There are a bunch of activities that I engage in when doing research. These include but are not limited to:

• Figuring out the best thing to do.
• Talking out loud to force my ideas into language.
  • For the last 3 months I have been working maybe 50 hours per week by meeting with people and doing stream-of-thought reasoning. That was very productive. Probably in large part because of this.
  • Even when working alone I try to use this. The main thing that holds me back from using it all the time when working alone is that it can be quite awkward.
• Recording myself explaining something, usually on a whiteboard. This is useful to check:
  • Check if my understanding is good enough yet to write a post.
  • Helps remove the awkwardness when talking to yourself (because you are not).
• Trying to explain an idea on the whiteboard.
  • I mainly use whiteboards when I am still at the stage of being confused.
• Writing pseudocode.
  • Similar to forcing yourself to explain something in natural language.
  • Notice where you are confused by not being able to express something.
• Writing a concrete implementation we can run.
  • I rarely do this because it is so slow, probably because I have not acquired sufficient software engineering skills yet.
  • I expect that writing programs can be very useful for getting observations that you could not easily generate in your head. E.g. Mandelbrot did make plot fractals.
• Writing down things that we have figured out on a whiteboard or any other process in rough notes.
• Writing a distillation of the thing I have figured out, such that I can understand these notes 1 year from now.
• Reflecting on how it went.
• Writing public posts, that convey concepts to other people.

My main questions are:

• What research processes do you use?
• When do you use them?
• What do you get out if it goes well?

Also, feel free to mention great posts about this. I am most interested in processes that you personally use on a regular basis.