I seem to bleed code. When asked a question or confronted with a problem, I almost immediately try and get a computer to do it for me. This is a useful instinct because it helps sharpen all my abilities and whatnot, but it turns out to not be the point of graduate school. In grad school in CS, the point is to bleed not code, but papers. Erudite discussions of the problem, not just a quick hack to explore the problem space. So now I must train myself so that, whenever I write some code to solve a problem, I also create a document talking about the problem, it’s history, my solution and the solutions of others. Ideally this would take 8 to 12 pages per problem so that I could submit it to a small conference, but really just forming this as a habit is the most important thing.
So I expect I will as a consequence of this be bleeding slightly less code, but I should also be making better and faster progress towards the unspoken “acculturation” requirement of my PhD.
