I'm currently working on a PhD dissertation part time, and this concept has been very helpful since I'm not working full days and therefore have more context switching. I end up drawing a lot of graphs and figures to think things through, so a big notebook of graph paper ends up working the best for me. Every couple of days or so I write out a list of the next few tasks as I currently see them, then think on paper to figure out how to implement the code for each task.

Funny thing is, writing it down helps it stick in my brain, so I need the write up less than I would if I didn't write it. That's got to be some kind of contrapositive of Murphy's law or something.