I started writing my thesis this week. The code I've been working on for ages is finally giving me some reasonable results. I think I will need to code a few more things in order to demonstrate certain results more clearly, but I'm feeling ok about everything.
My goal was to write 5 pages per day. While some suggest aiming for something more like 2 per day, I knew I was on a tight schedule for finishing things up and graduating in time to start a PhD in September (fortunately, at the same school). Considering that I lost about a day in total fiddling with my LaTeX templates and getting my preferred editor LyX up running on my Vista laptop, I managed to actually reach my goal by writing just over 20 pages (including appendices) in the four days I had remaining. Will things get harder later? It's hard to tell. I will have more results images and graphs for a couple of the chapters, but as mentioned, I may need to write some more code, too.
One technique that I have found useful is using a simple little kitchen timer to force myself to work for one hour, then force myself to take a break for ten or fifteen minutes. An hour seems to be just long enough to get something accomplished, so when the timer goes off, I don't just ignore it to finish some little thing. (I used to do this a lot when my timer was set to something shorter, and that some little thing often took enough time to cut into the next break.) During breaks I get food, go outside, tidy up the bedroom, read a leisure book... whatever. Anything that isn't thesis writing helps clear the mind.
When everything is completely finished, approved, defended, etc, I will post more about the actual results.