for this.

Task: Implement toList. You are not given tests this time, so you will need to write them!

All that’s left is to test and implement all the remaining methods. For the rest of this project, we’ll describe our suggested steps at a high level. We strongly encourage you to follow the remaining steps in the order given. In particular, write tests before you implement the method’s functionality. This is called “test-driven development,” and helps ensure that you know what your methods are supposed to do before you do them.

isEmpty and size #

These two methods must take constant time. That is, the time it takes to for either method to finish execution should not depend on how many elements are in the deque.

Task: Write tests for the isEmpty and size methods, and check that they fail. Then, implement the methods.

removeFirst and removeLast #

Lastly, write some tests that test the behavior of removeFirst and removeLast, and again ensure that the tests fail.

Do not maintain references to items that are no longer in the deque.

Resizing #

Before moving on to this section, you should test your code to make sure that everything works for Deques that don’t need to resize the backing array.

Resizing Up #

The exception to the “constant time” requirement is when the array fills, and you need to “resize” to have enough space to add the element. In this case, you can take “linear time” to resize the array before adding the element.

Correctly resizing your array is very tricky, and will require some deep thought. Try drawing out various approaches by hand. It may take you quite some time to come up with the right approach, and we encourage you to debate the big ideas with your fellow students or TAs. Make sure that your actual implementation is by you (or your partner) alone.

Make sure to resize by a geometric factor.

We do not recommend using arraycopy with a circular implementation. It will work, but results in a significantly more complex (and harder to debug!) implementation than necessary.

Instead, we suggest thinking about using get and a for loop in some way.

Task: Modify addFirst and addLast, and verify that they are correct using your tests and the Java visualizer. Make sure to add enough elements so that your backing array resizes! For more info on resizing, check out these slides.

Resizing Down #

The amount of memory that your program uses at any given time must be proportional to the number of items. For example, if you add 10,000 items to the deque, and then remove 9,999 items, you shouldn’t still be using an array that can hold 10,000 items. For arrays of length 16 or more, your usage factor should always be at least 25%. This means that before performing a remove operation that will bring the number of elements in the array under 25% the length of the array, you should resize the size of the array down. For arrays under length 16, your usage factor can be arbitrarily low.

We, again, do not recommend using arraycopy with a circular implementation. If you followed our advice above to use a for loop to resize up, resizing down should look very similar to resizing up (perhaps a helper method?).

Task: After you’ve written tests and verified that they fail, implement removeFirst and removeLast.

For the intended experience, follow these steps in order. If you do something else and ask us for help, we will refer you back to these steps.

Submit to the Autograder #

Once you’ve written local tests and passed them, try submitting to the autograder. You may or may not pass everything.

  • If you fail any of the coverage tests, it means that there is a case that your local tests did not cover. The autograder test name and the test coverage component will give you hints towards the missing case.
  • If you fail a correctness test, this means that there is a case that your local tests did not cover, despite having sufficient coverage for flags. This is expected. Coverage flags are an approximation! They also do not provide describe every single behavior that needs to be tested, nor do they guarantee that you assert everything.
  • If you fail any of the timing tests, it means that your implementation does not meet the timing constraints described above.

Scoring #

This project, similar to Project 0, is divided into individual components, each of which you must implement completely correctly to receive credit.

  1. Adding (25%): Correctly implement addFirst, addLast, and toList.
  2. isEmpty, size (5%): Correctly implement isEmpty and size with add methods working.
  3. get (10%): Correctly implement get.
  4. Removing (30%): Correctly implement removeFirst and removeLast.
  5. Memory (20%): Correctly implement resizing so that you do not use too much memory.

Additionally, there is a test coverage (10%) component. We will run your tests against a staff solution, and check how many scenarios and edge cases are tested. You can receive partial credit for this component.

In total, Project 1b is worth 12 points.

Next Section #

Congrats on completing Project 1a! We really suggest that you take a break before you move on to Project 1c.