Organizing a Thesis

Organizing your thesis is pretty simple when you get down to it. Don't make it more complicated than it needs to be. If you are confused about what to do, you might consider following the structure recommended below. This comes via way of Brian Smith, a former MIT faculty member now at Penn State. He was given this advice by Bruce Buchanan. I think it's reasonable as well. 

Remember that: 

  1. Writing takes time to do well
  2. You will need to rewrite sections. Writing is a learning process. It forces you to come to terms with your own ideas and to be precise about what you have done. Think of the writing as an evolutionary process and don't get upset when you need to rewrite sections and reorganize large chunks. 
  3. Your advisor will have high expectations and will not allow you to throw something together last minute and declare that it is done. 

Prof. Buchanan's advice is that any engineering thesis can be reduced to the following six chapters. Sometimes other structures can be warranted, but you can't go wrong with the structure below. 

Chapter Contents
1) Introduction What's the problem you're trying to solve? Argue why your work has a surprising, unexpected result or solves a problem that nobody has solved before or can be proven to work substantially better than any other solution anyone has tried or raises some questions that nobody has thought about before.
2) Extended Example Give an example of your system and how it works. 
3) Theory & Rationale Why are you doing what you're doing? Concisely but comprehensively, what is the prior work that your work is based upon or that your work supercedes? You are justifying your decisions based on prior work and arguing why your approach is needed, but you are not providing a full literature review. Look in the background sections of published papers to see how this compare and contrasting is typically achieved. You are connecting your work to the work of others and setting up your discussion of design decisions in the next section. That said, all the important prior work should be references. You just don't need a summary of each work. Only what makes it relevant. 
4) Design & implementation How was what you built built and why? What have you improved upon as described in Chapter 3? I
5) Evaluation How well does your system actually work with respect to solving the problem you lay out in Chapter 1? Explain this, being extremely carefully in how you document the experiments you have performed. How do you know that what you have created actually works?  Analyze the problems that you encountered: both those you solved and those you didn't. 
6) Conclusion Summarize your main arguments and how you have supported your hypotheses. What would you do differently if you were going to take another pass at this problem. You do not need an extensive section on pie-in-the-sky future work. 

In addition, I would suggest that for a MEng/SM thesis these chapters, in total, should be approximately the length of a long conference paper submission. Extra implementation detail should be included in an extensive set of appendices.


Stephen Intille's Thesis Development and Writing Tips

Last updated: 11/16/04