Memory edits the story
Research looks cleaner in retrospect than it felt in real time. Once a solution works, failed branches disappear from memory and the final idea starts to feel inevitable. A notebook preserves the mess before hindsight rewrites it.
I try to record three things: the question in plain language, the evidence that would change my mind, and the next smallest experiment. This is enough structure to resume work without turning the notebook into another project to maintain.
Failed attempts are data
A failed attempt can eliminate an entire family of ideas, but only if the reason for failure is recorded. “Didn’t work” is rarely enough. Was the assumption false? Was the code wrong? Was the effect smaller than the measurement noise?
The notebook becomes most valuable months later, when a new problem resembles an old dead end. A short, honest note can save a full afternoon.
Writing changes the work
The act of explaining a question exposes gaps that silent thinking can glide past. This blog is the public, edited edge of that process. The notebook is the private working surface underneath it.