I am rapidly approaching my preliminary defense, and facing the challenge of defining my thesis topic. My advisor has given me an initial direction, but now I need to turn that into a more specific question. The hardest part is trying to figure out what is "enough" for a thesis. I can think of a vast number of relevant systems to study, and ways to study them, but I'd rather not spend 10 years working on my thesis. Because a good deal of what I'm doing involves writing code, it can be difficult to predict how long any segment will really take.
It's also difficult to predict what my advisor will consider sufficiently substantial to be a paper. I know his standards are higher than most, and so I'm likely to end up with fewer, but more substantial papers. Unfortunately, the first project I was working on doesn't have funds or motivation to continue, so I'm switching topics midstream.
It doesn't help that I find it all fascinating... My hunting through the experimental literature keeps turning up awesome materials that are very closely related to what I was originally intending to study.