595/700 Guidelines

  1. You must take CSS 509/510. Rare exceptions may be made, but you will have to prove to me that your technical writing skills are superb.
  2. Your proposal should be thorough enough that I understand the background of what you are going to do, why you are going to do it, what you are going to do and how you will evaluate if you have completed your project sufficiently. I want to see that you have thought this through and have a plan. Make sure that your proposal looks professional. I do not want to see a poorly organized document. Make sure your schedule is in the form of a table and is easy to read.
  3. Make sure you include a schedule of what you plan to do on a 1-2 week granularity
  4. Estimate the time you expect to spend on each task at a # of hours granularity. The capstone or thesis should be 400 hours total
  5. Make sure to include the defense date in your schedule
  6. Make sure to include a final report in your schedule
What is the difference between the Project and the Thesis?

I differentiate by how you will be evaluated. For the project, you will be evaluated primarily based on the completeness of your design and implementation. You still need to evaluate your work and justify your design decisions. For the thesis, you will be evaluated primarily on contributing to the advancement of scientific knowledge. This means that your results need to improve the state of the art in some way. You will still need to do some implementation to run your experiments, but in many cases, a complete implementation will not be necessary to get your results. Since I mostly build systems, most capstones my students do could be considered for either a project or a thesis.

For Thesis
  1. I am only chairing theses for students that have regularly attended my research meetings (contact me if you would like to begin attending them) or worked with me on a 600 project. There may be exceptions on a case-by-case basis, but these are rare.
  2. You must use LaTeX.
  3. You must store your thesis and any code you write in a private VCS that you share with me.
  4. The thesis must be spread over at least two quarters.
For Project
  1. You'll need to convince me that the project is worth doing. You must present and defend your motivation, design, and project plan to me.
  2. This option must be spread over two quarters. Exceptions may be made if you've already worked on this in a class with me or in 600 with me.
Questions you need to be able to answer by the time you submit your proposal
  1. What is the problem? (note that you don't have to solve everything, you can solve a small part of a bigger problem)
  2. Why does it matter? (you need to justify that there's a reason why solving the problem is useful)
  3. Why does existing work not solve it? (show me that you've searched to see that this hasn't already been done)
  4. Why is the problem hard? (your problem can't just be a trivial addition to existing work)
  5. What is your new approach? (capstones must create something new)
  6. Why do you think it is going to work? (you've run some experiments already or you can cite a paper that has)
  7. How are you going to demonstrate that your approach is better? (you need to collect evidence of the validity of your solution)
Questions you need to be able to answer by your defense
  1. What is your contribution? What did you do that is worthy of being an MS Thesis/Capstone?
  2. What measurements are you taking? Part of your paper will be showing how "good" the system is. What are you using for comparison? How do you know that it is successful?
  3. What are the limitations? What could be done to improve those if you had more time? What are alternatives decisions you could have made as you developed this and why didn't you choose those options? We want to see that you thought through your work and that you made engineering decisions. This is one of the points of receiving an MS degree is the implication that you have rigorous decision making process within the field of your expertise.