Suggestions for Summer C.S. activities

Are you appalled at the thought of an entire Summer break with no programming? Or just bored by the lack of homework? Or even possibly wanting to get a headstart on the next semester? If so, I have some suggestions for you:

• For a language you have already learned, brush up or practice it, either with some homework exercises from the textbook, or another book you have access to, or from something you find on the net that sounds interesting, or from these projects.
• Learn a new language.
  • If you have not yet completed CS110, Introduction to C++, then you should probably first concentrate on C++, because the Redlands programming sequence is based on it.
  • Java
    • A worthy alternative to C++ for any except the most compute-bound problems.
    • You can write GUI programs that work on any platform.
    • You can write applets that run inside a browser.
  • Python: www.python.org/ IDE included. Huge libraries available. Quick start to writing Python code. Advanced mini-tutorial.
  • Haskell: haskell.org/ a functional language. See how to work with infinite sets!
  • Lisp/Scheme: www.plt-scheme.org/ the pioneering functional language. Write functions which create new functions!
  • Prolog: www.swi-prolog.org/ Logic programming. Write functions which can compute their own inverses!
  • Perl: www.perl.org/ often used to build dynamic web pages, and for housekeeping tasks. Huge libraries available.
• Professionalize your tools:
  • CVS: Maintain audit trail of your source code [required for CS220].
  • Linux: get an account on our Linux box [required for CS230, useful for CS220 and CS340].
  • SSH: Secure connections between machines
  • PuTTY: SSH for MS-Windows [required for CS230]
  • Server maintence: some students may be allowed to participate in the care and feeding of our Linux box: it needs to support Apache and MySql.
  • Databases are important.
• Analyze the interaction of computer science and society:
  • Should reliability requirements of medical support systems, voting systems, credit reporting and other systems with potentially high impact on individuals influence the ethics and morality expected of computer scientists? Can/should government require/sanction such behaviors?
  • Impact of encryption on speech, crime, opposition to government, e-commerce, intellectual property.

Some projects to consider

Examples and code for many of these can be found on the web. You will get the most out of the exercise if you plan your approach and debug your code before you look at someone else's solution. On the other hand, reading other people's code is often illuminating, and is a worthwhile skill to acquire.

• Don't get sued: Honor copyright and licenses.
• Be honest: Provide proper source attribution.
• Remember that you must understand any code that you submit for a grade.

Sieve of Eratosthenes:

great practice with arrays, loops, and integer arithmetic.

Shuffle:

To shuffle a deck of cards correctly is a bit subtle. Try your hand before peeking at the standard solution.

Number theory:

If you like number theory, you will love cryptography. To get serious with it, you'll need to use a "large integer" package. Random number generators fit here, too.

Buffton's needle:

compute pi with only straight lines, and NO trig functions!

Hangman:

grab a dictionary from the net.

Poker:

evaluating a poker hand is quite tricky. Define a class to represent a card, and a class that collects cards into hands.

Tic-tac-toe:

Fairly easy to code after you decide on a representation for the board.

Four-in-a-row:

Implementation is a lot like tic-tac-toe, but it does not have such an immediate complete winning strategy.

Battleship:

Can be done at many levels. See following game comments.

Series:

Keep track of wins and losses over multiple game.

Game graphics:

Take some text-mode game that works, and display it nicely: pictures for card games, hangman pictures, fancy tic-tac-toe layout, etc. There are lots of bitmapped pictures on the net that you can use, just be sensitive to copyrights.

Game sounds:

play an appropriate (or annoying!) noise after a hit or a miss, etc.

GUI interface:

Convert (or re-write) your game to provide a true GUI interface. This needs an event-driven program, and I don't recommend you do it in C or C++ because you will have to learn a lot of things peculiar to a specific operating system (whether MS-Windows, X-windows, or Macintosh).

Sockets:

allow opponent on a remote machine / different operating system.

Applet:

allow your game to run inside a browser (especially easy with Java).

Client-server:

allow multiple players. Don't try this with C/C++ if you have only one semester.

Symbolic computation:

create deriviative of an expression. Then simplify it. Hint: use Python, Haskell, Scheme or Prolog.

Strategy enhancements:

Write a good/optimal strategy for one of your games. This tends to be messy in an imperative language, so use Python, Haskell, Scheme or Prolog.

Machine learning:

Don't write the strategy yourself, let your program learn it. Decision trees, neural networks, and genetic algorithms can all be competitive.

Notice in the list above that there is a progression, or arc, to the game projects. That is one of the appeals of game projects: You can always find more to do. Of equal import when deadlines are imminent, you can often also cut back on the requirements for games. For instance, in checkers, start out without implementing Kings. The game itself will be less interesting, but you can still demonstrate most of the same programming competence without Kings.

In the area of game strategy there is a similar arc. In particular, when implementing a computer player you should always start out with the "random-mover" strategy. It will usually be a lousy player, but it does exercise all the remaining parts of your system. (Ensuring that it generates only legal moves is a great test of your legal-move-tester.)

Furthermore, the random mover provides a benchmark for measuring how good your advanced strategies are. In particular, if a supposedly advanced strategy is beaten by the random-mover, you have serious questions to address.

Finally, the random-mover is often the strategy to bootstrap your machine learning method from.

Notice that the arcs have a natural bifurcation:

1. increasingly attractive interfaces, and
2. increasingly competent play.