Scope

Your proposal is expected to draw upon some or all of the material you’ve seen in your courses, but should also involve an independent exploration of new intellectual territory. There are two broad categories of projects:

1. Those in which students pull together multiple topics from across courses and (hopefully) explore some new domain as part of constructing a large project. For example, you might build a tool that uses networking techniques to fetch content from web pages, stores it in a database, and then applies a machine-learning algorithm to look for patterns or meaning.
2. Those involving an in-depth investigation of one particular topic but to a much greater extent than you’ve seen it in class. For example, you could propose an improved algorithm for training a neural network, for example, then implement it and rigorously compare its effectiveness to existing approaches.

Regardless of the approach you take, you’ll be expected to write a final report and make an oral presentation at the end of the semester.

Teams

In the real world, programmers and computer scientists work in teams to solve problems. We expect that the vast majority of projects in this class will be done by teams of 2-4 people. Not only will this let you work on your teamwork skills, but it means that you can tackle a larger and more interesting project, of which you will be prouder.

Of course, there are some situations in which a team isn’t the right approach. For example, if you’re doing a “depth-style” research project, it might make more sense to go alone. You might also find that you’ve got a project that you really want to complete but that doesn’t get anyone else excited. However, for the most part you’re expected to work in groups, and you’ll have to justify going solo in your proposal.

Ideas

Start by thinking back on the projects and assignments you’ve done so far. Were any of them particularly enjoyable? What did they have in common? Alternatively, is there anything in one of your classes that you wish you had learned?

Another option is to look forward instead: what would you like to see yourself doing after you graduate? What kinds of projects could you work on that will give you an edge when it comes to interviewing, will give you some direct experience, or will give you a taste of what your chosen field is like.

Another option is to look for a client, and see what needs they have. You could contact your favorite charity in Tacoma and see if they need a custom-made program or app. Or you could find a faculty member or campus group that has a similar need.

Finally, don’t underestimate your fellow senior CS students. One of them might have a great idea, but is looking for partners. Alternatively, just talking to one of them might inspire you to do something related but distinct.

Games

Games are controversial, and require special thought. Some professors have a negative opinion toward games—they seem them as an uninteresting distraction. Others are more nuanced and circumspect.

My opinion of game development is more positive. I'm also aware that some of you would like to do this professionally, and having a game under your belt will give you a leg up during interviews. However, there are many aspects of game development of which you need to be aware.

• Games are extremely difficult to create, and take a lot of work. Big-studio games take literally thousands of person-hours to create. Game developers often work 60-70 hour weeks. (Game development is responsible for more breakups and divorces than any other programming job.) If you take on a game as your project, you will probably end up spending more hours than your classmates on it.
• Student efforts are often disappointing. If you don't have the time, the end result will often seem tired and cheap.
• It is very hard to create something truly new. For every Minecraft, Portal, or Braid, there are at least fifty forgettable games that have just thrown together old ideas, without doing anything truly interesting.

This is not meant to discourage you. But if you are going to take on this task, you need to do so with your eyes open.

Here are some ideas that might help make a game into an appropriate project:

• Incorporate advanced AI. Machine learning and intelligent search routines (e.g. minimax) can make your game into a very interesting project. This is most useful if you've taken AI.
• Use multithreading. Most modern games are still single-threaded, because of the inherent difficulty of creating a multithreaded game. If you have some experience with multithreaded programming, this can help your game run much more efficiently. (And impress future employers.)
• Make the game multiplayer, using networking. Negotiating network traffic in real time is tricky, but doing so will add a level of accomplishment to your game.

Past Examples

Here is a list of the projects that were done last year:

• Acoustic Snooping (2 people): use a recording of a person typing in order to recover the typed text. Used elements from AI, digital signal processing, and cryptography.
• Business Intelligence Analysis of the University (2 people): use open-source data-mining tools to analyze and predict student success. Used AI and databases.
• D&D Character Builder/Campaign Manager (1 person): create a management tool for the popular role-playing game. Used databases and networking.
• Fit N’ Lift App(4 people): create a mobile, social-media app facilitating communication between athletes and coaches. Used networking and databases.
• Imbue Reality App(4 people): an app to create an “augmented reality”, such as displaying textual information relevant to landmarks. Used databases and networking.
• Mashbot (1 person): a tool for automatically making mashups of two different songs. Used digital signal processing and advanced mathematics.
• Real-Space Game App (1 person): create an app to allow users to play “alternate reality” games while moving around in the real world. Used networking and databases.
• Snow Simulation (2 people): render and display realistic-looking and -acting snow. Used graphics and advanced mathematics.
• Tennis Scheduler (1 person): create a web-based application to manage scheduling of the university’s tennis resources. Used databases and networking.
• TransitBuilder (4 people): create a gamified pedagogical tool to teach users about the resources needed to operate an urban mass-transit network. Used databases and networking.

Getting Started

You should all start thinking about your projects now. Talk to other people in the class, to propose ideas or solicit partners. Next, run the project by me for tentative approval. Just e-mail me a paragraph or talk to me in person. Finally, when you have that, you can work up a detailed description of your idea. This must include:

• A brief summary of the project.
• A description of what will be hard or novel. Where is the complexity? Where is the Computer Science? Why is it worthy? Remember, this is the culmination of your years as an undergraduate. Make sure that it’s something of which you’ll be proud in the years to come.
• A justification of the “independent exploration” angle. What will you have to learn on your own? You might have to learn a new tool or system or algorithm. More commonly, you might have to explore independently to learn how to synthesize elements from different classes.
• A list of tools, packages, or other previous work that you'll be using.
• A tentative timeline for the project, broken down by week. What checkpoints will you set up, to ascertain your progress? Often a figure or table will help here.
• An emergency plan for what to do if you run out of time. Many projects ultimately don’t pan out; what will you do if yours is one of them? Very often, it’s useful to break your project into the “core functionality” and the “stretch goals” that you hope to accomplish (if there’s time).

You must have an approved idea before you are given an add code.