‘Tis the receptacle of our tinkering of the world, out of which we distill and crystallise knowledge.
Leonardo da Vinci's notebook / British Library
Whether one does research in academia and industry, the lab notebook is a constant companion. In fact, in industry the notebook recording would be even more enforced and regulated due to IP and patent and stuff, while in academia it tends to be more lax; or at least that is my impression anyway.
Now, for those doing computational work, how best to log and record our work? The wet lab notebook is definitely not a good fit, since we have some code snippets here and there, which are not so amenable to longhand. In my initial PhD years doing structural biology work, I was happy to adopt OneNote at the suggestion of my PI, who was a fan. It worked fine, but somehow when I transitioned to more computational work, I abandoned it. Perhaps it is too bloated with features. Perhaps it’s the change to Linux desktop. But there is the web version, you say; I know, I know, perhaps it’s just my lack of discipline after all…
Anyway, during my latter PhD years in Sweden I reverted back to pen and paper, plus some README files at the project folders. And this mostly worked fine, but looking back, I wondered how in the world I could find all my results back if asked by a reviewer or an auditor? I didn’t worry much about this since all my work during this time are already published, but I do wonder if in the future I can adopt a more organised and robust method of documentation.
With this in mind, I was happy to find Jupyter notebook in the beginning of my postdoc earlier this year. It is a notebook in which one can mix text (Markdown format, plus html and LaTeX if you need them) and executable code snippets – perfect for computational work, right? A few months ago I would nod enthusiastically, but here I am again with no update to my Jupyter notebook for months.
Don’t get me wrong, Jupyter notebook is very useful and I still use it, but not for my big picture notebook. I find it better to have one notebook of say, a particular data processing from raw data with some docking scores, to pretty graphs, to final selection to be passed on to a collaborator. I think Jupyter notebook would better fit work that is more scripting-heavy e.g. non-structural bioinformatics and/or deals with an enormous amount of data e.g. data science, astronomy. (See: Nature | Why Jupyter is data scientists’ computational notebook of choice)
I also still use pen and paper as scratchpad: brainstorming, sketching ideas, immediate todo list, doubts, and other unpolished stuff. I would not abandon pen and paper yet, since the current research says that pen is mightier than the keyboard (but I will copy-paste that code snippet, thank you).
Interestingly, what sticks for me as a big picture notebook is… wait for it… presentation slides! Wait, wait go back, don’t go away.
I’ve learned during my PhD that my weekly update with my supervisor keep me on my toes and give me focus on the direction to go. I think putting my progress in slides puts me in that state of mind of preparing to do progress update meeting with my supervisor, so it forces me to create a logical narrative and spot the holes. As a bonus, that work progress slide you need to make for lab meeting? You now will always have it on hand, constantly updated. In restrospect, this makes the most sense – you log and record your work in a notebook, then translate it to slides to communicate it to your colleagues and supervisor; why not skip the notebook altogether?
Another thing I want to mention, is that I also have a script that nags me every hour to log what I am doing at the hour, with the script automatically generating the timestamp. Besides jolting me out of procrastination, it helps me track of the distribution of work among the projects over the weeks, for example. I keep the entries simple, and I don’t fill every hour, so it comes down to a few entries per workday.
If you would like a similar setup, you can easily set one by setting hourly reminder and log the entries in your Google or Outlook calendar (I keep mine in a plain text file). You can see how this log file looks like below:
In Part 2, I will talk about alternative options out there and a new lab notebook setup that I am trying out.