If you're someone who thinks & processes information mainly in the form of Stories & Metaphors, then organization (and recall) becomes organic/natural.
Also on a side note, we must all be wary of epistemophilia!
I try to do mind maps sometimes, but on day to day basis when reading articles to save for later, Google keep with relevant tags.
But before that, always define whether the information you capture is useful or volatile. If useful, is it today, next week or someday? Someday is probably never ;-)
Information capture happens in a variety of media and goes into a physical or digital "inbox". I have both kinds because sometimes I take a note with pen and paper, or I have business card, or I have my cellphone and I can take a picture.
My inbox includes the "pictures" and "downloads" folders on all of my devices. I link my devices together using SyncThing.
Then I have an org-mode file I keep my projects, someday-maybe's, and project archives in. My reference "file" is a 1/2Tb drive split into folders. My file organization is pretty flat right now, so at the top level I have "school" and "recipes" and "day trips", etc.
This is an utter cludge, but it does most of what I need it to. I've actually adapted it some to use at work and that has allowed my to stay very organized with work projects.
It is intended for memorizing things, but the fact that it forces you to divide knowledge into pieces small enough to fit on a flash card also helps to organize it.
And you can have different card stacks in Anki, so there also is categorization of knowledge.
I just switched IDEs/operating systems, so I'm using Anki to help me learn that Cmd + Shift + Z is revert changelist and so on. When I'm waiting for something to compile, I tap Anki and let it reinforce my memories that were about to expire.
I crack open a book about the topic (usually from manning publishing). Self-document my learnings in the form of tree-structured bullet-point notes. Usually in a git book space.
I can always revisit the notes to quickly refresh myself on the topic (if a context-switch happens later down the line, or if the topic is not related to my day2day work).
I've successfully used this approach to onboard myself for a Spring project (known to have a steep learning curve). It was my second time using Spring, with an 8 month time separation. The ramp-up/refresh took about a week.
There were also some positive unintentional side-effects of my approach. I keep my gitbook notes public. I found that making my notes public ended up helping others on other teams in my company. Specifically, more experienced Spring devs who were new to Webflux were able to use my notes to jump right into a client's code base.
I've used personal wikis and knowledge bases for myself and they work okay. vimwiki has been useful for me.
I find that I do some pretty cool stuff and then totally forget how I did it a year or 5 years later, mostly with coding. The open source projects I documented and explained are things I myself google to help me set them up in the future, so a way of codifying my knowledge has been to try to write about it and put that in public.
You could do a lot worse than read about the concept of a Zettelkasten or Commonplace book. Here's my notes on the two. I am currently using the app Bear to take notes on everything and anything, tagging as I go, and trying my best to link notes to other meaningful notes. Less formal than a Zettelkasten but more digital than your average commonplace book.
A different concept, but also with relevance to storing information for a lifetime. From my brief readings, there’s more emphasis placed on two stages of note. The first is a quick jot - an addition. The second is a more thought out, self-contained existence of the note. There’s also focus placed here on _linking_ notes together.
Something about like... a syntax similar to Markdown that is [[deep-linkable]] or something that allows you to... create really deeply nested queries (almost like a GraphQL vibe) to finds things that are "loosely/tangentially related"
Something like... if you were to go to Wikipedia and play that game where you start on one page (a kind of bird) and try to get to a different page that is 50-60 deep/nested links away (a World War II event or something).
I personally am not a big note taker, but boy oh boy do some seem obsessed with special note-taking syntaxes/apps.
I've been really happy with Notion (https://www.notion.so). I use standard pages with subcategories for personal notes/knowledge.
I wanted to go deeper for my company, so I bought the Notion system called Bulletproof workspace (https://www.notion.vip/bulletproof/) in December after er doing. ton of research, but found it a bit overkill. In retrospect, I think it was just my newness to Notion that made it seem so.
However, once we reached a certain point where we were overwhelmed and completely unorganized. I finally invested the time to learn how to use Notion and Bulletproof Workspace and it was amazing, now I use it for everything from onboarding, to brainstorming, to-do's, etc. It's quite impressive.
I highly recommend checking them out if you're looking to get organized. The flexibility is what really made it work for me since it works for everything. It's nice just to have 1 tool.
A per project markdown file that usually lives somewhere inside /doc, called dev-note.md (pun intended). It is handled append only and therefore quite messy, but grep is your friend.
One's knowledge is too valuable and important to entrust it to a particular binary format that can soon no longer be read. Plain text is durable, portable, easy to process using UNIX command line tools, it can be full-text indexed with a reasonable overhead.
It can be version controlled easier than binary formats or formats with heavy markup.
I often summarize scientific articles, write down new ideas or need to preserve how I did something (run a system, install a tool) for later replay, and plain text is great for that for the most part. Occasionally, I used LaTeX
commands e.g. for embedded $ maths $ or #hastags to tie
together files by topic for indexing.
Importantly, my workflow is OS independent (I mostly use Linux and occasionally MacOS X/iOS) and editor agnostic (I use Sublime and Emacs).
I would very much like to hear from others how they address their KM needs.
Very interesting, I also tried several approaches and came finally to text files. For me the most important feature are full text search and some kind of formatting, including basic support for images.
As I couldn`t find a good tool I wrote my own Wiki server - 13 years ago and still using it every day:
https://moasdawiki.net/
Why not use markdown instead instead of .txt? All good editors (including vim etc) have great support for markdown and you get plenty of great simple features without much crap getting in the way.
I see two sides of this. There's information collection and knowledge keeping.
The information collection: When I'm learning new things and sorting by what is interesting I simple bookmark a lot of stuff. It could be a blog post, a research paper, a book recommendation etc. All this is raw information until I take the time to consume it and digest it. Here I use https://lxi.ai/ (disclaimer: I built this) to keep my bookmarks organized with a lil ML.
The knowledge keeping: While I consume/digest I take notes in obsidian. I like working in markdown and having everything stored locally is something that keeps me comfortable. The real key here is putting in the time and actually merging what I am learning with what I already have in my obsidian "second brain".
Obsidian was the mind opener for me. Simple to use. Backuped on Dropbox on my various machines, it is my ubiquitous personal knowledge base.
The fact that you can copy/paste HTML from web pages into it makes me rebuild knowledge I get from the internet into my own version.
Really really a fundamental tool for me.
The Vi of knowledge.
PS: I don't know the other tools (notion, roam,...)
PPS: a big downside in Obsidian, in my humble opinion, is publishing that knowledge on the web. Probably a static CMS on top of the .md file generated with Obsidian can do the trick. But it is a tedious step that I never had/wanted to investigate.
The is the Obsidian Publish service exactly for that usecase. (https://obsidian.md/publish), although it is paid. I'm sure there will be soon, if not already, an open source extension that allows you to do that.
Several people I know simply donate money to the Obsidian project directly. But I agree that paying for their publishing service is more a win-win. You help the project PLUS you get the publishing service. I will probably do that.
(and yell that I need this or that feature, that the service does not provide yet :)
I have not tried this with Obsidian yet, but I know that Markdeep allows you to render .md files on the fly as HTML by appending some javascript to the unmodified markdown.
The .md publishing step you’re asking about is very neatly handled by Neuron (https://neuron.zettel.page/). The community is helpful and the original author of the project is a great guy and accessible on matrix.
you’ll be up and publishing in no time (for free and automatically to GH pages if you want)
Yeah, it can tie in really nicely with a static site generator.
I use Notable rather than Obsidian, but it works the same (folder of .md files). My blog builds with Bridgetown, and I wrote a little bit of extra code to handle the wikilink formatting and build up an incoming link list, but it runs pretty cleanly. https://blog.tracefunc.com/notes/
Everything starts in my emacs scratch buffer (the default file opened upon startup), which is an org-mode file, and when some category of stuff gets big I move it into its own file in Dropbox. The content is just links and written-out thoughts and questions.
After what I suppose has been 15 years of searching, I have settled on the Zettelkasten method using Doom Emacs (Vim-like keybindings) with org-mode and org-roam (https://www.orgroam.com). I struggle to think of anything that could beat it.
I used to use trillium notes (which is really good) for GTD and idea mapping for a few months but stopped. I just use a big three ring binder with some five star hard to rip paper now with some plastic sticky tabs for categories and write down things that are important. The barrier to taking down and recalling info is both low and high enough in just the right way that I remember things I want to remember and can recall important enough forgotten things by referring to my notebook. I just reference other ideas by category and some other detailing info and flip through until I find what i'm looking for.
Digital notes are good for a lot of people and a lot of situations but not for my long term very general knowledge based tasks. For specific commands/documentation/syntax/techniques yeah digital notes are nice. For Brain storming, and connecting ideas though? I fully believe paper works best. It's more flexible.
Plus it is known that when you write things down with ink and paper, your recall tends to be higher. That doesnt mean its perfect for everyone, but a larger than just an average amount of people.
I like your idea - what ever happened to that e-ink note pad one could write on and it saves a digital copy?
110 comments
[ 2.8 ms ] story [ 178 ms ] threadAlso on a side note, we must all be wary of epistemophilia!
Information capture happens in a variety of media and goes into a physical or digital "inbox". I have both kinds because sometimes I take a note with pen and paper, or I have business card, or I have my cellphone and I can take a picture.
My inbox includes the "pictures" and "downloads" folders on all of my devices. I link my devices together using SyncThing.
Then I have an org-mode file I keep my projects, someday-maybe's, and project archives in. My reference "file" is a 1/2Tb drive split into folders. My file organization is pretty flat right now, so at the top level I have "school" and "recipes" and "day trips", etc.
This is an utter cludge, but it does most of what I need it to. I've actually adapted it some to use at work and that has allowed my to stay very organized with work projects.
Writing helps structure and encode information in my mind
If I forget it, it’s on google, and I often land back at my old posts that I reread.
It might also help other people, but really future me (and possibly close colleagues) are the primary audience.
It is intended for memorizing things, but the fact that it forces you to divide knowledge into pieces small enough to fit on a flash card also helps to organize it.
And you can have different card stacks in Anki, so there also is categorization of knowledge.
Cf. the "Zettelkasten", a physical flashcard system which is actually intended at organization, not memorization: https://en.wikipedia.org/wiki/Zettelkasten
Anki can be used to emulate that digitally.
I can always revisit the notes to quickly refresh myself on the topic (if a context-switch happens later down the line, or if the topic is not related to my day2day work).
I've successfully used this approach to onboard myself for a Spring project (known to have a steep learning curve). It was my second time using Spring, with an 8 month time separation. The ramp-up/refresh took about a week.
There were also some positive unintentional side-effects of my approach. I keep my gitbook notes public. I found that making my notes public ended up helping others on other teams in my company. Specifically, more experienced Spring devs who were new to Webflux were able to use my notes to jump right into a client's code base.
I find that I do some pretty cool stuff and then totally forget how I did it a year or 5 years later, mostly with coding. The open source projects I documented and explained are things I myself google to help me set them up in the future, so a way of codifying my knowledge has been to try to write about it and put that in public.
I set up development environments for myself a lot, so I wrote an article about how I do it https://modfoss.com/creating-my-development-environment.html and then put the code on GitHub as well https://github.com/symkat/modfoss_devel So if I don't do it for a while, I'll have a starting point and me-from-the-past explaining what I did and why.
1. Commonplace Book This is a book of thoughts, findings and general collections of information captured throughout your life and brought together into one place. [How And Why To Keep A “Commonplace Book” | Thought Catalog](https://thoughtcatalog.com/ryan-holiday/2013/08/how-and-why-...) [Commonplace Books Part 3: Choosing a System - GeekDad](https://geekdad.com/2020/03/commonplace-books-part-3-choosin...)
2. Zettelkasten
A different concept, but also with relevance to storing information for a lifetime. From my brief readings, there’s more emphasis placed on two stages of note. The first is a quick jot - an addition. The second is a more thought out, self-contained existence of the note. There’s also focus placed here on _linking_ notes together.
A good example of someone’s Zettelkasten is [[Andy Matuschak’s Note Collection]]: https://notes.andymatuschak.org/About_these_notes?stackedNot...
[Zettelkasten - Wikipedia](https://en.wikipedia.org/wiki/Zettelkasten) [Bear App: A Solid Zettelkasten Solution for iOS Users — Mental Pivot](https://mentalpivot.com/bear-app-a-solid-zettelkasten-soluti...)
Something about like... a syntax similar to Markdown that is [[deep-linkable]] or something that allows you to... create really deeply nested queries (almost like a GraphQL vibe) to finds things that are "loosely/tangentially related"
Something like... if you were to go to Wikipedia and play that game where you start on one page (a kind of bird) and try to get to a different page that is 50-60 deep/nested links away (a World War II event or something).
I personally am not a big note taker, but boy oh boy do some seem obsessed with special note-taking syntaxes/apps.
What I would love is a way to index & search ALL my digital assets: audio, video, images, docs, urls, code and binaries
And it would help if I could use a query syntax to create custom time ranges
Recently noticed Weezer frontman Rivers Cuomo taught himself Python via Harvard's CS50 to organize his music demos dating back decades!
And I've come to realize virtually every one could use such a personal digital archive interface ;)
https://techcrunch.com/2020/11/25/rock-star-programmer-river...
I wanted to go deeper for my company, so I bought the Notion system called Bulletproof workspace (https://www.notion.vip/bulletproof/) in December after er doing. ton of research, but found it a bit overkill. In retrospect, I think it was just my newness to Notion that made it seem so.
However, once we reached a certain point where we were overwhelmed and completely unorganized. I finally invested the time to learn how to use Notion and Bulletproof Workspace and it was amazing, now I use it for everything from onboarding, to brainstorming, to-do's, etc. It's quite impressive.
I highly recommend checking them out if you're looking to get organized. The flexibility is what really made it work for me since it works for everything. It's nice just to have 1 tool.
One's knowledge is too valuable and important to entrust it to a particular binary format that can soon no longer be read. Plain text is durable, portable, easy to process using UNIX command line tools, it can be full-text indexed with a reasonable overhead. It can be version controlled easier than binary formats or formats with heavy markup.
I often summarize scientific articles, write down new ideas or need to preserve how I did something (run a system, install a tool) for later replay, and plain text is great for that for the most part. Occasionally, I used LaTeX commands e.g. for embedded $ maths $ or #hastags to tie together files by topic for indexing.
Importantly, my workflow is OS independent (I mostly use Linux and occasionally MacOS X/iOS) and editor agnostic (I use Sublime and Emacs).
I would very much like to hear from others how they address their KM needs.
As I couldn`t find a good tool I wrote my own Wiki server - 13 years ago and still using it every day: https://moasdawiki.net/
The information collection: When I'm learning new things and sorting by what is interesting I simple bookmark a lot of stuff. It could be a blog post, a research paper, a book recommendation etc. All this is raw information until I take the time to consume it and digest it. Here I use https://lxi.ai/ (disclaimer: I built this) to keep my bookmarks organized with a lil ML.
The knowledge keeping: While I consume/digest I take notes in obsidian. I like working in markdown and having everything stored locally is something that keeps me comfortable. The real key here is putting in the time and actually merging what I am learning with what I already have in my obsidian "second brain".
PS: I don't know the other tools (notion, roam,...) PPS: a big downside in Obsidian, in my humble opinion, is publishing that knowledge on the web. Probably a static CMS on top of the .md file generated with Obsidian can do the trick. But it is a tedious step that I never had/wanted to investigate.
https://casual-effects.com/markdeep/
you’ll be up and publishing in no time (for free and automatically to GH pages if you want)
I use Notable rather than Obsidian, but it works the same (folder of .md files). My blog builds with Bridgetown, and I wrote a little bit of extra code to handle the wikilink formatting and build up an incoming link list, but it runs pretty cleanly. https://blog.tracefunc.com/notes/
Then I use this Custom New Tab URL extension to turn my new tab page to Spaceli: https://chrome.google.com/webstore/detail/custom-new-tab-url...
Digital notes are good for a lot of people and a lot of situations but not for my long term very general knowledge based tasks. For specific commands/documentation/syntax/techniques yeah digital notes are nice. For Brain storming, and connecting ideas though? I fully believe paper works best. It's more flexible.
I like your idea - what ever happened to that e-ink note pad one could write on and it saves a digital copy?
Probably https://remarkable.com/
Joplin is opensource and uses Markdown files. Syncs with WebDAV. It's the best I found till now.
LibreOffice is for tables since Markdown sucks for them