For a long time I've been struggling with what to do about organizing information. I'm something like a hoarder when it comes to information. To be charitable, I could say I'm a super-encounterer. I'd considered a number of pretty complicated setups. I wanted to track such things as who referred me to the thing, or if not a person, how I found it, when I found it, scrape the page and try to discern what subject it is, and a bunch of other nonsense.
This is a generalized XY problem (when you ask how to do X because you're trying to do Y, and you think X will help you do Y). I want to solve a few specific problems. I'd like to have fewer tabs open (the cmd - shift - [/] presses actually contribute to RSI, I have to do so many to switch tabs). I'd like to be able to track down things I found, or was interested in, in the past, and maybe know how I found them.
This is an XY problem since, when I wanted to basically have some notes on things I find, I decided I needed to build some software to manage this, and just gave up on doing it otherwise. I basically chose helplessness. Somehow, the frustration built up enough (or perhaps it was the motivation to move on and process more of this information, rather than just drowning in it) that I had a moment of clarity. I can just write some goddamn notes.
When you have the option to build some complicated, interesting, or powerful technology to solve a problem – especially a social, habits, or otherwise human problem – try to think of how you could do it without building anything at all. In my case, rather than writing any software at all, I decided to just put some notes in markdown files, in git. I'll write them up daily, and probably roll them up on a few other time scales. For now, weekly seems certain, but I'm not sure if monthly or quarterly (or both) will be the next appropriate unit. I'll figure it out as I go.
Credit for the daily/weekly rollup trick goes to Greg McKeown, as presented in his book Essentialism, and subconscious inspiration for the markdown/git pairing should probably go to Jamie Brandon, as I've been reading his notes on GitHub right before deciding to do this.