I have a confession to make: I have an extremely goofy note-taking setup, one that only I could love. And I do love it. I could be using a thousand-and-one other applications that would do this better or give me an app or ... whatever.
But: I'll tell you a secret. I like designing software for myself. I know this can be a liability in the long run, but when I build software I designed for myself, I get to build it exactly the way I like it. And it'll probably break over time as the other software I depend on breaks, but... that's fine. It's all sandcastles anyway.
So, with that said, what's likelike?
This is to say, as part of my ersatz note-taking setup, I tend to create "link dump" text files in markdown. They look something like this:
* [An Axiomatic Basis for Computer Programming - Hoare-CACM-69.pdf](http://sunnyday.mit.edu/16.355/Hoare-CACM-69.pdf "An Axiomatic Basis for Computer Programming - Hoare-CACM-69.pdf")
* [ascii](https://ia600606.us.archive.org/17/items/enf-ascii/ascii.pdf)
* [cfgrammar - Rust](https://docs.rs/cfgrammar/latest/cfgrammar/ "cfgrammar - Rust")
* [lrlex - Rust](https://docs.rs/lrlex/latest/lrlex/ "lrlex - Rust")
* [lrpar - Rust](https://docs.rs/lrpar/latest/lrpar/ "lrpar - Rust")
I build up links in Firefox as tabs, then I use a firefox
extension to copy
multiple tabs of links into markdown. I then use a note bash script to look
up my last note title "link dump" and append them to the file. Occasionally I
start new link-dump documents when the old ones grow too big.
When I've read links, I'll add little bullet-point list notes under the link:
* [An Axiomatic Basis for Computer Programming - Hoare-CACM-69.pdf](http://sunnyday.mit.edu/16.355/Hoare-CACM-69.pdf "An Axiomatic Basis for Computer Programming - Hoare-CACM-69.pdf")
- notes:
- axioms, huh? never heard of 'em.
- tags: pdfs, computer science, etc
* [ascii](https://ia600606.us.archive.org/17/items/enf-ascii/ascii.pdf)
* [cfgrammar - Rust](https://docs.rs/cfgrammar/latest/cfgrammar/ "cfgrammar - Rust")
* [lrlex - Rust](https://docs.rs/lrlex/latest/lrlex/ "lrlex - Rust")
* [lrpar - Rust](https://docs.rs/lrpar/latest/lrpar/ "lrpar - Rust")
Then I save it.
It occurred to me that I'd really love to share these results, as so much of my work lately has been "read a link, compose some thoughts, find some more links." I really admire Simon Willison's weblog and would like to use my notes to bake more useful content for my personal website. It occurs to me that it'd both be useful for me (hey! I get to lean on hypermedia tags!) and maybe-useful for other people. And with Twitter's recent decline, I'd really like to spend more time on my corner of the internet, which I can guarantee won't be purchased by a goofy hyperbillionaire anytime soon.
(My business email is, however, listed in my Git commits on this repo, should a goofy hyperbillionaire read this README and think, "hey, I'd love to buy a worthless link dump site." Who knows, maybe they want to speedrun Digg by skipping some steps? I'm not a hyperbillionaire.)
So this works, in theory, in four steps:
- There's a link ingestion CLI. You point it at markdown docs and a SQLite database and it updates or creates links in the database based on what it finds. Honestly, it didn't have to be SQLite, but I wanted a chance to play around with SQLX in rust again.
- There's a Zola content creation CLI. You point it at a Zola directory and a SQLite database and it spits out re-formatted markdown in a prescribed fashion with all of the frontmatter data set just so.
- There's some amount of bash that runs locally (somehow, maybe via launchctl?) and makes sure to run ingestion against my notes directory every so often; downloading, updating, and re-uploading the sqlite store to some internet bucket somewhere.
- There's some amount of bash-embedded in YAML that runs the second step whenever I deploy my blog by downloading the sqlite database from an internet bucket.
If you're thinking: "hey, this is a lot of steps to avoid just writing frontmatter in a blog on a regular basis"; buddy, you're right. But I've also proved to myself the value of removing ANY and ALL friction from habits I wish to develop; I warm myself with the trash fire of habits I tried to develop that were just "too hard" so I didn't end up practicing them.
All of this code is MIT licensed, but why would you want to use it? You almost certainly don't write notes in Vim then rely on Dropbox to send 'em every which way, including to a GitHub action to render them as a Zola project. Go look at Obsidian. You'll be a lot happier!