4 ms·

I've always been attracted to the idea of keeping a lab notebook for my work, but since I do engineering/scientific computing, the concepts never seem to direct

I've always been attracted to the idea of keeping a lab notebook for my work, but since I do engineering/scientific computing, the concepts never seem to directly translate. Does anyone have any resources for keeping a similarly structured log of research/work done on non hard science projects?

all2 5y ago
If you're in software engineering, Jupyter notebooks support a large variety of languages now. I keep handwritten notes in a notebook of recipes I particularly like. These are useful for making adjustments over many preparations. I guess the bottom line is being willing to enforce an organization on your notes, and then iterate towards a format that makes sense to you. For the recipe notes, time series (like journal entries) works well. Or inline. It depends on the note. Ingredient substitutions can be time series, but wrong measures specified in the recipe would be an inline correction. > engineering/scientific computing Definitely Jupyter notebooks or other "literate computing" methods. Emacs org-mode has literate programming stuff, but honestly Jupyter is easier to use.
- visviva 5y ago
  Thanks - I'm not a software engineer, but I write a lot of code anyway. I've used Jupyter notebooks to good effect in the past for smaller, self-contained projects, but they don't quite seem to fit my workflow well enough for me to use them for everything. You make a good point about picking something and enforcing it - I just have a hard time even finding something to start iterating on. Maybe I just need to be more creative about how I think about/use Jupyter notebooks...
  - rmnclmnt 5y ago
    You can use use Jupyter Notebooks like Markdown files, with the added benefit of code blocks being executable for reproducible work. Then, just throw your notebooks into MkDocs (using a plugin) and voila, you get a statically hostable version of your notebooks! Easily coupled with Git and CI/CD practices, you have a nice, easy to use setup for prodyctive work and knowledge sharing if needed!
  - all2 5y ago
    > something and start iterating Go with a standard time-series journal. This is about the simplest way to do things. It works well enough for me with minor additions. Eventually you've referred back to your journals enough to know what kinds of information you'll be looking for. I've gotten part way there, so far. I'm still trying to sort out details for a lot of stuff; where does the plan of a project live? In the journal? In the projects files? What about the day-to-day entries? What about troubleshooting epilogs? Like I said... I'm still trying to sort out where each of these live. I'm tempted to just throw all the things in project folders (because they're all projects anyway) and then grep for stuff if I need to find it.
    - all2 5y ago
      But putting everything in one folder still doesn't satisfy my itch to see my journal from two views: time-series (which answers "what did I do X days ago?") and project collections (which answers questions like "what's next?" and "how did I solve this obscure problem?!").

all2 5y ago

If you're in software engineering, Jupyter notebooks support a large variety of languages now. I keep handwritten notes in a notebook of recipes I particularly like. These are useful for making adjustments over many preparations. I guess the bottom line is being willing to enforce an organization on your notes, and then iterate towards a format that makes sense to you. For the recipe notes, time series (like journal entries) works well. Or inline. It depends on the note. Ingredient substitutions can be time series, but wrong measures specified in the recipe would be an inline correction. > engineering/scientific computing Definitely Jupyter notebooks or other "literate computing" methods. Emacs org-mode has literate programming stuff, but honestly Jupyter is easier to use.

visviva 5y ago
Thanks - I'm not a software engineer, but I write a lot of code anyway. I've used Jupyter notebooks to good effect in the past for smaller, self-contained projects, but they don't quite seem to fit my workflow well enough for me to use them for everything. You make a good point about picking something and enforcing it - I just have a hard time even finding something to start iterating on. Maybe I just need to be more creative about how I think about/use Jupyter notebooks...
- rmnclmnt 5y ago
  You can use use Jupyter Notebooks like Markdown files, with the added benefit of code blocks being executable for reproducible work. Then, just throw your notebooks into MkDocs (using a plugin) and voila, you get a statically hostable version of your notebooks! Easily coupled with Git and CI/CD practices, you have a nice, easy to use setup for prodyctive work and knowledge sharing if needed!
- all2 5y ago
  > something and start iterating Go with a standard time-series journal. This is about the simplest way to do things. It works well enough for me with minor additions. Eventually you've referred back to your journals enough to know what kinds of information you'll be looking for. I've gotten part way there, so far. I'm still trying to sort out details for a lot of stuff; where does the plan of a project live? In the journal? In the projects files? What about the day-to-day entries? What about troubleshooting epilogs? Like I said... I'm still trying to sort out where each of these live. I'm tempted to just throw all the things in project folders (because they're all projects anyway) and then grep for stuff if I need to find it.
  - all2 5y ago
    But putting everything in one folder still doesn't satisfy my itch to see my journal from two views: time-series (which answers "what did I do X days ago?") and project collections (which answers questions like "what's next?" and "how did I solve this obscure problem?!").