Evolving my note-taking processes(codethrasher.com) |
Evolving my note-taking processes(codethrasher.com) |
https://GitHub.com/samsquire/ideas https://GitHub.com/samsquire/ideas2 https://GitHub.com/samsquire/ideas3 https://GitHub.com/samsquire/ideas4 https://GitHub.com/samsquire/startups
I create a new markdown heading for each entry and write.
It's searchable and the data is easy to synchronize, backup and use and the solution shall last for multiple decades maybe even longer.
It's also indexed by Google.
I'm still tweaking my first journal that I created in 2013.
At one point I tried to love Emacs and I am yet to use org mode. I actually use the GitHub interface mostly to update my journal. And before that vim and lately IntelliJ which includes preview features.
I would recommend if you want to write notes or create an external mind to improve your thinking just write. The tool you use doesn't really matter. It's the quality and reward from writing and rereading what you wrote.
https://lancebachmeier.com/other.html
Then I moved to GH pages and a simple theme, so that I could type directly in the browser on any device:
https://bachmeil.github.io/online-writing/
There was still too much styling on that page for my taste. The full CSS for my new site is
body { max-width: 950px; margin: auto; padding: 8px;} pre { border: 1px solid gray; border-radius: 6px; padding: 8px; }
That feels more natural to me. Site here:
Agreed. Personally, I still use paper + pencil for a lot of things. Even if I never rustle through my stacks of notes to find it again, I already got a big chunk of the value from just writing it.
Aside: I find rustling through my notes to be useful in itself, even though it's inefficient on the surface. Electronic systems are so good at finding exactly what you want as quickly as possible that the "browse around and see what you bump into" experience can get lost.
Sometimes I'll just happen to see diagram X next to diagram Y and come up with a new insight, where if I only ever saw what I was specifically looking for, I'd lose that. Though I do waste a lot of time, too (:
My stack of paper drawings and diagrams and notes was thrown away by accident. So I only have my digital notes left.
As it was for Richard Feynman, writing is thinking itself.
If someone checks my journal on Monday, unless they scroll to the bottom to check for new journal entries on Tuesday, they shall miss the journal entries I added Tuesday. (Unless they come back but I feel I have one chance to gain a regular reader)
I submit to HN when I reach a new batch of 100-300 entries.
I don't want people to not see the new entries.
I only tweak or add to older journal repositories, I don't add new entries to the bottom of those repositories.
Interesting process, though.
thanks
I use orgmode for to-do lists too, but it's a little weak there. I think an effective to-do list management implementation is tightly coupled with calendars...
It was such a welcome surprise to me that when I got the M2 Macbook air, Apple calendar synced to my google calendar so easily... now my to-do list/calendar schedule is never too far away from me. The apple watch makes a vibration ding for events in my calendar... as does the iphone, and I can CRUD calendar entries on any of these interfaces. I would have _loved_ to be able to have org-mode as one other place where my calendar/todo list is staring at me in the face, as I now use emacs for everything, but sadly that's iffy. Yes, solutions purportedly exist (e.g. https://github.com/kidd/org-gcal.el9 ) but none work reliably and require too much finagling. Linked solution doesn't work as Google deprecated the ability to retrieve OAuth 2.0 tokens.
That is a fascinating take, because I am of the opposite opinion. I don’t think a todo system that involves scheduling tasks can work well, because it is impossible to figure out ahead of time whether too many or too few tasks are scheduled on a day, and because it is hard to reprioritize scheduled tasks. Instead, imho, the only role dates play in an effective todo system is to label tasks with their due date, but never to schedule a task on its due date (as that is too late) but only to inform the setting of priorities. Instead in my view the critical aspect of an effective todo system is a clear priority of tasks, easily reprioritized, so that working time involves only tackling the flat list of tasks in order of priority until the completion of the working day.
For me integration with email instead of calendar is super convenient though, because that is where tasks often stream into my todo list. But I never use unread or flagged status to mark mails as a todo, because email does not afford reprioritizing. I see email as a way station between other people and my todo list.
Like a mix between Logseq and Nextcloud, where I can still use CalDav and CardDav but I can link to tasks and calendar events and contacts from my notes.
And i can link to a certain pdf page right from my notes.. or even a certain page of a word document. Anything needs to be treated as a generic object of content that is linkable.
It's a shame the Ruby renderer used by GitHub and GitLab doesn't understand them. I've been intending to patch it.
In the meantime, one can always define a CI task that uses Emacs to export to HTML.
[1] https://en.m.wikipedia.org/wiki/Zettelkasten#:~:text=A%20Zet....
:)
[1] https://discord.gg/VsqazDhsES
[2] https://github.com/brettkromkamp/awesome-knowledge-managemen...
But I don't think any one note taking style or software would apply to 2 people. Everyone's thought process is so different. I tried doing the Zettlekasten but it didn't work out (for me). The worst problem with taking notes is finding them (and God knows all these tools don't make it easier). It's tragic to see that search is often last on the list.
What I wish for is search like Google (when it was better than it is now) for my own notes - intelligent, fast, and predictable.
I'm a software engineer. My thinking process gets better as I develop and implement things. I cannot write notes and then implement. There were times when I wished I took notes/planned before implementing, but nothing to make me regret. It is just more refactor and time consuming to rewrite things.
I feel like note taking is overrated and trendy like AI/ML (like DALL-E or Stable Diffusion). I wish someone can mentor me to takes notes and make understand the value of it. And I read a lot of blog post about it.
I've been using it for months for basically everything from project ideas to journaling and I've been impressed.
My iPad is still confined to paper replacement duty, and it excels.
eInk leaves a lot to be desired still but the input latency on the supernote seems comparable to the iPad. I think the tactile experience is superior and for me that sacrifice was worth it. I can go a few weeks without charging my supernote and my iPad requires a charge at least once every couple days. The supernote also seems better set up for bullet Journaling and what not. It definitely is exactly as offered - a notebook replacement. Nothing more and nothing less. I find that charming.
I often wonder who I'm writing them for, as apparently it's not me.
Question for anyone that consumes a lot of PDFs as I do: what do you use instead? Manual data insertion is an absolute no. So that leaves the Zotero alternatives, but quick search shows 0 mentions for any of them too.
FWIW, I settled on https://github.com/zadam/trilium which is very bare-bones but fully open source and self-hostable.
I want to take create a chronological journal where I track tutorials I've done (and how well, etc), but I want to add some sort of a tag wherever I feel like so that with a click I can find all my entries related to (say) React, or R (statistics), and so on.
If you're reading this, and have had luck with such a tagging mechanism, please share your solution.
FOSS only, of course.
I'd need my PC and 3 NAS drives (all different vendors/batches) to fail AND Backblaze and GitHub would need to both break before I worry about losing my notes.
I doubt I'd be able to lose them if I tried honestly. There'll be a copy on my laptop too now I'm thinking about it.
As for why? Because I can build my own systems no bother. I also like to have many ways of adding to my notes such as via phone, via command line, via the share menu in iOS, via web browser (github.dev), via Siri if I really wanted to.
Wrapping all this up into a nice working system took the better part of a few hours and I don't have to worry about anything being discontinued. Text files will probably still be usable 100 years from now, which is plenty for what I need.
E: Oh yeah there's another copy of the repo on both my phone and my iPad haha. Both autosync using iOS Shortcuts whenever I write a note on that device.
The likelihood of Google locking you out of your account seems higher than your own drive failing.
Using Google for this strikes me as much riskier, as well as just creepier.
If you buy an executive notebook and a really nice flowing pen like the G2, it’s a pleasurable practice.
Once I have it written down I feel much less stressed. I’ve never felt hat way about online notetsking
I started keeping a “log” back in 1999 just for work (I too am a software engineer). I note what I worked on and/or what problems I was trying to solve and/or what solutions I found. And some personal stuff too.
My inspiration is the idea of the scientist’s lab book where “if you didn’t write it down it didn’t happen.”
Countless times — even this past week - it has proven invaluable when I wanted to look up how I solved something before (“how do we generate those CSP headers on the fly?” “How did I mount that Docker container?”) or when no one knows why something was built a certain way (and I noted the date and time a manager made that decision).
After a few years I wind up with a few novels worth of entries (by volume) so it’s amazing how writing a little each day adds up.
As to format, it started as a .txt file in Emacs but has evolved into a moderately simple org-mode file.
Here's how I look at it. My programming notes have 2 purposes. #1 to keep me on track. #2 for a concrete log of events. I then rewind, playback and observe what and how I was thinking (almost like debugging by reading logs).
#1 helps with focus. I waste way too much time trying new and shiny libraries, reading articles, etc without making progress on the task at hand. If I simply focus on the task, I don't learn new things. This helps me find interesting tools, libs etc. and come back to them later. Also helps to organize important things first.
#2 Helps me find patterns. In the type of work I do, how I approach them, what I was thinking yesterday, how I debugged something 2 months ago. It has made me an overall better, or at least it gave me a feeling that I was getting better. Both are worth the effort.
So when you start, you will either overdo or underdo note taking. Both ends make the resulting notes less useful (usually called "waste"). Getting it right is a matter of knowing which end you are starting, make changes, see results and repeat till you see no improvement.
yyyymmdd-topic-title
I don't need tags. Willing to come across patronizing and say you don't either (not for personal knowledge management). Pretty soon you need to sort tags, find tags, cluster related tags, create tags of tags and other goofy gunk (see tiddlywiki to scratch this itch)
I keep things organized under 10 topics (so far used just 7). All plain text files. At the moment it lives in simplenote [2] (thats how I found the link above as a demonstration of its utility). I found notational velocity [3] and its alter ego NValt [4] useful when I had a Mac. Did not look for alternatives on windows. I just email myself these days. Look in my comment history for some tip around how I use that.
All of this is good for digital stuff. But for the deep, analog stuff that really matters, nothing, nothing beats the utility of taking notes in a 1 subject college notebook with a Pilot G2 ( i use a .38). I use the right side only (trade off for using cheap books). First page is the index. Each page numbered. I just make clusters of related words on each page. The main topic is circled to draw attention (4-5 per page). Anything related to it gets words around it. Details, references, long thoughts, go on the backside. Stacks well. I take pictures of good stuff. Don't fret if I lose any books (not lost any permanently till now). Do explore this route and see where that leads you. (hey a collge notebook is $00.75 in staples last I checked. What have you got to lose ?)
[1]: https://forum.literatureandlatte.com/t/large-scale-info-mana...
I struggled with them for years because in the back of my head they felt like a necessity.
And when I actively banished them from my note-taking flow, it was like the skies opened up. I realized that what I should be doing is simple "being verbose enough in the note taking such that searching later is robust against me trying to remember what I tagged things."
As in, it's a minimal marginal amount of effort to write for "future you" as if that person were a stranger, not someone who remembered your tags.
(fyi it's http://zim-wiki.org all day every day for what I do. I've tried most all the others and I keep coming back.)
[Foam](https://github.com/foambubble/foam) seems very similar but I haven't personally tried it.
I use neovim+vimwiki(in markdown mode)+telescope[0] and that setup has been working well for me. I don’t do tags per se but telescope should be able to search those.
The best outcome I’ve had for finding stuff in notes related to a particular concept is just by grepping through my notes, which are just plain text (and/or markdown) files, in a directory, that I “manage” with vim. If I know that I want something to be findable in the future I just make sure to sprinkle some keywords in the note. Since it’s just my notes and it doesn’t need to conform to any formal system, I can also drop images into the dirs (diagrams, screenshots, etc) and either link them from markdown or just name them so they come up in a ‘find’.
For a long time this even worked well when I had irc logs saved near the notes so I can grep them too, because I often remember having a conversation but not the details. But now with less irc and more slack/etc this isn’t as fruitful.
However, rendering is missing some features. By using a CI task that generates HTML (using Emacs, but on the cloud) it's possible to overcome this (hopefully temporal) limitation.
They are terrible. Not a single useful catalog of information uses the concept of willy-nilly off-the-fly generated tags. Their degree of terrible-ness is exemplified on any wordpress website. Absolute junk.
Folders are awesome, otherwise. Not in this case. Because you need to "open" a folder to look in. Can't eyeball closed folders. Easy to overdo with too many folders within folders. Easy to underdo putting too many unrelated things into one. Easy to fo wrong with naming the folders. Naming folders and find the "right home" for a file becomes a chore. Moving from one folder to another needs both to be well defined not overlapping.
A list of files is much harder to get wrong. In one extreme case, you have 1 big text file. The other, you have a huge list of text files. Both are not a nightmare to fix (sed+awk+grep is all you need). Works well on all OSes including those from the 70s and the ones on all my phones. Copy and paste friendly. Super easy to backup. Also diffs well. Only issue i have is not being able to use images. Thats why I use email now. Just works. Every-fking-where.
I think it's because you put a number in your titles so they all looked like blog spam articles '100 ways to improve you sex life'
I am always looking for high quality things to read. So I'm interested in other people's blogs such as Joel on Software or Jeff Atwood or Joe Duffy.
If you don't find it interesting then you are not the intended audience.