Zettelkasten method and math related sciences

ulver48

Hello everyone,

I am an electrical engineer PhD student and I try to use my Zettelkasten effectively, but there is a thing that puzzles me. In his book "How to take smart notes", Sohren suggests taking small notes for your slip-box. In my case, many of the ideas involve mathematical equations and graphs. I currently use the Zkn3 app, which sadly does not have math notation support. I would like to hear, how the physicists, mathematicians and engineers among you, build a Zettelkasten with equations and stuff.

Is it a good idea to have figures and mathematical equations on a Zettel ?

Any advice would be welcome.

Thank you.

boxcariii

My research area is in machine learning and I deal frequently with equations and figures. For the figures, I just add them as images within the note. If it is a table from a paper, normally I just copy the relevant information into a markdown table.

For equations, I use latex. Many programs will render latex equations within markdown. My current workflow is to use The Archive. If I am looking at a note with a significant amount of math (or figures), then I use the shortcut to open it in an external editor (Typora). So far this is working well for me, but would be interested if anyone has a better approach.

GeoEng51

@ulver48 You might have a look at a similar thread that I started about 6 months ago:

https://forum.zettelkasten.de/discussion/1192/is-there-a-way-to-insert-mathematical-functions-while-writing-a-note-in-ta#latest

Splattack

Hey Ulver, I'm a physicist. Keeping everything simple, you could have your ZK notes in

• one folder as LaTeX files (light-weight text files)
• another folder with photos from the blackboard, book pages from the library, screenshots from articles. Graphs as png images.
• a third folder with all the pdf articles you accumulate. The standard convention for the filename of articles is (Author'sLastName)(YearPublished). Remember Smith2003.pdf and you'll be set. If there are more authors then SmithJ2003 for Smith and Jones and if Smith published two papers in the same year go for Smith2003a and Smith2003b. Keeping the convention for filenames like this is great for referencing to those articles with other people, everyone will know which paper you're talking about.

Lastly, write the equations in LaTeX in your text files. Ask faculty members for a little guidance on LaTeX if you're new to it. Then hit StackExchange or the hundreds of tutorials online. LaTeX templates are a great resource to learn from.

Use BibDesk, Mendeley or something similar for managing your articles and books. In those reference managers your Smith2003.pdf will have the full title, year, authors, abstract, tags.

Use structure notes as much as possible so your notes are organized and allow your ZK to be scalable.

Will

@ulver48 said:
Is it a good idea to have figures and mathematical equations on a Zettel ?

Any advice would be welcome.

My workflow is much like @boxcariii's

This screenshot shows my Latex and Mathjax workflow. The Archive is on the left, and the right is Marked 2 in the "Stream to Marked Preview" node. I haven't yet needed to delve into the latex and mathjax syntax much beyond mostly chemical formulas. There is lots of help here and all over the web on the syntax to use. You probably know all the correct markdown syntax already.

I was surprised to learn that most markdown previewers also support HTML. I have, at times, resorted to this as a time saver for simple, quick formulas.

How I use latex and mathjax may change and hopefully be easier when The Archive switches to Multi-Markdown as its editor. We'll see.

ulver48

Thanks guys for your responses.

I am using Latex for 5 years, so I am quite familiar with it. The thing is, that when you create a PDF from your tex document in Texmaker or another program, a whole bunch of files are created along with a PDF file. Doesn't it make the folder with your slip-box look cluttered ?

With respect to literature notes, I use a folder where I store all my pdf research articles with the format [Year-Author] Title, whereas I use the same format on Joplin, an opensource alternative to Evernote that allows creating notes with support for Latex based mathematical equations.

But with respect to the Zettelkasten, the only software that I found easy to use is the zkn3. I am on Linux, so I cannot use the Archive. Obsidian led to a very confusing slip-box and I quickly abandoned it. Is there any articles for building a Zettelkasten from tex files ?

Splattack

From what I found here https://tex.stackexchange.com/questions/11123/prevent-pdflatex-from-writing-a-bunch-of-files it seems like there's no way to prevent the extra files. If you were using macOS I'll recommend you Hazel https://www.noodlesoft.com/whats-new-in-hazel-5/. This application will filter out automatically any file inside a directory according to rules that you set up. In this case you'll filter any non pdf to be deleted or moved into another folder.

bmooers

Two follow-up comments:

Bibtex cite keys: I too use the first author's last name and year in the cite key, but I also append the title written in PascalCase because I want to be certain that I am selecting the correct citation. It is easy to configure JabRef to autogenerate such cite keys. I start typing the cite key name, and the editor (e.g., emacs, vim, sublime text, VSC, overleaf, etc.) will autosuggest some candidates. I select the correct one, and the editor autocompletes the cite key. This does lead to long cite keys, but this had not been a problem.

Equation rendering: One approach that avoids creating the intermediate LaTeX files is to rely on MathJaX to render the equations written in LaTeX in the Note file. The org-roam-server in emacs provides a Zettelkasten knowledge graph. You can hover the mouse cursor over the node representing a note and obtain a preview of the note with the equation rendered (see below).

Other zettelkasten software may also take advantage of MathJax in this fashion. Of course, this approach's disadvantage is that you are limited to viewing the rendered equation with the org-roam server. On the other hand, the org-roam server has been my main avenue of conversing with my zettelkasten.

If you turn on the emacs server, you can also open the corresponding Note file in emacs for editing when you click on the corresponding node in the graph. This is very useful for adding missing backlinks or for changing links. The graphical view is useful for ensuring that every Note has a forward and backward link because each link is displayed.

Org-roam uses org files for the notes. Org files are a form of markdown on steroids. Tables are easy to generate. You can import images. You can use LaTeX environments or org-markdown or both in the same file.

Of course, the problem with org-roam is the need to learn the fundamentals of emacs. You only need to use about ten key sequences to add Notes and edit them efficiently in emacs. I am a novice emacs user and became productive in org-roam in a couple of days.

davecan

For anyone reading this later, Obsidian does support MathJax natively (via the usual $ and $$ delimiters) and renders it immediately in preview mode. You can also pin notes to panes and have preview mode side by side with editing.

I'm assuming other tools support this as well but I'm only directly familiar with Obsidian.

Will

@bmooers said:

Cool!

Might I request the plain text of these formulas so as I can experiment? Extracting this MathJax from an image is tedious. Thanks.

learning_ran

Extracting this MathJax from an image is tedious

"free" (free beer) software is ready for you: https://mathpix.com/

Will

@learning_ran said:

>

"free" (free beer) software is ready for you: https://mathpix.com/

Thanks, I've got to keep reminding myself that I'm not the first to have these issues and realize that there probably is an "app for that".

Snip worked great.

Here is what Snip scrapped from the image.

\begin{array}{l}
V_{\text {electrostatic }}=k_{e} q_{i} q_{j}\left(\frac{1}{r}+k_{r f} r^{2}-c_{r f}\right) \\
k_{r f}=\left(\frac{1}{r_{c}^{3}}\right)\left(\frac{e_{m l}-1}{2 e_{k d}+1}\right) \\
c_{r f}=\left(\frac{1}{r_{c}}\right)\left(\frac{3 \epsilon_{w l}}{2 e_{s o l}+1}\right)
\end{array}

Here is the output in Marked 2.

carolyn

+1 for Obsidian.

I clearly need to learn about what I can do with emacs. It's nice to have a Plan B.

davecan

Mathpix is pretty amazing, thanks for that link.

bmooers

I did use MathPix to translate the above three equations into LaTeX. I like how it gives you four alternatives for the LaTeX environment including the equation environment. a
I have been using Mathpix for over a year. Mathpix does remarkably well with blurry images from photocopies. It does make mistakes sometimes, especially with multi-line equations.
It can take some time to repair these mistakes, but in general, it saves time.

I found that I could be up and running in Obsidian faster than in emacs. However, I have to confess that I have additional reasons to learn emacs: https://www.youtube.com/watch?v=Zya8SfmCtFA&t=58s for efficient pdf and BibTeX file management, [ markdown), literate programming in multiple programming languages in one document (polyglot literate programming, like in Jupyter), access to snippet libraries, and autocompletion, even of English words written in prose and not just computer code.

I want to reduce the barriers between writing a manuscript and having those conversations with my kettelkasten.

I have been aware of zettelkastens for about two weeks. I understand that a couple of social scientists have benefited from using zettelkastens. Where are the examples of physical, chemical, biological, or biomedical experimentalists or engineers benefiting via increased productivity? Was it been more valuable for generating new hypotheses, writing literature reviews, generating points of discussion, or amplifying the significance of one's work in grant applications?

ctietze

Last weekend, a guide to get a Jupyter-like notebook in emacs org-mode was published that might be relevant: https://sqrtminusone.xyz/posts/2021-05-01-org-python/