Things to cover at a thesis defense?

Depends on your own personal grading of the project. My (bachelor's) project was theoretical. By the end I was a little skeptical of my theoretical results because of the scale of the project. Indeed, I had about as many hand-written proofs as a good student should remember for an Algorithms and Data-Structures exam.

So I wrote a compiler demoing my theoretical results, discovering subtle mistakes in my theory along the way (10-day Haskell coding spree :) ). In my presentation I presented the compiler (it generated pretty graphs, rather than text), and those mistakes. In the end I got two questions:

1. Very high level. Couldn't it have been easier to do my proofs if I knew Category Theory? Yes, but I didn't know it at the time.
2. Terribly picky. A subtle hick-up in one of the proofs. Turned out I was simply not being formal enough.

Otherwise I attempted to show the general qualities of someone with a BSc degree in Computer Science, rather than show that I fulfilled the goals specified for the course "bachelorprojekt"[0], somehow this seemed like a better strategy. For instance, I ended up discussing complexities, although my project was largely about feasibility in general.

So I guess my advice is know your project and your report inside out, and show that you're BSc quality when reflecting upon the handed in work and your results.

As a last note, make sure you both have a projector and a board to write on. You'll need both.

P.S. I got a 12.

[0] http://sis.ku.dk/kurser/viskursus.aspx?knr=142725