On keeping a lab notebook after ten years
The habits that stuck were the boring ones. Dated entries, no tags, one file per week, append-only. Everything else turned into ceremony and got quietly dropped within a month or two of adoption.
I think the mistake I made, for a long time, was treating the notebook as a piece of software rather than as a piece of writing. I kept wanting to structure it: schema the entries, enforce tagging, build a little tool to query it. Every time I did this, I stopped writing in the notebook. Every time I dropped back to plain prose in flat files, I started writing in it again.
The version I've had for about three years now is almost embarrassingly simple. One Markdown file per week, named by ISO date. At the top, a few lines about what I'm trying to do. Underneath, dated paragraphs as things happen. I go back and search it with grep. There is no index, there are no tags, and I have never once wished there were.
The part that I did not predict is how much the notebook helps during the weeks when nothing is working. It is easier to keep writing when you have been writing for a while. The notebook, by existing, asks to be continued.